\input texinfo   @c -*-texinfo-*-
@c vim: filetype=texinfo
@c %**start of header (This is for running Texinfo on a region.)
@setfilename gawk.info
@settitle The GNU Awk User's Guide
@c %**end of header (This is for running Texinfo on a region.)

@dircategory Text creation and manipulation
@direntry
* Gawk: (gawk).                 A text scanning and processing language.
@end direntry
@dircategory Individual utilities
@direntry
* awk: (gawk)Invoking Gawk.                     Text scanning and processing.
@end direntry

@c Enable better indexing, requires texindex from Texinfo 6 or later.
@tex
\global\usebracesinindexestrue
@end tex

@ifset FOR_PRINT
@tex
\gdef\xrefprintnodename#1{``#1''}
@end tex
@end ifset

@ifclear FOR_PRINT
@c With early 2014 texinfo.tex, restore PDF links and colors
@tex
\gdef\linkcolor{0.5 0.09 0.12} % Dark Red
\gdef\urlcolor{0.5 0.09 0.12} % Also
\global\urefurlonlylinktrue
@end tex
@end ifclear

@ifnotdocbook
@set BULLET @bullet{}
@set MINUS @minus{}
@end ifnotdocbook

@ifdocbook
@set BULLET
@set MINUS
@end ifdocbook

@iftex
@set TIMES @times
@end iftex
@ifnottex
@set TIMES *
@end ifnottex
        

@set xref-automatic-section-title

@c The following information should be updated here only!
@c This sets the edition of the document, the version of gawk it
@c applies to and all the info about who's publishing this edition

@c These apply across the board.
@set UPDATE-MONTH February, 2018
@set VERSION 4.2
@set PATCHLEVEL 1

@set GAWKINETTITLE TCP/IP Internetworking with @command{gawk}
@ifset FOR_PRINT
@set TITLE Effective awk Programming
@end ifset
@ifclear FOR_PRINT
@set TITLE GAWK: Effective AWK Programming
@end ifclear
@set SUBTITLE A User's Guide for GNU Awk
@set EDITION 4.2

@iftex
@set DOCUMENT book
@set CHAPTER chapter
@set APPENDIX appendix
@set SECTION section
@set SUBSECTION subsection
@set DARKCORNER @inmargin{@image{lflashlight,1cm}, @image{rflashlight,1cm}}
@set COMMONEXT (c.e.)
@set PAGE page
@end iftex
@ifinfo
@set DOCUMENT Info file
@set CHAPTER major node
@set APPENDIX major node
@set SECTION minor node
@set SUBSECTION node
@set DARKCORNER (d.c.)
@set COMMONEXT (c.e.)
@set PAGE screen
@end ifinfo
@ifhtml
@set DOCUMENT Web page
@set CHAPTER chapter
@set APPENDIX appendix
@set SECTION section
@set SUBSECTION subsection
@set DARKCORNER (d.c.)
@set COMMONEXT (c.e.)
@set PAGE screen
@end ifhtml
@ifdocbook
@set DOCUMENT book
@set CHAPTER chapter
@set APPENDIX appendix
@set SECTION section
@set SUBSECTION subsection
@set DARKCORNER (d.c.)
@set COMMONEXT (c.e.)
@set PAGE page
@end ifdocbook
@ifxml
@set DOCUMENT book
@set CHAPTER chapter
@set APPENDIX appendix
@set SECTION section
@set SUBSECTION subsection
@set DARKCORNER (d.c.)
@set COMMONEXT (c.e.)
@set PAGE page
@end ifxml
@ifplaintext
@set DOCUMENT book
@set CHAPTER chapter
@set APPENDIX appendix
@set SECTION section
@set SUBSECTION subsection
@set DARKCORNER (d.c.)
@set COMMONEXT (c.e.)
@set PAGE page
@end ifplaintext

@ifdocbook
@c empty on purpose
@set PART1
@set PART2
@set PART3
@set PART4
@end ifdocbook

@ifnotdocbook
@set PART1 Part I:@*
@set PART2 Part II:@*
@set PART3 Part III:@*
@set PART4 Part IV:@*
@end ifnotdocbook

@c some special symbols
@iftex
@set LEQ @math{@leq}
@set PI @math{@pi}
@end iftex
@ifdocbook
@set LEQ @inlineraw{docbook, ≤}
@set PI @inlineraw{docbook, &pgr;}
@end ifdocbook
@ifnottex
@ifnotdocbook
@set LEQ <=
@set PI @i{pi}
@end ifnotdocbook
@end ifnottex

@ifnottex
@ifnotdocbook
@macro ii{text}
@i{\text\}
@end macro
@end ifnotdocbook
@end ifnottex

@ifdocbook
@macro ii{text}
@inlineraw{docbook,<lineannotation>\text\</lineannotation>}
@end macro
@end ifdocbook

@ifclear FOR_PRINT
@set FN file name
@set FFN File name
@set DF data file
@set DDF Data file
@set PVERSION version
@end ifclear
@ifset FOR_PRINT
@set FN filename
@set FFN Filename
@set DF datafile
@set DDF Datafile
@set PVERSION version
@end ifset

@c For HTML, spell out email addresses, to avoid problems with
@c address harvesters for spammers.
@ifhtml
@macro EMAIL{real,spelled}
``\spelled\''
@end macro
@end ifhtml
@ifnothtml
@macro EMAIL{real,spelled}
@email{\real\}
@end macro
@end ifnothtml

@c Indexing macros
@ifinfo

@macro cindexawkfunc{name}
@cindex @code{\name\}
@end macro

@macro cindexgawkfunc{name}
@cindex @code{\name\}
@end macro

@end ifinfo

@ifnotinfo

@macro cindexawkfunc{name}
@cindex @code{\name\()} function
@end macro

@macro cindexgawkfunc{name}
@cindex @code{\name\()} function (@command{gawk})
@end macro
@end ifnotinfo

@ignore
Some comments on the layout for TeX.
1. Use at least texinfo.tex 2016-02-05.07.
@end ignore

@c merge the function and variable indexes into the concept index
@ifinfo
@synindex fn cp
@synindex vr cp
@end ifinfo
@iftex
@syncodeindex fn cp
@syncodeindex vr cp
@end iftex
@ifxml
@syncodeindex fn cp
@syncodeindex vr cp
@end ifxml
@ifdocbook
@synindex fn cp
@synindex vr cp
@end ifdocbook

@c If "finalout" is commented out, the printed output will show
@c black boxes that mark lines that are too long.  Thus, it is
@c unwise to comment it out when running a master in case there are
@c overfulls which are deemed okay.

@iftex
@finalout
@end iftex

@copying
@docbook
<para>
&ldquo;To boldly go where no man has gone before&rdquo; is a
Registered Trademark of Paramount Pictures Corporation.</para>

<para>Published by:</para>

<literallayout class="normal">Free Software Foundation
51 Franklin Street, Fifth Floor
Boston, MA  02110-1301 USA
Phone: +1-617-542-5942
Fax: +1-617-542-2652
Email: <email>gnu@@gnu.org</email>
URL: <ulink url="https://www.gnu.org">https://www.gnu.org/</ulink></literallayout>

<literallayout class="normal">Copyright &copy; 1989, 1991, 1992, 1993, 1996&ndash;2005, 2007, 2009&ndash;2018
Free Software Foundation, Inc.
All Rights Reserved.</literallayout>
@end docbook

@ifnotdocbook
Copyright @copyright{} 1989, 1991, 1992, 1993, 1996--2005, 2007, 2009--2018 @*
Free Software Foundation, Inc.
@end ifnotdocbook
@sp 2

This is Edition @value{EDITION} of @cite{@value{TITLE}: @value{SUBTITLE}},
for the @value{VERSION}.@value{PATCHLEVEL} (or later) version of the GNU
implementation of AWK.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``GNU General Public License'', with the
Front-Cover Texts being ``A GNU Manual'', and with the Back-Cover Texts
as in (a) below.
@ifclear FOR_PRINT
A copy of the license is included in the section entitled
``GNU Free Documentation License''.
@end ifclear
@ifset FOR_PRINT
A copy of the license
may be found on the Internet at
@uref{https://www.gnu.org/software/gawk/manual/html_node/GNU-Free-Documentation-License.html,
the GNU Project's website}.
@end ifset

@enumerate a
@item
The FSF's Back-Cover Text is: ``You have the freedom to
copy and modify this GNU manual.''
@end enumerate
@end copying

@c Comment out the "smallbook" for technical review.  Saves
@c considerable paper.  Remember to turn it back on *before*
@c starting the page-breaking work.

@c 4/2002: Karl Berry recommends commenting out this and the
@c `@setchapternewpage odd', and letting users use `texi2dvi -t'
@c if they want to waste paper.
@c @smallbook


@c Uncomment this for the release.  Leaving it off saves paper
@c during editing and review.
@setchapternewpage odd

@shorttitlepage GNU Awk
@titlepage
@title @value{TITLE}
@subtitle @value{SUBTITLE}
@subtitle Edition @value{EDITION}
@subtitle @value{UPDATE-MONTH}
@author Arnold D. Robbins

@ifnotdocbook
@c Include the Distribution inside the titlepage environment so
@c that headings are turned off.  Headings on and off do not work.

@page
@vskip 0pt plus 1filll
``To boldly go where no man has gone before'' is a
Registered Trademark of Paramount Pictures Corporation. @*
@c sorry, i couldn't resist
@sp 3
Published by:
@sp 1

Free Software Foundation @*
51 Franklin Street, Fifth Floor @*
Boston, MA  02110-1301 USA @*
Phone: +1-617-542-5942 @*
Fax: +1-617-542-2652 @*
Email: @email{gnu@@gnu.org} @*
URL: @uref{https://www.gnu.org/} @*

@c This one is correct for gawk 3.1.0 from the FSF
ISBN 1-882114-28-0 @*
@sp 2
@insertcopying
@end ifnotdocbook
@end titlepage

@c Thanks to Bob Chassell for directions on doing dedications.
@iftex
@headings off
@page
@w{ }
@sp 9
@center @i{To my parents, for their love, and for the wonderful example they set for me.}
@sp 1
@center @i{To my wife, Miriam, for making me complete.
Thank you for building your life together with me.}
@sp 1
@center @i{To our children, Chana, Rivka, Nachum, and Malka, for enrichening our lives in innumerable ways.}
@sp 1
@w{ }
@page
@w{ }
@page
@headings on
@end iftex

@docbook
<dedication>
<para>To my parents, for their love, and for the wonderful
example they set for me.</para>
<para>To my wife Miriam, for making me complete.
Thank you for building your life together with me.</para>
<para>To our children Chana, Rivka, Nachum and Malka,
for enrichening our lives in innumerable ways.</para>
</dedication>
@end docbook

@iftex
@headings off
@evenheading @thispage@ @ @ @strong{@value{TITLE}} @| @|
@oddheading  @| @| @strong{@thischapter}@ @ @ @thispage
@end iftex

@ifnottex
@ifnotxml
@ifnotdocbook
@node Top
@top General Introduction
@c Preface node should come right after the Top
@c node, in `unnumbered' sections, then the chapter, `What is gawk'.
@c Licensing nodes are appendices, they're not central to AWK.

This file documents @command{awk}, a program that you can use to select
particular records in a file and perform operations upon them.

@insertcopying

@end ifnotdocbook
@end ifnotxml
@end ifnottex

@menu
* Foreword3::                      Some nice words about this
                                   @value{DOCUMENT}.
* Foreword4::                      More nice words.
* Preface::                        What this @value{DOCUMENT} is about; brief
                                   history and acknowledgments.
* Getting Started::                A basic introduction to using
                                   @command{awk}. How to run an @command{awk}
                                   program. Command-line syntax.
* Invoking Gawk::                  How to run @command{gawk}.
* Regexp::                         All about matching things using regular
                                   expressions.
* Reading Files::                  How to read files and manipulate fields.
* Printing::                       How to print using @command{awk}. Describes
                                   the @code{print} and @code{printf}
                                   statements. Also describes redirection of
                                   output.
* Expressions::                    Expressions are the basic building blocks
                                   of statements.
* Patterns and Actions::           Overviews of patterns and actions.
* Arrays::                         The description and use of arrays. Also
                                   includes array-oriented control statements.
* Functions::                      Built-in and user-defined functions.
* Library Functions::              A Library of @command{awk} Functions.
* Sample Programs::                Many @command{awk} programs with complete
                                   explanations.
* Advanced Features::              Stuff for advanced users, specific to
                                   @command{gawk}.
* Internationalization::           Getting @command{gawk} to speak your
                                   language.
* Debugger::                       The @command{gawk} debugger.
* Arbitrary Precision Arithmetic:: Arbitrary precision arithmetic with
                                   @command{gawk}.
* Dynamic Extensions::             Adding new built-in functions to
                                   @command{gawk}.
* Language History::               The evolution of the @command{awk}
                                   language.
* Installation::                   Installing @command{gawk} under various
                                   operating systems.
* Notes::                          Notes about adding things to @command{gawk}
                                   and possible future work.
* Basic Concepts::                 A very quick introduction to programming
                                   concepts.
* Glossary::                       An explanation of some unfamiliar terms.
* Copying::                        Your right to copy and distribute
                                   @command{gawk}.
* GNU Free Documentation License:: The license for this @value{DOCUMENT}.
* Index::                          Concept and Variable Index.

@detailmenu
* History::                             The history of @command{gawk} and
                                        @command{awk}.
* Names::                               What name to use to find
                                        @command{awk}.
* This Manual::                         Using this @value{DOCUMENT}. Includes
                                        sample input files that you can use.
* Conventions::                         Typographical Conventions.
* Manual History::                      Brief history of the GNU project and
                                        this @value{DOCUMENT}.
* How To Contribute::                   Helping to save the world.
* Acknowledgments::                     Acknowledgments.
* Running gawk::                        How to run @command{gawk} programs;
                                        includes command-line syntax.
* One-shot::                            Running a short throwaway
                                        @command{awk} program.
* Read Terminal::                       Using no input files (input from the
                                        keyboard instead).
* Long::                                Putting permanent @command{awk}
                                        programs in files.
* Executable Scripts::                  Making self-contained @command{awk}
                                        programs.
* Comments::                            Adding documentation to @command{gawk}
                                        programs.
* Quoting::                             More discussion of shell quoting
                                        issues.
* DOS Quoting::                         Quoting in Windows Batch Files.
* Sample Data Files::                   Sample data files for use in the
                                        @command{awk} programs illustrated in
                                        this @value{DOCUMENT}.
* Very Simple::                         A very simple example.
* Two Rules::                           A less simple one-line example using
                                        two rules.
* More Complex::                        A more complex example.
* Statements/Lines::                    Subdividing or combining statements
                                        into lines.
* Other Features::                      Other Features of @command{awk}.
* When::                                When to use @command{gawk} and when to
                                        use other things.
* Intro Summary::                       Summary of the introduction.
* Command Line::                        How to run @command{awk}.
* Options::                             Command-line options and their
                                        meanings.
* Other Arguments::                     Input file names and variable
                                        assignments.
* Naming Standard Input::               How to specify standard input with
                                        other files.
* Environment Variables::               The environment variables
                                        @command{gawk} uses.
* AWKPATH Variable::                    Searching directories for
                                        @command{awk} programs.
* AWKLIBPATH Variable::                 Searching directories for
                                        @command{awk} shared libraries.
* Other Environment Variables::         The environment variables.
* Exit Status::                         @command{gawk}'s exit status.
* Include Files::                       Including other files into your
                                        program.
* Loading Shared Libraries::            Loading shared libraries into your
                                        program.
* Obsolete::                            Obsolete Options and/or features.
* Undocumented::                        Undocumented Options and Features.
* Invoking Summary::                    Invocation summary.
* Regexp Usage::                        How to Use Regular Expressions.
* Escape Sequences::                    How to write nonprinting characters.
* Regexp Operators::                    Regular Expression Operators.
* Bracket Expressions::                 What can go between @samp{[...]}.
* Leftmost Longest::                    How much text matches.
* Computed Regexps::                    Using Dynamic Regexps.
* GNU Regexp Operators::                Operators specific to GNU software.
* Case-sensitivity::                    How to do case-insensitive matching.
* Regexp Summary::                      Regular expressions summary.
* Records::                             Controlling how data is split into
                                        records.
* awk split records::                   How standard @command{awk} splits
                                        records.
* gawk split records::                  How @command{gawk} splits records.
* Fields::                              An introduction to fields.
* Nonconstant Fields::                  Nonconstant Field Numbers.
* Changing Fields::                     Changing the Contents of a Field.
* Field Separators::                    The field separator and how to change
                                        it.
* Default Field Splitting::             How fields are normally separated.
* Regexp Field Splitting::              Using regexps as the field separator.
* Single Character Fields::             Making each character a separate
                                        field.
* Command Line Field Separator::        Setting @code{FS} from the command
                                        line.
* Full Line Fields::                    Making the full line be a single
                                        field.
* Field Splitting Summary::             Some final points and a summary table.
* Constant Size::                       Reading constant width data.
* Fixed width data::                    Processing fixed-width data.
* Skipping intervening::                Skipping intervening fields.
* Allowing trailing data::              Capturing optional trailing data.
* Fields with fixed data::              Field values with fixed-width data.
* Splitting By Content::                Defining Fields By Content
* Testing field creation::              Checking how @command{gawk} is
                                        splitting records.
* Multiple Line::                       Reading multiline records.
* Getline::                             Reading files under explicit program
                                        control using the @code{getline}
                                        function.
* Plain Getline::                       Using @code{getline} with no
                                        arguments.
* Getline/Variable::                    Using @code{getline} into a variable.
* Getline/File::                        Using @code{getline} from a file.
* Getline/Variable/File::               Using @code{getline} into a variable
                                        from a file.
* Getline/Pipe::                        Using @code{getline} from a pipe.
* Getline/Variable/Pipe::               Using @code{getline} into a variable
                                        from a pipe.
* Getline/Coprocess::                   Using @code{getline} from a coprocess.
* Getline/Variable/Coprocess::          Using @code{getline} into a variable
                                        from a coprocess.
* Getline Notes::                       Important things to know about
                                        @code{getline}.
* Getline Summary::                     Summary of @code{getline} Variants.
* Read Timeout::                        Reading input with a timeout.
* Retrying Input::                      Retrying input after certain errors.
* Command-line directories::            What happens if you put a directory on
                                        the command line.
* Input Summary::                       Input summary.
* Input Exercises::                     Exercises.
* Print::                               The @code{print} statement.
* Print Examples::                      Simple examples of @code{print}
                                        statements.
* Output Separators::                   The output separators and how to
                                        change them.
* OFMT::                                Controlling Numeric Output With
                                        @code{print}.
* Printf::                              The @code{printf} statement.
* Basic Printf::                        Syntax of the @code{printf} statement.
* Control Letters::                     Format-control letters.
* Format Modifiers::                    Format-specification modifiers.
* Printf Examples::                     Several examples.
* Redirection::                         How to redirect output to multiple
                                        files and pipes.
* Special FD::                          Special files for I/O.
* Special Files::                       File name interpretation in
                                        @command{gawk}. @command{gawk} allows
                                        access to inherited file descriptors.
* Other Inherited Files::               Accessing other open files with
                                        @command{gawk}.
* Special Network::                     Special files for network
                                        communications.
* Special Caveats::                     Things to watch out for.
* Close Files And Pipes::               Closing Input and Output Files and
                                        Pipes.
* Nonfatal::                            Enabling Nonfatal Output.
* Output Summary::                      Output summary.
* Output Exercises::                    Exercises.
* Values::                              Constants, Variables, and Regular
                                        Expressions.
* Constants::                           String, numeric and regexp constants.
* Scalar Constants::                    Numeric and string constants.
* Nondecimal-numbers::                  What are octal and hex numbers.
* Regexp Constants::                    Regular Expression constants.
* Using Constant Regexps::              When and how to use a regexp constant.
* Standard Regexp Constants::           Regexp constants in standard
                                        @command{awk}.
* Strong Regexp Constants::             Strongly typed regexp constants.
* Variables::                           Variables give names to values for
                                        later use.
* Using Variables::                     Using variables in your programs.
* Assignment Options::                  Setting variables on the command line
                                        and a summary of command-line syntax.
                                        This is an advanced method of input.
* Conversion::                          The conversion of strings to numbers
                                        and vice versa.
* Strings And Numbers::                 How @command{awk} Converts Between
                                        Strings And Numbers.
* Locale influences conversions::       How the locale may affect conversions.
* All Operators::                       @command{gawk}'s operators.
* Arithmetic Ops::                      Arithmetic operations (@samp{+},
                                        @samp{-}, etc.)
* Concatenation::                       Concatenating strings.
* Assignment Ops::                      Changing the value of a variable or a
                                        field.
* Increment Ops::                       Incrementing the numeric value of a
                                        variable.
* Truth Values and Conditions::         Testing for true and false.
* Truth Values::                        What is ``true'' and what is
                                        ``false''.
* Typing and Comparison::               How variables acquire types and how
                                        this affects comparison of numbers and
                                        strings with @samp{<}, etc.
* Variable Typing::                     String type versus numeric type.
* Comparison Operators::                The comparison operators.
* POSIX String Comparison::             String comparison with POSIX rules.
* Boolean Ops::                         Combining comparison expressions using
                                        boolean operators @samp{||} (``or''),
                                        @samp{&&} (``and'') and @samp{!}
                                        (``not'').
* Conditional Exp::                     Conditional expressions select between
                                        two subexpressions under control of a
                                        third subexpression.
* Function Calls::                      A function call is an expression.
* Precedence::                          How various operators nest.
* Locales::                             How the locale affects things.
* Expressions Summary::                 Expressions summary.
* Pattern Overview::                    What goes into a pattern.
* Regexp Patterns::                     Using regexps as patterns.
* Expression Patterns::                 Any expression can be used as a
                                        pattern.
* Ranges::                              Pairs of patterns specify record
                                        ranges.
* BEGIN/END::                           Specifying initialization and cleanup
                                        rules.
* Using BEGIN/END::                     How and why to use BEGIN/END rules.
* I/O And BEGIN/END::                   I/O issues in BEGIN/END rules.
* BEGINFILE/ENDFILE::                   Two special patterns for advanced
                                        control.
* Empty::                               The empty pattern, which matches every
                                        record.
* Using Shell Variables::               How to use shell variables with
                                        @command{awk}.
* Action Overview::                     What goes into an action.
* Statements::                          Describes the various control
                                        statements in detail.
* If Statement::                        Conditionally execute some
                                        @command{awk} statements.
* While Statement::                     Loop until some condition is
                                        satisfied.
* Do Statement::                        Do specified action while looping
                                        until some condition is satisfied.
* For Statement::                       Another looping statement, that
                                        provides initialization and increment
                                        clauses.
* Switch Statement::                    Switch/case evaluation for conditional
                                        execution of statements based on a
                                        value.
* Break Statement::                     Immediately exit the innermost
                                        enclosing loop.
* Continue Statement::                  Skip to the end of the innermost
                                        enclosing loop.
* Next Statement::                      Stop processing the current input
                                        record.
* Nextfile Statement::                  Stop processing the current file.
* Exit Statement::                      Stop execution of @command{awk}.
* Built-in Variables::                  Summarizes the predefined variables.
* User-modified::                       Built-in variables that you change to
                                        control @command{awk}.
* Auto-set::                            Built-in variables where @command{awk}
                                        gives you information.
* ARGC and ARGV::                       Ways to use @code{ARGC} and
                                        @code{ARGV}.
* Pattern Action Summary::              Patterns and Actions summary.
* Array Basics::                        The basics of arrays.
* Array Intro::                         Introduction to Arrays
* Reference to Elements::               How to examine one element of an
                                        array.
* Assigning Elements::                  How to change an element of an array.
* Array Example::                       Basic Example of an Array
* Scanning an Array::                   A variation of the @code{for}
                                        statement. It loops through the
                                        indices of an array's existing
                                        elements.
* Controlling Scanning::                Controlling the order in which arrays
                                        are scanned.
* Numeric Array Subscripts::            How to use numbers as subscripts in
                                        @command{awk}.
* Uninitialized Subscripts::            Using Uninitialized variables as
                                        subscripts.
* Delete::                              The @code{delete} statement removes an
                                        element from an array.
* Multidimensional::                    Emulating multidimensional arrays in
                                        @command{awk}.
* Multiscanning::                       Scanning multidimensional arrays.
* Arrays of Arrays::                    True multidimensional arrays.
* Arrays Summary::                      Summary of arrays.
* Built-in::                            Summarizes the built-in functions.
* Calling Built-in::                    How to call built-in functions.
* Numeric Functions::                   Functions that work with numbers,
                                        including @code{int()}, @code{sin()}
                                        and @code{rand()}.
* String Functions::                    Functions for string manipulation,
                                        such as @code{split()}, @code{match()}
                                        and @code{sprintf()}.
* Gory Details::                        More than you want to know about
                                        @samp{\} and @samp{&} with
                                        @code{sub()}, @code{gsub()}, and
                                        @code{gensub()}.
* I/O Functions::                       Functions for files and shell
                                        commands.
* Time Functions::                      Functions for dealing with timestamps.
* Bitwise Functions::                   Functions for bitwise operations.
* Type Functions::                      Functions for type information.
* I18N Functions::                      Functions for string translation.
* User-defined::                        Describes User-defined functions in
                                        detail.
* Definition Syntax::                   How to write definitions and what they
                                        mean.
* Function Example::                    An example function definition and
                                        what it does.
* Function Caveats::                    Things to watch out for.
* Calling A Function::                  Don't use spaces.
* Variable Scope::                      Controlling variable scope.
* Pass By Value/Reference::             Passing parameters.
* Return Statement::                    Specifying the value a function
                                        returns.
* Dynamic Typing::                      How variable types can change at
                                        runtime.
* Indirect Calls::                      Choosing the function to call at
                                        runtime.
* Functions Summary::                   Summary of functions.
* Library Names::                       How to best name private global
                                        variables in library functions.
* General Functions::                   Functions that are of general use.
* Strtonum Function::                   A replacement for the built-in
                                        @code{strtonum()} function.
* Assert Function::                     A function for assertions in
                                        @command{awk} programs.
* Round Function::                      A function for rounding if
                                        @code{sprintf()} does not do it
                                        correctly.
* Cliff Random Function::               The Cliff Random Number Generator.
* Ordinal Functions::                   Functions for using characters as
                                        numbers and vice versa.
* Join Function::                       A function to join an array into a
                                        string.
* Getlocaltime Function::               A function to get formatted times.
* Readfile Function::                   A function to read an entire file at
                                        once.
* Shell Quoting::                       A function to quote strings for the
                                        shell.
* Data File Management::                Functions for managing command-line
                                        data files.
* Filetrans Function::                  A function for handling data file
                                        transitions.
* Rewind Function::                     A function for rereading the current
                                        file.
* File Checking::                       Checking that data files are readable.
* Empty Files::                         Checking for zero-length files.
* Ignoring Assigns::                    Treating assignments as file names.
* Getopt Function::                     A function for processing command-line
                                        arguments.
* Passwd Functions::                    Functions for getting user
                                        information.
* Group Functions::                     Functions for getting group
                                        information.
* Walking Arrays::                      A function to walk arrays of arrays.
* Library Functions Summary::           Summary of library functions.
* Library Exercises::                   Exercises.
* Running Examples::                    How to run these examples.
* Clones::                              Clones of common utilities.
* Cut Program::                         The @command{cut} utility.
* Egrep Program::                       The @command{egrep} utility.
* Id Program::                          The @command{id} utility.
* Split Program::                       The @command{split} utility.
* Tee Program::                         The @command{tee} utility.
* Uniq Program::                        The @command{uniq} utility.
* Wc Program::                          The @command{wc} utility.
* Miscellaneous Programs::              Some interesting @command{awk}
                                        programs.
* Dupword Program::                     Finding duplicated words in a
                                        document.
* Alarm Program::                       An alarm clock.
* Translate Program::                   A program similar to the @command{tr}
                                        utility.
* Labels Program::                      Printing mailing labels.
* Word Sorting::                        A program to produce a word usage
                                        count.
* History Sorting::                     Eliminating duplicate entries from a
                                        history file.
* Extract Program::                     Pulling out programs from Texinfo
                                        source files.
* Simple Sed::                          A Simple Stream Editor.
* Igawk Program::                       A wrapper for @command{awk} that
                                        includes files.
* Anagram Program::                     Finding anagrams from a dictionary.
* Signature Program::                   People do amazing things with too much
                                        time on their hands.
* Programs Summary::                    Summary of programs.
* Programs Exercises::                  Exercises.
* Nondecimal Data::                     Allowing nondecimal input data.
* Array Sorting::                       Facilities for controlling array
                                        traversal and sorting arrays.
* Controlling Array Traversal::         How to use PROCINFO["sorted_in"].
* Array Sorting Functions::             How to use @code{asort()} and
                                        @code{asorti()}.
* Two-way I/O::                         Two-way communications with another
                                        process.
* TCP/IP Networking::                   Using @command{gawk} for network
                                        programming.
* Profiling::                           Profiling your @command{awk} programs.
* Advanced Features Summary::           Summary of advanced features.
* I18N and L10N::                       Internationalization and Localization.
* Explaining gettext::                  How GNU @command{gettext} works.
* Programmer i18n::                     Features for the programmer.
* Translator i18n::                     Features for the translator.
* String Extraction::                   Extracting marked strings.
* Printf Ordering::                     Rearranging @code{printf} arguments.
* I18N Portability::                    @command{awk}-level portability
                                        issues.
* I18N Example::                        A simple i18n example.
* Gawk I18N::                           @command{gawk} is also
                                        internationalized.
* I18N Summary::                        Summary of I18N stuff.
* Debugging::                           Introduction to @command{gawk}
                                        debugger.
* Debugging Concepts::                  Debugging in General.
* Debugging Terms::                     Additional Debugging Concepts.
* Awk Debugging::                       Awk Debugging.
* Sample Debugging Session::            Sample debugging session.
* Debugger Invocation::                 How to Start the Debugger.
* Finding The Bug::                     Finding the Bug.
* List of Debugger Commands::           Main debugger commands.
* Breakpoint Control::                  Control of Breakpoints.
* Debugger Execution Control::          Control of Execution.
* Viewing And Changing Data::           Viewing and Changing Data.
* Execution Stack::                     Dealing with the Stack.
* Debugger Info::                       Obtaining Information about the
                                        Program and the Debugger State.
* Miscellaneous Debugger Commands::     Miscellaneous Commands.
* Readline Support::                    Readline support.
* Limitations::                         Limitations and future plans.
* Debugging Summary::                   Debugging summary.
* Computer Arithmetic::                 A quick intro to computer math.
* Math Definitions::                    Defining terms used.
* MPFR features::                       The MPFR features in @command{gawk}.
* FP Math Caution::                     Things to know.
* Inexactness of computations::         Floating point math is not exact.
* Inexact representation::              Numbers are not exactly represented.
* Comparing FP Values::                 How to compare floating point values.
* Errors accumulate::                   Errors get bigger as they go.
* Getting Accuracy::                    Getting more accuracy takes some work.
* Try To Round::                        Add digits and round.
* Setting precision::                   How to set the precision.
* Setting the rounding mode::           How to set the rounding mode.
* Arbitrary Precision Integers::        Arbitrary Precision Integer Arithmetic
                                        with @command{gawk}.
* Checking for MPFR::                   How to check if MPFR is available.
* POSIX Floating Point Problems::       Standards Versus Existing Practice.
* Floating point summary::              Summary of floating point discussion.
* Extension Intro::                     What is an extension.
* Plugin License::                      A note about licensing.
* Extension Mechanism Outline::         An outline of how it works.
* Extension API Description::           A full description of the API.
* Extension API Functions Introduction:: Introduction to the API functions.
* General Data Types::                  The data types.
* Memory Allocation Functions::         Functions for allocating memory.
* Constructor Functions::               Functions for creating values.
* Registration Functions::              Functions to register things with
                                        @command{gawk}.
* Extension Functions::                 Registering extension functions.
* Exit Callback Functions::             Registering an exit callback.
* Extension Version String::            Registering a version string.
* Input Parsers::                       Registering an input parser.
* Output Wrappers::                     Registering an output wrapper.
* Two-way processors::                  Registering a two-way processor.
* Printing Messages::                   Functions for printing messages.
* Updating @code{ERRNO}::               Functions for updating @code{ERRNO}.
* Requesting Values::                   How to get a value.
* Accessing Parameters::                Functions for accessing parameters.
* Symbol Table Access::                 Functions for accessing global
                                        variables.
* Symbol table by name::                Accessing variables by name.
* Symbol table by cookie::              Accessing variables by ``cookie''.
* Cached values::                       Creating and using cached values.
* Array Manipulation::                  Functions for working with arrays.
* Array Data Types::                    Data types for working with arrays.
* Array Functions::                     Functions for working with arrays.
* Flattening Arrays::                   How to flatten arrays.
* Creating Arrays::                     How to create and populate arrays.
* Redirection API::                     How to access and manipulate
                                        redirections.
* Extension API Variables::             Variables provided by the API.
* Extension Versioning::                API Version information.
* Extension GMP/MPFR Versioning::       Version information about GMP and
                                        MPFR.
* Extension API Informational Variables:: Variables providing information about
                                        @command{gawk}'s invocation.
* Extension API Boilerplate::           Boilerplate code for using the API.
* Changes from API V1::                 Changes from V1 of the API.
* Finding Extensions::                  How @command{gawk} finds compiled
                                        extensions.
* Extension Example::                   Example C code for an extension.
* Internal File Description::           What the new functions will do.
* Internal File Ops::                   The code for internal file operations.
* Using Internal File Ops::             How to use an external extension.
* Extension Samples::                   The sample extensions that ship with
                                        @command{gawk}.
* Extension Sample File Functions::     The file functions sample.
* Extension Sample Fnmatch::            An interface to @code{fnmatch()}.
* Extension Sample Fork::               An interface to @code{fork()} and
                                        other process functions.
* Extension Sample Inplace::            Enabling in-place file editing.
* Extension Sample Ord::                Character to value to character
                                        conversions.
* Extension Sample Readdir::            An interface to @code{readdir()}.
* Extension Sample Revout::             Reversing output sample output
                                        wrapper.
* Extension Sample Rev2way::            Reversing data sample two-way
                                        processor.
* Extension Sample Read write array::   Serializing an array to a file.
* Extension Sample Readfile::           Reading an entire file into a string.
* Extension Sample Time::               An interface to @code{gettimeofday()}
                                        and @code{sleep()}.
* Extension Sample API Tests::          Tests for the API.
* gawkextlib::                          The @code{gawkextlib} project.
* Extension summary::                   Extension summary.
* Extension Exercises::                 Exercises.
* V7/SVR3.1::                           The major changes between V7 and
                                        System V Release 3.1.
* SVR4::                                Minor changes between System V
                                        Releases 3.1 and 4.
* POSIX::                               New features from the POSIX standard.
* BTL::                                 New features from Brian Kernighan's
                                        version of @command{awk}.
* POSIX/GNU::                           The extensions in @command{gawk} not
                                        in POSIX @command{awk}.
* Feature History::                     The history of the features in
                                        @command{gawk}.
* Common Extensions::                   Common Extensions Summary.
* Ranges and Locales::                  How locales used to affect regexp
                                        ranges.
* Contributors::                        The major contributors to
                                        @command{gawk}.
* History summary::                     History summary.
* Gawk Distribution::                   What is in the @command{gawk}
                                        distribution.
* Getting::                             How to get the distribution.
* Extracting::                          How to extract the distribution.
* Distribution contents::               What is in the distribution.
* Unix Installation::                   Installing @command{gawk} under
                                        various versions of Unix.
* Quick Installation::                  Compiling @command{gawk} under Unix.
* Shell Startup Files::                 Shell convenience functions.
* Additional Configuration Options::    Other compile-time options.
* Configuration Philosophy::            How it's all supposed to work.
* Non-Unix Installation::               Installation on Other Operating
                                        Systems.
* PC Installation::                     Installing and Compiling
                                        @command{gawk} on Microsoft Windows.
* PC Binary Installation::              Installing a prepared distribution.
* PC Compiling::                        Compiling @command{gawk} for
                                        Windows32.
* PC Using::                            Running @command{gawk} on Windows32.
* Cygwin::                              Building and running @command{gawk}
                                        for Cygwin.
* MSYS::                                Using @command{gawk} In The MSYS
                                        Environment.
* VMS Installation::                    Installing @command{gawk} on VMS.
* VMS Compilation::                     How to compile @command{gawk} under
                                        VMS.
* VMS Dynamic Extensions::              Compiling @command{gawk} dynamic
                                        extensions on VMS.
* VMS Installation Details::            How to install @command{gawk} under
                                        VMS.
* VMS Running::                         How to run @command{gawk} under VMS.
* VMS GNV::                             The VMS GNV Project.
* VMS Old Gawk::                        An old version comes with some VMS
                                        systems.
* Bugs::                                Reporting Problems and Bugs.
* Bug address::                         Where to send reports to.
* Usenet::                              Where not to send reports to.
* Maintainers::                         Maintainers of non-*nix ports.
* Other Versions::                      Other freely available @command{awk}
                                        implementations.
* Installation summary::                Summary of installation.
* Compatibility Mode::                  How to disable certain @command{gawk}
                                        extensions.
* Additions::                           Making Additions To @command{gawk}.
* Accessing The Source::                Accessing the Git repository.
* Adding Code::                         Adding code to the main body of
                                        @command{gawk}.
* New Ports::                           Porting @command{gawk} to a new
                                        operating system.
* Derived Files::                       Why derived files are kept in the Git
                                        repository.
* Future Extensions::                   New features that may be implemented
                                        one day.
* Implementation Limitations::          Some limitations of the
                                        implementation.
* Extension Design::                    Design notes about the extension API.
* Old Extension Problems::              Problems with the old mechanism.
* Extension New Mechanism Goals::       Goals for the new mechanism.
* Extension Other Design Decisions::    Some other design decisions.
* Extension Future Growth::             Some room for future growth.
* Old Extension Mechanism::             Some compatibility for old extensions.
* Notes summary::                       Summary of implementation notes.
* Basic High Level::                    The high level view.
* Basic Data Typing::                   A very quick intro to data types.
@end detailmenu
@end menu

@c dedication for Info file
@ifinfo
To my parents, for their love, and for the wonderful
example they set for me.
@sp 1
To my wife Miriam, for making me complete.
Thank you for building your life together with me.
@sp 1
To our children Chana, Rivka, Nachum and Malka,
for enrichening our lives in innumerable ways.
@end ifinfo

@summarycontents
@contents

@node Foreword3
@unnumbered Foreword to the Third Edition

@c This bit is post-processed by a script which turns the chapter
@c tag into a preface tag, and moves this stuff to before the title.
@c Bleah.
@docbook
  <prefaceinfo>
    <author>
      <firstname>Michael</firstname>
      <surname>Brennan</surname>
      <!-- can't put mawk into command tags. sigh. -->
      <affiliation><jobtitle>Author of mawk</jobtitle></affiliation>
    </author>
    <date>March 2001</date>
   </prefaceinfo>
@end docbook

Arnold Robbins and I are good friends. We were introduced
@c 11 years ago
in 1990
by circumstances---and our favorite programming language, AWK.
The circumstances started a couple of years
earlier. I was working at a new job and noticed an unplugged
Unix computer sitting in the corner.  No one knew how to use it,
and neither did I.  However,
a couple of days later, it was running, and
I was @code{root} and the one-and-only user.
That day, I began the transition from statistician to Unix programmer.

On one of many trips to the library or bookstore in search of
books on Unix, I found the gray AWK book, a.k.a.@:
Alfred V.@: Aho, Brian W.@: Kernighan, and
Peter J.@: Weinberger's @cite{The AWK Programming Language} (Addison-Wesley,
1988).  @command{awk}'s simple programming paradigm---find a pattern in the
input and then perform an action---often reduced complex or tedious
data manipulations to a few lines of code.  I was excited to try my
hand at programming in AWK.

Alas,  the @command{awk} on my computer was a limited version of the
language described in the gray book.  I discovered that my computer
had ``old @command{awk}'' and the book described
``new @command{awk}.''
I learned that this was typical; the old version refused to step
aside or relinquish its name.  If a system had a new @command{awk}, it was
invariably called @command{nawk}, and few systems had it.
The best way to get a new @command{awk} was to @command{ftp} the source code for
@command{gawk} from @code{prep.ai.mit.edu}.  @command{gawk} was a version of
new @command{awk} written by David Trueman and Arnold, and available under
the GNU General Public License.

(Incidentally,
it's no longer difficult to find a new @command{awk}. @command{gawk} ships with
GNU/Linux, and you can download binaries or source code for almost
any system; my wife uses @command{gawk} on her VMS box.)

My Unix system started out unplugged from the wall; it certainly was not
plugged into a network.  So, oblivious to the existence of @command{gawk}
and the Unix community in general, and desiring a new @command{awk}, I wrote
my own, called @command{mawk}.
Before I was finished, I knew about @command{gawk},
but it was too late to stop, so I eventually posted
to a @code{comp.sources} newsgroup.

A few days after my posting, I got a friendly email
from Arnold introducing
himself.   He suggested we share design and algorithms and
attached a draft of the POSIX standard so
that I could update @command{mawk} to support language extensions added
after publication of @cite{The AWK Programming Language}.

Frankly, if our roles had
been reversed, I would not have been so open and we probably would
have never met.  I'm glad we did meet.
He is an AWK expert's AWK expert and a genuinely nice person.
Arnold contributes significant amounts of his
expertise and time to the Free Software Foundation.

This book is the @command{gawk} reference manual, but at its core it
is a book about AWK programming that
will appeal to a wide audience.
It is a definitive reference to the AWK language as defined by the
1987 Bell Laboratories release and codified in the 1992 POSIX Utilities
standard.

On the other hand, the novice AWK programmer can study
a wealth of practical programs that emphasize
the power of AWK's basic idioms:
data-driven control flow, pattern matching with regular expressions,
and associative arrays.
Those looking for something new can try out @command{gawk}'s
interface to network protocols via special @file{/inet} files.

The programs in this book make clear that an AWK program is
typically much smaller and faster to develop than
a counterpart written in C.
Consequently, there is often a payoff to prototyping an
algorithm or design in AWK to get it running quickly and expose
problems early. Often, the interpreted performance is adequate
and the AWK prototype becomes the product.

The new @command{pgawk} (profiling @command{gawk}), produces
program execution counts.
I recently experimented with an algorithm that for
@ifnotdocbook
@math{n}
@end ifnotdocbook
@ifdocbook
@i{n}
@end ifdocbook
lines of input, exhibited
@tex
$\sim\! Cn^2$
@end tex
@ifnottex
@ifnotdocbook
~ C n^2
@end ifnotdocbook
@end ifnottex
@docbook
<emphasis>&sim; Cn<superscript>2</superscript></emphasis>
@end docbook
performance, while
theory predicted
@tex
$\sim\! Cn\log n$
@end tex
@ifnottex
@ifnotdocbook
~ C n log n
@end ifnotdocbook
@end ifnottex
@docbook
<emphasis>&sim; Cn log n</emphasis>
@end docbook
behavior. A few minutes poring
over the @file{awkprof.out} profile pinpointed the problem to
a single line of code.  @command{pgawk} is a welcome addition to
my programmer's toolbox.

Arnold has distilled over a decade of experience writing and
using AWK programs, and developing @command{gawk}, into this book.  If you use
AWK or want to learn how, then read this book.

@ifnotdocbook
@cindex Brennan, Michael
@display
Michael Brennan
Author of @command{mawk}
March 2001
@end display
@end ifnotdocbook

@node Foreword4
@unnumbered Foreword to the Fourth Edition

@c This bit is post-processed by a script which turns the chapter
@c tag into a preface tag, and moves this stuff to before the title.
@c Bleah.
@docbook
  <prefaceinfo>
    <author>
      <firstname>Michael</firstname>
      <surname>Brennan</surname>
      <!-- can't put mawk into command tags. sigh. -->
      <affiliation><jobtitle>Author of mawk</jobtitle></affiliation>
    </author>
    <date>October 2014</date>
   </prefaceinfo>
@end docbook

Some things don't change.  Thirteen years ago I wrote:
``If you use AWK or want to learn how, then read this book.''
True then, and still true today.

Learning to use a programming language is about more than mastering the
syntax.  One needs to acquire an understanding of how to use the
features of the language to solve practical programming problems.
A focus of this book is many examples that show how to use AWK.

Some things do change. Our computers are much faster and have more memory.
Consequently, speed and storage inefficiencies of a high-level language
matter less.  Prototyping in AWK and then rewriting in C for performance
reasons happens less, because more often the prototype is fast enough.

Of course, there are computing operations that are best done in C or C++.
With @command{gawk} 4.1 and later, you do not have to choose between writing
your program in AWK or in C/C++.  You can write most of your
program in AWK and the aspects that require C/C++ capabilities can be written
in C/C++, and then the pieces glued together when the @command{gawk} module loads
the C/C++ module as a dynamic plug-in.
@c Chapter 16
@ref{Dynamic Extensions},
has all the
details, and, as expected, many examples to help you learn the ins and outs.

I enjoy programming in AWK and had fun (re)reading this book.
I think you will too.

@ifnotdocbook
@cindex Brennan, Michael
@display
Michael Brennan
Author of @command{mawk}
October 2014
@end display
@end ifnotdocbook

@node Preface
@unnumbered Preface
@c I saw a comment somewhere that the preface should describe the book itself,
@c and the introduction should describe what the book covers.
@c
@c 12/2000: Chuck wants the preface & intro combined.

@c This bit is post-processed by a script which turns the chapter
@c tag into a preface tag, and moves this stuff to before the title.
@c Bleah.
@docbook
  <prefaceinfo>
    <author>
      <firstname>Arnold</firstname>
      <surname>Robbins</surname>
      <affiliation><jobtitle>Nof Ayalon</jobtitle></affiliation>
      <affiliation><jobtitle>Israel</jobtitle></affiliation>
    </author>
    <date>February 2015</date>
   </prefaceinfo>
@end docbook

Several kinds of tasks occur repeatedly when working with text files.
You might want to extract certain lines and discard the rest.  Or you
may need to make changes wherever certain patterns appear, but leave the
rest of the file alone.  Such jobs are often easy with @command{awk}.
The @command{awk} utility interprets a special-purpose programming
language that makes it easy to handle simple data-reformatting jobs.

The GNU implementation of @command{awk} is called @command{gawk}; if you
invoke it with the proper options or environment variables,
it is fully compatible with
the POSIX@footnote{The 2008 POSIX standard is accessible online at
@w{@url{http://www.opengroup.org/onlinepubs/9699919799/}.}}
specification of the @command{awk} language
and with the Unix version of @command{awk} maintained
by Brian Kernighan.
This means that all
properly written @command{awk} programs should work with @command{gawk}.
So most of the time, we don't distinguish between @command{gawk} and other
@command{awk} implementations.

@cindex @command{awk}, POSIX and, See Also POSIX @command{awk}
@cindex @command{awk}, POSIX and
@cindex POSIX, @command{awk} and
@cindex @command{gawk}, @command{awk} and
@cindex @command{awk}, @command{gawk} and
@cindex @command{awk}, uses for
Using @command{awk} you can:

@itemize @value{BULLET}
@item
Manage small, personal databases

@item
Generate reports

@item
Validate data

@item
Produce indexes and perform other document-preparation tasks

@item
Experiment with algorithms that you can adapt later to other computer
languages
@end itemize

@cindex @command{awk}, See Also @command{gawk}
@cindex @command{gawk}, See Also @command{awk}
@cindex @command{gawk}, uses for
In addition,
@command{gawk}
provides facilities that make it easy to:

@itemize @value{BULLET}
@item
Extract bits and pieces of data for processing

@item
Sort data

@item
Perform simple network communications

@item
Profile and debug @command{awk} programs

@item
Extend the language with functions written in C or C++
@end itemize

This @value{DOCUMENT} teaches you about the @command{awk} language and
how you can use it effectively.  You should already be familiar with basic
system commands, such as @command{cat} and @command{ls},@footnote{These utilities
are available on POSIX-compliant systems, as well as on traditional
Unix-based systems. If you are using some other operating system, you still need to
be familiar with the ideas of I/O redirection and pipes.} as well as basic shell
facilities, such as input/output (I/O) redirection and pipes.

@cindex GNU @command{awk}, See @command{gawk}
Implementations of the @command{awk} language are available for many
different computing environments.  This @value{DOCUMENT}, while describing
the @command{awk} language in general, also describes the particular
implementation of @command{awk} called @command{gawk} (which stands for
``GNU @command{awk}'').  @command{gawk} runs on a broad range of Unix systems,
ranging from Intel-architecture PC-based computers
up through large-scale systems.
@command{gawk} has also been ported to Mac OS X,
Microsoft Windows
(all versions),
and OpenVMS.@footnote{Some other, obsolete systems to which @command{gawk}
was once ported are no longer supported and the code for those systems
has been removed.}

@menu
* History::                     The history of @command{gawk} and
                                @command{awk}.
* Names::                       What name to use to find @command{awk}.
* This Manual::                 Using this @value{DOCUMENT}. Includes sample
                                input files that you can use.
* Conventions::                 Typographical Conventions.
* Manual History::              Brief history of the GNU project and this
                                @value{DOCUMENT}.
* How To Contribute::           Helping to save the world.
* Acknowledgments::             Acknowledgments.
@end menu

@node History
@unnumberedsec History of @command{awk} and @command{gawk}
@cindex recipe for a programming language
@cindex programming language, recipe for
@sidebar Recipe for a Programming Language

@multitable {2 parts} {1 part  @code{egrep}} {1 part  @code{snobol}}
@item @tab 1 part  @code{egrep} @tab 1 part  @code{snobol}
@item @tab 2 parts @code{ed} @tab 3 parts C
@end multitable

Blend all parts well using @code{lex} and @code{yacc}.
Document minimally and release.

After eight years, add another part @code{egrep} and two
more parts C.  Document very well and release.
@end sidebar

@cindex Aho, Alfred
@cindex Weinberger, Peter
@cindex Kernighan, Brian
@cindex @command{awk}, history of
The name @command{awk} comes from the initials of its designers: Alfred V.@:
Aho, Peter J.@: Weinberger, and Brian W.@: Kernighan.  The original version of
@command{awk} was written in 1977 at AT&T Bell Laboratories.
In 1985, a new version made the programming
language more powerful, introducing user-defined functions, multiple input
streams, and computed regular expressions.
This new version became widely available with Unix System V
Release 3.1 (1987).
The version in System V Release 4 (1989) added some new features and cleaned
up the behavior in some of the ``dark corners'' of the language.
The specification for @command{awk} in the POSIX Command Language
and Utilities standard further clarified the language.
Both the @command{gawk} designers and the original @command{awk} designers at Bell Laboratories
provided feedback for the POSIX specification.

@cindex Rubin, Paul
@cindex Fenlason, Jay
@cindex Trueman, David
Paul Rubin wrote @command{gawk} in 1986.
Jay Fenlason completed it, with advice from Richard Stallman.  John Woods
contributed parts of the code as well.  In 1988 and 1989, David Trueman, with
help from me, thoroughly reworked @command{gawk} for compatibility
with the newer @command{awk}.
Circa 1994, I became the primary maintainer.
Current development focuses on bug fixes,
performance improvements, standards compliance, and, occasionally, new features.

In May 1997, J@"urgen Kahrs felt the need for network access
from @command{awk}, and with a little help from me, set about adding
features to do this for @command{gawk}.  At that time, he also
wrote the bulk of
@cite{@value{GAWKINETTITLE}}
(a separate document, available as part of the @command{gawk} distribution).
His code finally became part of the main @command{gawk} distribution
with @command{gawk} @value{PVERSION} 3.1.

John Haque rewrote the @command{gawk} internals, in the process providing
an @command{awk}-level debugger. This version became available as
@command{gawk} @value{PVERSION} 4.0 in 2011.

@xref{Contributors}
for a full list of those who have made important contributions to @command{gawk}.

@node Names
@unnumberedsec A Rose by Any Other Name

@cindex @command{awk}, new vs.@: old
The @command{awk} language has evolved over the years. Full details are
provided in @ref{Language History}.
The language described in this @value{DOCUMENT}
is often referred to as ``new @command{awk}.''
By analogy, the original version of @command{awk} is
referred to as ``old @command{awk}.''

On most current systems, when you run the @command{awk} utility
you get some version of new @command{awk}.@footnote{Only
Solaris systems still use an old @command{awk} for the
default @command{awk} utility. A more modern @command{awk} lives in
@file{/usr/xpg6/bin} on these systems.} If your system's standard
@command{awk} is the old one, you will see something like this
if you try the test program:

@example
@group
$ @kbd{awk 1 /dev/null}
@error{} awk: syntax error near line 1
@error{} awk: bailing out near line 1
@end group
@end example

@noindent
In this case, you should find a version of new @command{awk},
or just install @command{gawk}!

Throughout this @value{DOCUMENT}, whenever we refer to a language feature
that should be available in any complete implementation of POSIX @command{awk},
we simply use the term @command{awk}.  When referring to a feature that is
specific to the GNU implementation, we use the term @command{gawk}.

@node This Manual
@unnumberedsec Using This Book
@cindex @command{awk}, terms describing

The term @command{awk} refers to a particular program as well as to the language you
use to tell this program what to do.  When we need to be careful, we call
the language ``the @command{awk} language,''
and the program ``the @command{awk} utility.''
This @value{DOCUMENT} explains
both how to write programs in the @command{awk} language and how to
run the @command{awk} utility.
The term ``@command{awk} program'' refers to a program written by you in
the @command{awk} programming language.

@cindex @command{gawk}, @command{awk} and
@cindex @command{awk}, @command{gawk} and
@cindex POSIX @command{awk}
Primarily, this @value{DOCUMENT} explains the features of @command{awk}
as defined in the POSIX standard.  It does so in the context of the
@command{gawk} implementation.  While doing so, it also
attempts to describe important differences between @command{gawk}
and other @command{awk}
@ifclear FOR_PRINT
implementations.@footnote{All such differences
appear in the index under the
entry ``differences in @command{awk} and @command{gawk}.''}
@end ifclear
@ifset FOR_PRINT
implementations.
@end ifset
Finally, it notes any @command{gawk} features that are not in
the POSIX standard for @command{awk}.

@ifnotinfo
This @value{DOCUMENT} has the difficult task of being both a tutorial and a reference.
If you are a novice, feel free to skip over details that seem too complex.
You should also ignore the many cross-references; they are for the
expert user and for the Info and
@uref{https://www.gnu.org/software/gawk/manual/, HTML}
versions of the @value{DOCUMENT}.
@end ifnotinfo

There are sidebars
scattered throughout the @value{DOCUMENT}.
They add a more complete explanation of points that are relevant, but not likely
to be of interest on first reading.
@ifclear FOR_PRINT
All appear in the index, under the heading ``sidebar.''
@end ifclear

Most of the time, the examples use complete @command{awk} programs.
Some of the more advanced @value{SECTION}s show only the part of the @command{awk}
program that illustrates the concept being described.

Although this @value{DOCUMENT} is aimed principally at people who have not been
exposed
to @command{awk}, there is a lot of information here that even the @command{awk}
expert should find useful.  In particular, the description of POSIX
@command{awk} and the example programs in
@ref{Library Functions}, and
@ifnotdocbook
in
@end ifnotdocbook
@ref{Sample Programs},
should be of interest.

This @value{DOCUMENT} is split into several parts, as follows:

@c FULLXREF ON

@itemize @value{BULLET}
@item
Part I describes the @command{awk} language and the @command{gawk} program in detail.
It starts with the basics, and continues through all of the features of @command{awk}.
It contains the following chapters:

@c nested
@itemize @value{MINUS}
@item
@ref{Getting Started},
provides the essentials you need to know to begin using @command{awk}.

@item
@ref{Invoking Gawk},
describes how to run @command{gawk}, the meaning of its
command-line options, and how it finds @command{awk}
program source files.

@item
@ref{Regexp},
introduces regular expressions in general, and in particular the flavors
supported by POSIX @command{awk} and @command{gawk}.

@item
@ref{Reading Files},
describes how @command{awk} reads your data.
It introduces the concepts of records and fields, as well
as the @code{getline} command.
I/O redirection is first described here.
Network I/O is also briefly introduced here.

@item
@ref{Printing},
describes how @command{awk} programs can produce output with
@code{print} and @code{printf}.

@item
@ref{Expressions},
describes expressions, which are the basic building blocks
for getting most things done in a program.

@item
@ref{Patterns and Actions},
describes how to write patterns for matching records, actions for
doing something when a record is matched, and the predefined variables
@command{awk} and @command{gawk} use.

@item
@ref{Arrays},
covers @command{awk}'s one-and-only data structure: the associative array.
Deleting array elements and whole arrays is described, as well as
sorting arrays in @command{gawk}.  The @value{CHAPTER} also describes how
@command{gawk} provides arrays of arrays.

@item
@ref{Functions},
describes the built-in functions @command{awk} and @command{gawk} provide,
as well as how to define your own functions.  It also discusses how
@command{gawk} lets you call functions indirectly.
@end itemize

@item
Part II shows how to use @command{awk} and @command{gawk} for problem solving.
There is lots of code here for you to read and learn from.
This part contains the following chapters:

@c nested
@itemize @value{MINUS}
@item
@ref{Library Functions}, provides a number of functions meant to
be used from main @command{awk} programs.

@item
@ref{Sample Programs},
provides many sample @command{awk} programs.
@end itemize

Reading these two chapters allows you to see @command{awk}
solving real problems.

@item
Part III focuses on features specific to @command{gawk}.
It contains the following chapters:

@c nested
@itemize @value{MINUS}
@item
@ref{Advanced Features},
describes a number of advanced features.
Of particular note
are the abilities to control the order of array traversal,
have two-way communications with another process,
perform TCP/IP networking, and
profile your @command{awk} programs.

@item
@ref{Internationalization},
describes special features for translating program
messages into different languages at runtime.

@item
@ref{Debugger}, describes the @command{gawk} debugger.

@item
@ref{Arbitrary Precision Arithmetic},
describes advanced arithmetic facilities.

@item
@ref{Dynamic Extensions}, describes how to add new variables and
functions to @command{gawk} by writing extensions in C or C++.
@end itemize

@item
@ifclear FOR_PRINT
Part IV provides the appendices, the Glossary, and two licenses that cover
the @command{gawk} source code and this @value{DOCUMENT}, respectively.
It contains the following appendices:
@end ifclear
@ifset FOR_PRINT
Part IV provides the following appendices,
including the GNU General Public License:
@end ifset

@itemize @value{MINUS}
@item
@ref{Language History},
describes how the @command{awk} language has evolved since
its first release to the present.  It also describes how @command{gawk}
has acquired features over time.

@item
@ref{Installation},
describes how to get @command{gawk}, how to compile it
on POSIX-compatible systems,
and how to compile and use it on different
non-POSIX systems.  It also describes how to report bugs
in @command{gawk} and where to get other freely
available @command{awk} implementations.

@ifset FOR_PRINT
@item
@ref{Copying},
presents the license that covers the @command{gawk} source code.
@end ifset

@ifclear FOR_PRINT
@item
@ref{Notes},
describes how to disable @command{gawk}'s extensions, as
well as how to contribute new code to @command{gawk},
and some possible future directions for @command{gawk} development.

@item
@ref{Basic Concepts},
provides some very cursory background material for those who
are completely unfamiliar with computer programming.

The @ref{Glossary}, defines most, if not all, of the significant terms used
throughout the @value{DOCUMENT}.  If you find terms that you aren't familiar with,
try looking them up here.

@item
@ref{Copying}, and
@ref{GNU Free Documentation License},
present the licenses that cover the @command{gawk} source code
and this @value{DOCUMENT}, respectively.
@end ifclear
@end itemize
@end itemize

@ifset FOR_PRINT
The version of this @value{DOCUMENT} distributed with @command{gawk}
contains additional appendices and other end material.
To save space, we have omitted them from the
printed edition. You may find them online, as follows:

@itemize @value{BULLET}
@item
@uref{https://www.gnu.org/software/gawk/manual/html_node/Notes.html,
The appendix on implementation notes}
describes how to disable @command{gawk}'s extensions, how to contribute
new code to @command{gawk}, where to find information on some possible
future directions for @command{gawk} development, and the design decisions
behind the extension API.

@item
@uref{https://www.gnu.org/software/gawk/manual/html_node/Basic-Concepts.html,
The appendix on basic concepts}
provides some very cursory background material for those who
are completely unfamiliar with computer programming.

@item
@uref{https://www.gnu.org/software/gawk/manual/html_node/Glossary.html,
The Glossary}
defines most, if not all, of the significant terms used
throughout the @value{DOCUMENT}.  If you find terms that you aren't familiar with,
try looking them up here.

@item
@uref{https://www.gnu.org/software/gawk/manual/html_node/GNU-Free-Documentation-License.html,
The GNU FDL}
is the license that covers this @value{DOCUMENT}.
@end itemize

@c ok not to use CHAPTER / SECTION here
Some of the chapters have exercise sections; these have also been
omitted from the print edition but are available online.
@end ifset

@c FULLXREF OFF

@node Conventions
@unnumberedsec Typographical Conventions

@cindex Texinfo
This @value{DOCUMENT} is written in @uref{https://www.gnu.org/software/texinfo/, Texinfo},
the GNU documentation formatting language.
A single Texinfo source file is used to produce both the printed and online
versions of the documentation.
@ifnotinfo
Because of this, the typographical conventions
are slightly different than in other books you may have read.
@end ifnotinfo
@ifinfo
This @value{SECTION} briefly documents the typographical conventions used in Texinfo.
@end ifinfo

Examples you would type at the command line are preceded by the common
shell primary and secondary prompts, @samp{$} and @samp{>}.
Input that you type is shown @kbd{like this}.
@c 8/2014: @print{} is stripped from the texi to make docbook.
@ifclear FOR_PRINT
Output from the command is preceded by the glyph ``@print{}''.
This typically represents the command's standard output.
@end ifclear
@ifset FOR_PRINT
Output from the command, usually its standard output, appears
@code{like this}.
@end ifset
Error messages and other output on the command's standard error are preceded
by the glyph ``@error{}''.  For example:

@example
$ @kbd{echo hi on stdout}
@print{} hi on stdout
$ @kbd{echo hello on stderr 1>&2}
@error{} hello on stderr
@end example

@ifnotinfo
In the text, almost anything related to programming, such as
command names,
variable and function names, and string, numeric and regexp constants
appear in @code{this font}. Code fragments
appear in the same font and quoted, @samp{like this}.
Things that are replaced by the user or programmer
appear in @var{this font}.
Options look like this: @option{-f}.
@value{FFN}s are indicated like this: @file{/path/to/ourfile}.
@ifclear FOR_PRINT
Some things are
emphasized @emph{like this}, and if a point needs to be made
strongly, it is done @strong{like this}.
@end ifclear
The first occurrence of
a new term is usually its @dfn{definition} and appears in the same
font as the previous occurrence of ``definition'' in this sentence.
@end ifnotinfo

Characters that you type at the keyboard look @kbd{like this}.  In particular,
there are special characters called ``control characters.''  These are
characters that you type by holding down both the @kbd{CONTROL} key and
another key, at the same time.  For example, a @kbd{Ctrl-d} is typed
by first pressing and holding the @kbd{CONTROL} key, next
pressing the @kbd{d} key, and finally releasing both keys.

For the sake of brevity, throughout this @value{DOCUMENT}, we refer to
Brian Kernighan's version of @command{awk} as ``BWK @command{awk}.''
(@xref{Other Versions} for information on his and other versions.)

@ifset FOR_PRINT
@quotation NOTE
Notes of interest look like this.
@end quotation

@quotation CAUTION
Cautionary or warning notes look like this.
@end quotation
@end ifset

@c fakenode --- for prepinfo
@unnumberedsubsec Dark Corners
@cindex Kernighan, Brian
@quotation
@i{Dark corners are basically fractal---no matter how much
you illuminate, there's always a smaller but darker one.}
@author Brian Kernighan
@end quotation

@cindex d.c., See dark corner
@cindex dark corner
Until the POSIX standard (and @cite{@value{TITLE}}),
many features of @command{awk} were either poorly documented or not
documented at all.  Descriptions of such features
(often called ``dark corners'') are noted in this @value{DOCUMENT} with
@iftex
the picture of a flashlight in the margin, as shown here.
@value{DARKCORNER}
@end iftex
@ifnottex
``(d.c.).''
@end ifnottex
@ifclear FOR_PRINT
They also appear in the index under the heading ``dark corner.''
@end ifclear

But, as noted by the opening quote, any coverage of dark
corners is by definition incomplete.

@cindex c.e., See common extensions
Extensions to the standard @command{awk} language that are supported by
more than one @command{awk} implementation are marked
@ifclear FOR_PRINT
``@value{COMMONEXT},'' and listed in the index under ``common extensions''
and ``extensions, common.''
@end ifclear
@ifset FOR_PRINT
``@value{COMMONEXT}'' for ``common extension.''
@end ifset

@node Manual History
@unnumberedsec The GNU Project and This Book

@cindex FSF (Free Software Foundation)
@cindex Free Software Foundation (FSF)
@cindex Stallman, Richard
The Free Software Foundation (FSF) is a nonprofit organization dedicated
to the production and distribution of freely distributable software.
It was founded by Richard M.@: Stallman, the author of the original
Emacs editor.  GNU Emacs is the most widely used version of Emacs today.

@cindex GNU Project
@cindex GPL (General Public License)
@cindex General Public License, See GPL
@cindex documentation, online
The GNU@footnote{GNU stands for ``GNU's Not Unix.''}
Project is an ongoing effort on the part of the Free Software
Foundation to create a complete, freely distributable, POSIX-compliant
computing environment.
The FSF uses the GNU General Public License (GPL) to ensure that
its software's
source code is always available to the end user.
@ifclear FOR_PRINT
A copy of the GPL is included
@ifnotinfo
in this @value{DOCUMENT}
@end ifnotinfo
for your reference
(@pxref{Copying}).
@end ifclear
The GPL applies to the C language source code for @command{gawk}.
To find out more about the FSF and the GNU Project online,
see @uref{https://www.gnu.org, the GNU Project's home page}.
This @value{DOCUMENT} may also be read from
@uref{https://www.gnu.org/software/gawk/manual/, GNU's website}.

@ifclear FOR_PRINT
A shell, an editor (Emacs), highly portable optimizing C, C++, and
Objective-C compilers, a symbolic debugger and dozens of large and
small utilities (such as @command{gawk}), have all been completed and are
freely available.  The GNU operating
system kernel (the HURD), has been released but remains in an early
stage of development.

@cindex Linux
@cindex GNU/Linux
@cindex operating systems, BSD-based
Until the GNU operating system is more fully developed, you should
consider using GNU/Linux, a freely distributable, Unix-like operating
system for Intel,
Power Architecture,
Sun SPARC, IBM S/390, and other
systems.@footnote{The terminology ``GNU/Linux'' is explained
in the @ref{Glossary}.}
Many GNU/Linux distributions are
available for download from the Internet.
@end ifclear

@ifnotinfo
The @value{DOCUMENT} you are reading is actually free---at least, the
information in it is free to anyone.  The machine-readable
source code for the @value{DOCUMENT} comes with @command{gawk}.
@ifclear FOR_PRINT
(Take a moment to check the Free Documentation
License in @ref{GNU Free Documentation License}.)
@end ifclear
@end ifnotinfo

@cindex Close, Diane
The @value{DOCUMENT} itself has gone through multiple previous editions.
Paul Rubin wrote the very first draft of @cite{The GAWK Manual};
it was around 40 pages long.
Diane Close and Richard Stallman improved it, yielding a
version that was
around 90 pages and barely described the original, ``old''
version of @command{awk}.

I started working with that version in the fall of 1988.
As work on it progressed,
the FSF published several preliminary versions (numbered 0.@var{x}).
In 1996, edition 1.0 was released with @command{gawk} 3.0.0.
The FSF published the first two editions under
the title @cite{The GNU Awk User's Guide}.
@ifset FOR_PRINT
SSC published two editions of the @value{DOCUMENT} under the
title @cite{Effective awk Programming}, and O'Reilly published
the third edition in 2001.
@end ifset

This edition maintains the basic structure of the previous editions.
For FSF edition 4.0, the content was thoroughly reviewed and updated. All
references to @command{gawk} versions prior to 4.0 were removed.
Of significant note for that edition was the addition of @ref{Debugger}.

For FSF edition
@ifclear FOR_PRINT
@value{EDITION},
@end ifclear
@ifset FOR_PRINT
@value{EDITION}
(the fourth edition as published by O'Reilly),
@end ifset
the content has been reorganized into parts,
and the major new additions are @ref{Arbitrary Precision Arithmetic},
and @ref{Dynamic Extensions}.

This @value{DOCUMENT} will undoubtedly continue to evolve.  If you
find an error in the @value{DOCUMENT}, please report it!  @xref{Bugs}
for information on submitting problem reports electronically.

@ifset FOR_PRINT
@c fakenode --- for prepinfo
@unnumberedsec How to Stay Current

You may have a newer version of @command{gawk} than the
one described here.  To find out what has changed,
you should first look at the @file{NEWS} file in the @command{gawk}
distribution, which provides a high-level summary of the changes in
each release.

You can then look at the @uref{https://www.gnu.org/software/gawk/manual/,
online version} of this @value{DOCUMENT} to read about any new features.
@end ifset

@ifclear FOR_PRINT
@node How To Contribute
@unnumberedsec How to Contribute

As the maintainer of GNU @command{awk}, I once thought that I would be
able to manage a collection of publicly available @command{awk} programs
and I even solicited contributions.  Making things available on the Internet
helps keep the @command{gawk} distribution down to manageable size.

The initial collection of material, such as it is, is still available
at @uref{ftp://ftp.freefriends.org/arnold/Awkstuff}.

In the hopes of doing something more broad, I acquired the
@code{awklang.org} domain.  Late in 2017, a volunteer took on the task
of managing it.

If you have written an interesting @command{awk} program, that
you would like to share with the rest of the world, please see
@uref{http://www.awklang.org} and use the ``Contact'' link.

If you have written a @command{gawk} extension, please see
@ref{gawkextlib}.
@end ifclear

@node Acknowledgments
@unnumberedsec Acknowledgments

The initial draft of @cite{The GAWK Manual} had the following acknowledgments:

@quotation
Many people need to be thanked for their assistance in producing this
manual.  Jay Fenlason contributed many ideas and sample programs.  Richard
Mlynarik and Robert Chassell gave helpful comments on drafts of this
manual.  The paper @cite{A Supplemental Document for AWK} by John W.@:
Pierce of the Chemistry Department at UC San Diego, pinpointed several
issues relevant both to @command{awk} implementation and to this manual, that
would otherwise have escaped us.
@end quotation

@cindex Stallman, Richard
I would like to acknowledge Richard M.@: Stallman, for his vision of a
better world and for his courage in founding the FSF and starting the
GNU Project.

@ifclear FOR_PRINT
Earlier editions of this @value{DOCUMENT} had the following acknowledgements:
@end ifclear
@ifset FOR_PRINT
The previous edition of this @value{DOCUMENT} had
the following acknowledgements:
@end ifset

@quotation
The following people (in alphabetical order)
provided helpful comments on various
versions of this book:
Rick Adams,
Dr.@: Nelson H.F. Beebe,
Karl Berry,
Dr.@: Michael Brennan,
Rich Burridge,
Claire Cloutier,
Diane Close,
Scott Deifik,
Christopher (``Topher'') Eliot,
Jeffrey Friedl,
Dr.@: Darrel Hankerson,
Michal Jaegermann,
Dr.@: Richard J.@: LeBlanc,
Michael Lijewski,
Pat Rankin,
Miriam Robbins,
Mary Sheehan,
and
Chuck Toporek.

@cindex Berry, Karl
@cindex Chassell, Robert J.@:
@c @cindex Texinfo
Robert J.@: Chassell provided much valuable advice on
the use of Texinfo.
He also deserves special thanks for
convincing me @emph{not} to title this @value{DOCUMENT}
@cite{How to Gawk Politely}.
Karl Berry helped significantly with the @TeX{} part of Texinfo.

@cindex Hartholz, Marshall
@cindex Hartholz, Elaine
@cindex Schreiber, Bert
@cindex Schreiber, Rita
I would like to thank Marshall and Elaine Hartholz of Seattle and
Dr.@: Bert and Rita Schreiber of Detroit for large amounts of quiet vacation
time in their homes, which allowed me to make significant progress on
this @value{DOCUMENT} and on @command{gawk} itself.

@cindex Hughes, Phil
Phil Hughes of SSC
contributed in a very important way by loaning me his laptop GNU/Linux
system, not once, but twice, which allowed me to do a lot of work while
away from home.

@cindex Trueman, David
David Trueman deserves special credit; he has done a yeoman job
of evolving @command{gawk} so that it performs well and without bugs.
Although he is no longer involved with @command{gawk},
working with him on this project was a significant pleasure.

@cindex Drepper, Ulrich
@cindex GNITS mailing list
@cindex mailing list, GNITS
The intrepid members of the GNITS mailing list, and most notably Ulrich
Drepper, provided invaluable help and feedback for the design of the
internationalization features.

Chuck Toporek, Mary Sheehan, and Claire Cloutier of O'Reilly & Associates contributed
significant editorial help for this @value{DOCUMENT} for the
3.1 release of @command{gawk}.
@end quotation

@cindex Beebe, Nelson H.F.@:
@cindex Buening, Andreas
@cindex Collado, Manuel
@cindex Colombo, Antonio
@cindex Davies, Stephen
@cindex Deifik, Scott
@cindex Demaille, Akim
@cindex G., Daniel Richard
@cindex Guerrero, Juan Manuel
@cindex Hankerson, Darrel
@cindex Jaegermann, Michal
@cindex Kahrs, J@"urgen
@cindex Kasal, Stepan
@cindex Malmberg, John
@cindex Pitts, Dave
@cindex Ramey, Chet
@cindex Rankin, Pat
@cindex Schorr, Andrew
@cindex Vinschen, Corinna
@cindex Zaretskii, Eli

Dr.@: Nelson Beebe,
Andreas Buening,
Dr.@: Manuel Collado,
Antonio Colombo,
Stephen Davies,
Scott Deifik,
Akim Demaille,
Daniel Richard G.,
Juan Manuel Guerrero,
Darrel Hankerson,
Michal Jaegermann,
J@"urgen Kahrs,
Stepan Kasal,
John Malmberg,
Dave Pitts,
Chet Ramey,
Pat Rankin,
Andrew Schorr,
Corinna Vinschen,
and Eli Zaretskii
(in alphabetical order)
make up the current @command{gawk} ``crack portability team.''  Without
their hard work and help, @command{gawk} would not be nearly the robust,
portable program it is today.  It has been and continues to be a pleasure
working with this team of fine people.

Notable code and documentation contributions were made by
a number of people. @xref{Contributors} for the full list.

@ifset FOR_PRINT
@cindex Oram, Andy
Thanks to Andy Oram of O'Reilly Media for initiating
the fourth edition and for his support during the work.
Thanks to Jasmine Kwityn for her copyediting work.
@end ifset

Thanks to Michael Brennan for the Forewords.

@cindex Duman, Patrice
@cindex Berry, Karl
Thanks to Patrice Dumas for the new @command{makeinfo} program.
Thanks to Karl Berry, who continues to work to keep
the Texinfo markup language sane.

@cindex Kernighan, Brian
@cindex Brennan, Michael
@cindex Day, Robert P.J.@:
Robert P.J.@: Day, Michael Brennan, and Brian Kernighan kindly acted as
reviewers for the 2015 edition of this @value{DOCUMENT}. Their feedback
helped improve the final work.

I would also like to thank Brian Kernighan for his invaluable assistance during the
testing and debugging of @command{gawk}, and for his ongoing
help and advice in clarifying numerous points about the language.
We could not have done nearly as good a job on either @command{gawk}
or its documentation without his help.

Brian is in a class by himself as a programmer and technical
author.  I have to thank him (yet again) for his ongoing friendship
and for being a role model to me for close to 30 years!
Having him as a reviewer is an exciting privilege. It has also
been extremely humbling@enddots{}

@cindex Robbins, Miriam
@cindex Robbins, Jean
@cindex Robbins, Harry
@cindex G-d
I must thank my wonderful wife, Miriam, for her patience through
the many versions of this project, for her proofreading,
and for sharing me with the computer.
I would like to thank my parents for their love, and for the grace with
which they raised and educated me.
Finally, I also must acknowledge my gratitude to G-d, for the many opportunities
He has sent my way, as well as for the gifts He has given me with which to
take advantage of those opportunities.
@ifnotdocbook
@sp 2
@noindent
Arnold Robbins @*
Nof Ayalon @*
Israel @*
February 2015
@end ifnotdocbook

@ifnotinfo
@part @value{PART1}The @command{awk} Language
@end ifnotinfo

@ifdocbook

Part I describes the @command{awk} language and @command{gawk} program
in detail.  It starts with the basics, and continues through all of
the features of @command{awk}. Included also are many, but not all,
of the features of @command{gawk}.  This part contains the
following chapters:

@itemize @value{BULLET}
@item
@ref{Getting Started}

@item
@ref{Invoking Gawk}

@item
@ref{Regexp}

@item
@ref{Reading Files}

@item
@ref{Printing}

@item
@ref{Expressions}

@item
@ref{Patterns and Actions}

@item
@ref{Arrays}

@item
@ref{Functions}
@end itemize
@end ifdocbook

@node Getting Started
@chapter Getting Started with @command{awk}
@c @cindex script, definition of
@c @cindex rule, definition of
@c @cindex program, definition of
@c @cindex basic function of @command{awk}
@cindex @command{awk}, function of

The basic function of @command{awk} is to search files for lines (or other
units of text) that contain certain patterns.  When a line matches one
of the patterns, @command{awk} performs specified actions on that line.
@command{awk} continues to process input lines in this way until it reaches
the end of the input files.

@cindex @command{awk}, uses for
@cindex programming languages@comma{} data-driven vs.@: procedural
@cindex @command{awk} programs
Programs in @command{awk} are different from programs in most other languages,
because @command{awk} programs are @dfn{data driven} (i.e., you describe
the data you want to work with and then what to do when you find it).
Most other languages are @dfn{procedural}; you have to describe, in great
detail, every step the program should take.  When working with procedural
languages, it is usually much
harder to clearly describe the data your program will process.
For this reason, @command{awk} programs are often refreshingly easy to
read and write.

@cindex program, definition of
@cindex rule, definition of
When you run @command{awk}, you specify an @command{awk} @dfn{program} that
tells @command{awk} what to do.  The program consists of a series of
@dfn{rules} (it may also contain @dfn{function definitions},
an advanced feature that we will ignore for now;
@pxref{User-defined}).  Each rule specifies one
pattern to search for and one action to perform
upon finding the pattern.

Syntactically, a rule consists of a @dfn{pattern} followed by an
@dfn{action}.  The action is enclosed in braces to separate it from the
pattern.  Newlines usually separate rules.  Therefore, an @command{awk}
program looks like this:

@example
@var{pattern} @{ @var{action} @}
@var{pattern} @{ @var{action} @}
@dots{}
@end example

@menu
* Running gawk::                How to run @command{gawk} programs; includes
                                command-line syntax.
* Sample Data Files::           Sample data files for use in the @command{awk}
                                programs illustrated in this @value{DOCUMENT}.
* Very Simple::                 A very simple example.
* Two Rules::                   A less simple one-line example using two
                                rules.
* More Complex::                A more complex example.
* Statements/Lines::            Subdividing or combining statements into
                                lines.
* Other Features::              Other Features of @command{awk}.
* When::                        When to use @command{gawk} and when to use
                                other things.
* Intro Summary::               Summary of the introduction.
@end menu

@node Running gawk
@section How to Run @command{awk} Programs

@cindex @command{awk} programs, running
There are several ways to run an @command{awk} program.  If the program is
short, it is easiest to include it in the command that runs @command{awk},
like this:

@example
awk '@var{program}' @var{input-file1} @var{input-file2} @dots{}
@end example

@cindex command line, formats
When the program is long, it is usually more convenient to put it in a file
and run it with a command like this:

@example
awk -f @var{program-file} @var{input-file1} @var{input-file2} @dots{}
@end example

This @value{SECTION} discusses both mechanisms, along with several
variations of each.

@menu
* One-shot::                    Running a short throwaway @command{awk}
                                program.
* Read Terminal::               Using no input files (input from the keyboard
                                instead).
* Long::                        Putting permanent @command{awk} programs in
                                files.
* Executable Scripts::          Making self-contained @command{awk} programs.
* Comments::                    Adding documentation to @command{gawk}
                                programs.
* Quoting::                     More discussion of shell quoting issues.
@end menu

@node One-shot
@subsection One-Shot Throwaway @command{awk} Programs

Once you are familiar with @command{awk}, you will often type in simple
programs the moment you want to use them.  Then you can write the
program as the first argument of the @command{awk} command, like this:

@example
awk '@var{program}' @var{input-file1} @var{input-file2} @dots{}
@end example

@noindent
where @var{program} consists of a series of patterns and
actions, as described earlier.

@cindex single quote (@code{'})
@cindex @code{'} (single quote)
This command format instructs the @dfn{shell}, or command interpreter,
to start @command{awk} and use the @var{program} to process records in the
input file(s).  There are single quotes around @var{program} so
the shell won't interpret any @command{awk} characters as special shell
characters.  The quotes also cause the shell to treat all of @var{program} as
a single argument for @command{awk}, and allow @var{program} to be more
than one line long.

@cindex shells, scripts
@cindex @command{awk} programs, running, from shell scripts
This format is also useful for running short or medium-sized @command{awk}
programs from shell scripts, because it avoids the need for a separate
file for the @command{awk} program.  A self-contained shell script is more
reliable because there are no other files to misplace.

Later in this chapter, in
@ifdocbook
the @value{SECTION}
@end ifdocbook
@ref{Very Simple},
we'll see examples of several short,
self-contained programs.

@node Read Terminal
@subsection Running @command{awk} Without Input Files

@cindex standard input
@cindex input, standard
@cindex input files, running @command{awk} without
You can also run @command{awk} without any input files.  If you type the
following command line:

@example
awk '@var{program}'
@end example

@noindent
@command{awk} applies the @var{program} to the @dfn{standard input},
which usually means whatever you type on the keyboard.  This continues
until you indicate end-of-file by typing @kbd{Ctrl-d}.
(On non-POSIX operating systems, the end-of-file character may be different.)

@cindex files, input, See input files
@cindex input files, running @command{awk} without
@cindex @command{awk} programs, running, without input files
As an example, the following program prints a friendly piece of advice
(from Douglas Adams's @cite{The Hitchhiker's Guide to the Galaxy}),
to keep you from worrying about the complexities of computer
programming:

@example
$ @kbd{awk 'BEGIN @{ print "Don\47t Panic!" @}'}
@print{} Don't Panic!
@end example

@command{awk} executes statements associated with @code{BEGIN} before
reading any input.  If there are no other statements in your program,
as is the case here, @command{awk} just stops, instead of trying to read
input it doesn't know how to process.
The @samp{\47} is a magic way (explained later) of getting a single quote into
the program, without having to engage in ugly shell quoting tricks.

@quotation NOTE
If you use Bash as your shell, you should execute the
command @samp{set +H} before running this program interactively, to
disable the C shell-style command history, which treats @samp{!} as a
special character. We recommend putting this command into your personal
startup file.
@end quotation

This next simple @command{awk} program
emulates the @command{cat} utility; it copies whatever you type on the
keyboard to its standard output (why this works is explained shortly):

@example
$ @kbd{awk '@{ print @}'}
@kbd{Now is the time for all good men}
@print{} Now is the time for all good men
@kbd{to come to the aid of their country.}
@print{} to come to the aid of their country.
@kbd{Four score and seven years ago, ...}
@print{} Four score and seven years ago, ...
@kbd{What, me worry?}
@print{} What, me worry?
@kbd{Ctrl-d}
@end example

@node Long
@subsection Running Long Programs

@cindex @command{awk} programs, running
@cindex @command{awk} programs, lengthy
@cindex files, @command{awk} programs in
Sometimes @command{awk} programs are very long.  In these cases, it is
more convenient to put the program into a separate file.  In order to tell
@command{awk} to use that file for its program, you type:

@example
awk -f @var{source-file} @var{input-file1} @var{input-file2} @dots{}
@end example

@cindex @option{-f} option
@cindex command line, option @option{-f}
The @option{-f} instructs the @command{awk} utility to get the
@command{awk} program from the file @var{source-file} (@pxref{Options}).
Any @value{FN} can be used for @var{source-file}.  For example, you
could put the program:

@example
BEGIN @{ print "Don't Panic!" @}
@end example

@noindent
into the file @file{advice}.  Then this command:

@example
awk -f advice
@end example

@noindent
does the same thing as this one:

@example
awk 'BEGIN @{ print "Don\47t Panic!" @}'
@end example

@cindex quoting in @command{gawk} command lines
@noindent
This was explained earlier
(@pxref{Read Terminal}).
Note that you don't usually need single quotes around the @value{FN} that you
specify with @option{-f}, because most @value{FN}s don't contain any of the shell's
special characters.  Notice that in @file{advice}, the @command{awk}
program did not have single quotes around it.  The quotes are only needed
for programs that are provided on the @command{awk} command line.
(Also, placing the program in a file allows us to use a literal single quote in the program
text, instead of the magic @samp{\47}.)

@cindex single quote (@code{'}) in @command{gawk} command lines
@cindex @code{'} (single quote) in @command{gawk} command lines
If you want to clearly identify an @command{awk} program file as such,
you can add the extension @file{.awk} to the @value{FN}.  This doesn't
affect the execution of the @command{awk} program but it does make
``housekeeping'' easier.

@node Executable Scripts
@subsection Executable @command{awk} Programs
@cindex @command{awk} programs
@cindex @code{#} (number sign), @code{#!} (executable scripts)
@cindex Unix, @command{awk} scripts and
@cindex number sign (@code{#}), @code{#!} (executable scripts)

Once you have learned @command{awk}, you may want to write self-contained
@command{awk} scripts, using the @samp{#!} script mechanism.  You can do
this on many systems.@footnote{The @samp{#!} mechanism works on
GNU/Linux systems, BSD-based systems, and commercial Unix systems.}
For example, you could update the file @file{advice} to look like this:

@example
#! /bin/awk -f

BEGIN @{ print "Don't Panic!" @}
@end example

@noindent
After making this file executable (with the @command{chmod} utility),
simply type @samp{advice}
at the shell and the system arranges to run @command{awk} as if you had
typed @samp{awk -f advice}:

@example
$ @kbd{chmod +x advice}
$ @kbd{advice}
@print{} Don't Panic!
@end example

@noindent
(We assume you have the current directory in your shell's search
path variable [typically @code{$PATH}].  If not, you may need
to type @samp{./advice} at the shell.)

Self-contained @command{awk} scripts are useful when you want to write a
program that users can invoke without their having to know that the program is
written in @command{awk}.

@sidebar Understanding @samp{#!}
@cindex portability, @code{#!} (executable scripts)

@command{awk} is an @dfn{interpreted} language. This means that the
@command{awk} utility reads your program and then processes your data
according to the instructions in your program. (This is different
from a @dfn{compiled} language such as C, where your program is first
compiled into machine code that is executed directly by your system's
processor.)  The @command{awk} utility is thus termed an @dfn{interpreter}.
Many modern languages are interpreted.

The line beginning with @samp{#!} lists the full @value{FN} of an
interpreter to run and a single optional initial command-line argument
to pass to that interpreter.  The operating system then runs the
interpreter with the given argument and the full argument list of the
executed program.  The first argument in the list is the full @value{FN}
of the @command{awk} program.  The rest of the argument list contains
either options to @command{awk}, or @value{DF}s, or both. (Note that on
many systems @command{awk} may be found in @file{/usr/bin} instead of
in @file{/bin}.)

Some systems limit the length of the interpreter name to 32 characters.
Often, this can be dealt with by using a symbolic link.

You should not put more than one argument on the @samp{#!}
line after the path to @command{awk}. It does not work. The operating system
treats the rest of the line as a single argument and passes it to @command{awk}.
Doing this leads to confusing behavior---most likely a usage diagnostic
of some sort from @command{awk}.

@cindex @code{ARGC}/@code{ARGV} variables, portability and
@cindex portability, @code{ARGV} variable
@cindex dark corner, @code{ARGV} variable, value of
Finally, the value of @code{ARGV[0]}
(@pxref{Built-in Variables})
varies depending upon your operating system.
Some systems put @samp{awk} there, some put the full pathname
of @command{awk} (such as @file{/bin/awk}), and some put the name
of your script (@samp{advice}).  @value{DARKCORNER}
Don't rely on the value of @code{ARGV[0]}
to provide your script name.
@end sidebar

@node Comments
@subsection Comments in @command{awk} Programs
@cindex @code{#} (number sign), commenting
@cindex number sign (@code{#}), commenting
@cindex commenting
@cindex @command{awk} programs, documenting

A @dfn{comment} is some text that is included in a program for the sake
of human readers; it is not really an executable part of the program.  Comments
can explain what the program does and how it works.  Nearly all
programming languages have provisions for comments, as programs are
typically hard to understand without them.

In the @command{awk} language, a comment starts with the number sign
character (@samp{#}) and continues to the end of the line.
The @samp{#} does not have to be the first character on the line. The
@command{awk} language ignores the rest of a line following a number sign.
For example, we could have put the following into @file{advice}:

@example
# This program prints a nice, friendly message.  It helps
# keep novice users from being afraid of the computer.
BEGIN    @{ print "Don't Panic!" @}
@end example

You can put comment lines into keyboard-composed throwaway @command{awk}
programs, but this usually isn't very useful; the purpose of a
comment is to help you or another person understand the program
when reading it at a later time.

@cindex quoting, for small awk programs
@cindex single quote (@code{'}), vs.@: apostrophe
@cindex @code{'} (single quote), vs.@: apostrophe
@quotation CAUTION
As mentioned in
@ref{One-shot},
you can enclose short to medium-sized programs in single quotes,
in order to keep
your shell scripts self-contained.  When doing so, @emph{don't} put
an apostrophe (i.e., a single quote) into a comment (or anywhere else
in your program). The shell interprets the quote as the closing
quote for the entire program. As a result, usually the shell
prints a message about mismatched quotes, and if @command{awk} actually
runs, it will probably print strange messages about syntax errors.
For example, look at the following:

@example
$ @kbd{awk 'BEGIN @{ print "hello" @} # let's be cute'}
>
@end example

The shell sees that the first two quotes match, and that
a new quoted object begins at the end of the command line.
It therefore prompts with the secondary prompt, waiting for more input.
With Unix @command{awk}, closing the quoted string produces this result:

@example
$ @kbd{awk '@{ print "hello" @} # let's be cute'}
> @kbd{'}
@error{} awk: can't open file be
@error{}  source line number 1
@end example

@cindex @code{\} (backslash)
@cindex backslash (@code{\})
Putting a backslash before the single quote in @samp{let's} wouldn't help,
because backslashes are not special inside single quotes.
The next @value{SUBSECTION} describes the shell's quoting rules.
@end quotation

@node Quoting
@subsection Shell Quoting Issues
@cindex shell quoting, rules for

@menu
* DOS Quoting::                 Quoting in Windows Batch Files.
@end menu

For short to medium-length @command{awk} programs, it is most convenient
to enter the program on the @command{awk} command line.
This is best done by enclosing the entire program in single quotes.
This is true whether you are entering the program interactively at
the shell prompt, or writing it as part of a larger shell script:

@example
awk '@var{program text}' @var{input-file1} @var{input-file2} @dots{}
@end example

@cindex shells, quoting, rules for
@cindex Bourne shell, quoting rules for
Once you are working with the shell, it is helpful to have a basic
knowledge of shell quoting rules.  The following rules apply only to
POSIX-compliant, Bourne-style shells (such as Bash, the GNU Bourne-Again
Shell).  If you use the C shell, you're on your own.

Before diving into the rules, we introduce a concept that appears
throughout this @value{DOCUMENT}, which is that of the @dfn{null},
or empty, string.

The null string is character data that has no value.
In other words, it is empty.  It is written in @command{awk} programs
like this: @code{""}. In the shell, it can be written using single
or double quotes: @code{""} or @code{''}. Although the null string has
no characters in it, it does exist. For example, consider this command:

@example
$ @kbd{echo ""}
@end example

@noindent
Here, the @command{echo} utility receives a single argument, even
though that argument has no characters in it. In the rest of this
@value{DOCUMENT}, we use the terms @dfn{null string} and @dfn{empty string}
interchangeably.  Now, on to the quoting rules:

@itemize @value{BULLET}
@item
Quoted items can be concatenated with nonquoted items as well as with other
quoted items.  The shell turns everything into one argument for
the command.

@item
Preceding any single character with a backslash (@samp{\}) quotes
that character.  The shell removes the backslash and passes the quoted
character on to the command.

@item
@cindex @code{\} (backslash), in shell commands
@cindex backslash (@code{\}), in shell commands
@cindex single quote (@code{'}), in shell commands
@cindex @code{'} (single quote), in shell commands
Single quotes protect everything between the opening and closing quotes.
The shell does no interpretation of the quoted text, passing it on verbatim
to the command.
It is @emph{impossible} to embed a single quote inside single-quoted text.
Refer back to
@ref{Comments}
for an example of what happens if you try.

@item
@cindex double quote (@code{"}), in shell commands
@cindex @code{"} (double quote), in shell commands
Double quotes protect most things between the opening and closing quotes.
The shell does at least variable and command substitution on the quoted text.
Different shells may do additional kinds of processing on double-quoted text.

Because certain characters within double-quoted text are processed by the shell,
they must be @dfn{escaped} within the text.  Of note are the characters
@samp{$}, @samp{`}, @samp{\}, and @samp{"}, all of which must be preceded by
a backslash within double-quoted text if they are to be passed on literally
to the program.  (The leading backslash is stripped first.)
Thus, the example seen
@ifnotinfo
previously
@end ifnotinfo
in @ref{Read Terminal}:

@example
awk 'BEGIN @{ print "Don\47t Panic!" @}'
@end example

@noindent
could instead be written this way:

@example
$ @kbd{awk "BEGIN @{ print \"Don't Panic!\" @}"}
@print{} Don't Panic!
@end example

@cindex single quote (@code{'}), with double quotes
@cindex @code{'} (single quote), with double quotes
Note that the single quote is not special within double quotes.

@item
Null strings are removed when they occur as part of a non-null
command-line argument, while explicit null objects are kept.
For example, to specify that the field separator @code{FS} should
be set to the null string, use:

@example
awk -F "" '@var{program}' @var{files} # correct
@end example

@noindent
@cindex null strings in @command{gawk} arguments, quoting and
Don't use this:

@example
awk -F"" '@var{program}' @var{files}  # wrong!
@end example

@noindent
In the second case, @command{awk} attempts to use the text of the program
as the value of @code{FS}, and the first @value{FN} as the text of the program!
This results in syntax errors at best, and confusing behavior at worst.
@end itemize

@cindex quoting in @command{gawk} command lines, tricks for
Mixing single and double quotes is difficult.  You have to resort
to shell quoting tricks, like this:

@example
$ @kbd{awk 'BEGIN @{ print "Here is a single quote <'"'"'>" @}'}
@print{} Here is a single quote <'>
@end example

@noindent
This program consists of three concatenated quoted strings.  The first and the
third are single-quoted, and the second is double-quoted.

This can be ``simplified'' to:

@example
$ @kbd{awk 'BEGIN @{ print "Here is a single quote <'\''>" @}'}
@print{} Here is a single quote <'>
@end example

@noindent
Judge for yourself which of these two is the more readable.

Another option is to use double quotes, escaping the embedded, @command{awk}-level
double quotes:

@example
$ @kbd{awk "BEGIN @{ print \"Here is a single quote <'>\" @}"}
@print{} Here is a single quote <'>
@end example

@noindent
This option is also painful, because double quotes, backslashes, and dollar signs
are very common in more advanced @command{awk} programs.

A third option is to use the octal escape sequence equivalents
(@pxref{Escape Sequences})
for the
single- and double-quote characters, like so:

@example
@group
$ @kbd{awk 'BEGIN @{ print "Here is a single quote <\47>" @}'}
@print{} Here is a single quote <'>
$ @kbd{awk 'BEGIN @{ print "Here is a double quote <\42>" @}'}
@print{} Here is a double quote <">
@end group
@end example

@noindent
This works nicely, but you should comment clearly what the
escapes mean.

A fourth option is to use command-line variable assignment, like this:

@example
$ @kbd{awk -v sq="'" 'BEGIN @{ print "Here is a single quote <" sq ">" @}'}
@print{} Here is a single quote <'>
@end example

(Here, the two string constants and the value of @code{sq} are concatenated
into a single string that is printed by @code{print}.)

If you really need both single and double quotes in your @command{awk}
program, it is probably best to move it into a separate file, where
the shell won't be part of the picture and you can say what you mean.

@node DOS Quoting
@subsubsection Quoting in MS-Windows Batch Files

@ignore
Date: Wed, 21 May 2008 09:58:43 +0200 (CEST)
From: jeroen.brink@inter.NL.net
Subject: (g)awk "contribution"
To: arnold@skeeve.com
Message-id: <42220.193.172.132.34.1211356723.squirrel@webmail.internl.net>

Hello Arnold,

maybe you can help me out. Found your email on the GNU/awk online manual
pages.

I've searched hard to figure out how, on Windows, to print double quotes.
Couldn't find it in the Quotes area, nor on google or elsewhere. Finally i
figured out how to do this myself.

How to print all lines in a file surrounded by double quotes (on Windows):

gawk "{ print \"\042\" $0 \"\042\" }" <file>

Maybe this is a helpfull tip for other (Windows) gawk users. However, i
don't have a clue as to where to "publish" this tip! Do you?

Kind regards,

Jeroen Brink
@end ignore

Although this @value{DOCUMENT} generally only worries about POSIX systems and the
POSIX shell, the following issue arises often enough for many users that
it is worth addressing.

@cindex Brink, Jeroen
The ``shells'' on Microsoft Windows systems use the double-quote
character for quoting, and make it difficult or impossible to include an
escaped double-quote character in a command-line script.  The following
example, courtesy of Jeroen Brink, shows how to escape the double quotes
from this one liner script that prints all lines in a file surrounded by
double quotes:

@example
@{ print "\"" $0 "\"" @}
@end example

@noindent
In an MS-Windows command-line the one-liner script above may be passed as
follows:

@example
gawk "@{ print \"\042\" $0 \"\042\" @}" @var{file}
@end example

In this example the @samp{\042} is the octal code for a double-quote;
@command{gawk} converts it into a real double-quote for output by
the @code{print} statement.

In MS-Windows escaping double-quotes is a little tricky because you use
backslashes to escape double-quotes, but backslashes themselves are not
escaped in the usual way; indeed they are either duplicated or not,
depending upon whether there is a subsequent double-quote.  The MS-Windows
rule for double-quoting a string is the following:

@enumerate
@item
For each double quote in the original string, let @var{N} be the number
of backslash(es) before it, @var{N} might be zero. Replace these @var{N}
backslash(es) by @math{2@value{TIMES}@var{N}+1} backslash(es)

@item
Let @var{N} be the number of backslash(es) tailing the original string,
@var{N} might be zero. Replace these @var{N} backslash(es) by
@math{2@value{TIMES}@var{N}} backslash(es)

@item
Surround the resulting string by double-quotes.
@end enumerate

So to double-quote the one-liner script @samp{@{ print "\"" $0 "\"" @}}
from the previous example you would do it this way:

@example
gawk "@{ print \"\\\"\" $0 \"\\\"\" @}" @var{file}
@end example

@noindent
However, the use of @samp{\042} instead of @samp{\\\"} is also possible
and easier to read, because backslashes that are not followed by a
double-quote don't need duplication.

@node Sample Data Files
@section @value{DDF}s for the Examples

@cindex input files, examples
@cindex @code{mail-list} file
Many of the examples in this @value{DOCUMENT} take their input from two sample
@value{DF}s.  The first, @file{mail-list}, represents a list of peoples' names
together with their email addresses and information about those people.
The second @value{DF}, called @file{inventory-shipped}, contains
information about monthly shipments.  In both files,
each line is considered to be one @dfn{record}.

In @file{mail-list}, each record contains the name of a person,
his/her phone number, his/her email address, and a code for his/her relationship
with the author of the list.
The columns are aligned using spaces.
An @samp{A} in the last column
means that the person is an acquaintance.  An @samp{F} in the last
column means that the person is a friend.
An @samp{R} means that the person is a relative:

@example
@c system if test ! -d eg      ; then mkdir eg      ; fi
@c system if test ! -d eg/lib  ; then mkdir eg/lib  ; fi
@c system if test ! -d eg/data ; then mkdir eg/data ; fi
@c system if test ! -d eg/prog ; then mkdir eg/prog ; fi
@c system if test ! -d eg/misc ; then mkdir eg/misc ; fi
@c file eg/data/mail-list
Amelia       555-5553     amelia.zodiacusque@@gmail.com    F
Anthony      555-3412     anthony.asserturo@@hotmail.com   A
Becky        555-7685     becky.algebrarum@@gmail.com      A
Bill         555-1675     bill.drowning@@hotmail.com       A
Broderick    555-0542     broderick.aliquotiens@@yahoo.com R
Camilla      555-2912     camilla.infusarum@@skynet.be     R
Fabius       555-1234     fabius.undevicesimus@@ucb.edu    F
Julie        555-6699     julie.perscrutabor@@skeeve.com   F
Martin       555-6480     martin.codicibus@@hotmail.com    A
Samuel       555-3430     samuel.lanceolis@@shu.edu        A
Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@c endfile
@end example

@cindex @code{inventory-shipped} file
The @value{DF} @file{inventory-shipped} represents
information about shipments during the year.
Each record contains the month, the number
of green crates shipped, the number of red boxes shipped, the number of
orange bags shipped, and the number of blue packages shipped,
respectively.  There are 16 entries, covering the 12 months of last year
and the first four months of the current year.
An empty line separates the data for the two years:

@example
@c file eg/data/inventory-shipped
Jan  13  25  15 115
Feb  15  32  24 226
Mar  15  24  34 228
Apr  31  52  63 420
May  16  34  29 208
Jun  31  42  75 492
Jul  24  34  67 436
Aug  15  34  47 316
Sep  13  55  37 277
Oct  29  54  68 525
Nov  20  87  82 577
Dec  17  35  61 401

Jan  21  36  64 620
Feb  26  58  80 652
Mar  24  75  70 495
Apr  21  70  74 514
@c endfile
@end example

The sample files are included in the @command{gawk} distribution,
in the directory @file{awklib/eg/data}.

@node Very Simple
@section Some Simple Examples

The following command runs a simple @command{awk} program that searches the
input file @file{mail-list} for the character string @samp{li} (a
grouping of characters is usually called a @dfn{string};
the term @dfn{string} is based on similar usage in English, such
as ``a string of pearls'' or ``a string of cars in a train''):

@example
awk '/li/ @{ print $0 @}' mail-list
@end example

@noindent
When lines containing @samp{li} are found, they are printed because
@w{@samp{print $0}} means print the current line.  (Just @samp{print} by
itself means the same thing, so we could have written that
instead.)

You will notice that slashes (@samp{/}) surround the string @samp{li}
in the @command{awk} program.  The slashes indicate that @samp{li}
is the pattern to search for.  This type of pattern is called a
@dfn{regular expression}, which is covered in more detail later
(@pxref{Regexp}).
The pattern is allowed to match parts of words.
There are
single quotes around the @command{awk} program so that the shell won't
interpret any of it as special shell characters.

Here is what this program prints:

@example
$ @kbd{awk '/li/ @{ print $0 @}' mail-list}
@print{} Amelia       555-5553     amelia.zodiacusque@@gmail.com    F
@print{} Broderick    555-0542     broderick.aliquotiens@@yahoo.com R
@print{} Julie        555-6699     julie.perscrutabor@@skeeve.com   F
@print{} Samuel       555-3430     samuel.lanceolis@@shu.edu        A
@end example

@cindex actions, default
@cindex patterns, default
In an @command{awk} rule, either the pattern or the action can be omitted,
but not both.  If the pattern is omitted, then the action is performed
for @emph{every} input line.  If the action is omitted, the default
action is to print all lines that match the pattern.

@cindex actions, empty
Thus, we could leave out the action (the @code{print} statement and the
braces) in the previous example and the result would be the same:
@command{awk} prints all lines matching the pattern @samp{li}.  By comparison,
omitting the @code{print} statement but retaining the braces makes an
empty action that does nothing (i.e., no lines are printed).

@cindex @command{awk} programs, one-line examples
Many practical @command{awk} programs are just a line or two long.  Following is a
collection of useful, short programs to get you started.  Some of these
programs contain constructs that haven't been covered yet. (The description
of the program will give you a good idea of what is going on, but you'll
need to read the rest of the @value{DOCUMENT} to become an @command{awk} expert!)
Most of the examples use a @value{DF} named @file{data}.  This is just a
placeholder; if you use these programs yourself, substitute
your own @value{FN}s for @file{data}.
For future reference, note that there is often more than
one way to do things in @command{awk}.  At some point, you may want
to look back at these examples and see if
you can come up with different ways to do the same things shown here:

@itemize @value{BULLET}
@item
Print every line that is longer than 80 characters:

@example
awk 'length($0) > 80' data
@end example

The sole rule has a relational expression as its pattern and has no
action---so it uses the default action, printing the record.

@item
Print the length of the longest input line:

@example
@group
awk '@{ if (length($0) > max) max = length($0) @}
     END @{ print max @}' data
@end group
@end example

The code associated with @code{END} executes after all
input has been read; it's the other side of the coin to @code{BEGIN}.

@cindex @command{expand} utility
@item
Print the length of the longest line in @file{data}:

@example
expand data | awk '@{ if (x < length($0)) x = length($0) @}
                   END @{ print "maximum line length is " x @}'
@end example

This example differs slightly from the previous one:
the input is processed by the @command{expand} utility to change TABs
into spaces, so the widths compared are actually the right-margin columns,
as opposed to the number of input characters on each line.

@item
Print every line that has at least one field:

@example
awk 'NF > 0' data
@end example

This is an easy way to delete blank lines from a file (or rather, to
create a new file similar to the old file but from which the blank lines
have been removed).

@item
Print seven random numbers from 0 to 100, inclusive:

@example
awk 'BEGIN @{ for (i = 1; i <= 7; i++)
                 print int(101 * rand()) @}'
@end example

@item
Print the total number of bytes used by @var{files}:

@example
ls -l @var{files} | awk '@{ x += $5 @}
                   END @{ print "total bytes: " x @}'
@end example

@item
Print the total number of kilobytes used by @var{files}:

@c Don't use \ continuation, not discussed yet
@c Remember that awk does floating point division,
@c no need for (x+1023) / 1024
@example
ls -l @var{files} | awk '@{ x += $5 @}
   END @{ print "total K-bytes:", x / 1024 @}'
@end example

@item
Print a sorted list of the login names of all users:

@example
awk -F: '@{ print $1 @}' /etc/passwd | sort
@end example

@item
Count the lines in a file:

@example
awk 'END @{ print NR @}' data
@end example

@item
Print the even-numbered lines in the @value{DF}:

@example
awk 'NR % 2 == 0' data
@end example

If you used the expression @samp{NR % 2 == 1} instead,
the program would print the odd-numbered lines.
@end itemize

@node Two Rules
@section An Example with Two Rules
@cindex @command{awk} programs

The @command{awk} utility reads the input files one line at a
time.  For each line, @command{awk} tries the patterns of each rule.
If several patterns match, then several actions execute in the order in
which they appear in the @command{awk} program.  If no patterns match, then
no actions run.

After processing all the rules that match the line (and perhaps there are none),
@command{awk} reads the next line.  (However,
@pxref{Next Statement}
@ifdocbook
and @ref{Nextfile Statement}.)
@end ifdocbook
@ifnotdocbook
and also @pxref{Nextfile Statement}.)
@end ifnotdocbook
This continues until the program reaches the end of the file.
For example, the following @command{awk} program contains two rules:

@example
/12/  @{ print $0 @}
/21/  @{ print $0 @}
@end example

@noindent
The first rule has the string @samp{12} as the
pattern and @samp{print $0} as the action.  The second rule has the
string @samp{21} as the pattern and also has @samp{print $0} as the
action.  Each rule's action is enclosed in its own pair of braces.

This program prints every line that contains the string
@samp{12} @emph{or} the string @samp{21}.  If a line contains both
strings, it is printed twice, once by each rule.

This is what happens if we run this program on our two sample @value{DF}s,
@file{mail-list} and @file{inventory-shipped}:

@example
$ @kbd{awk '/12/ @{ print $0 @}}
>      @kbd{/21/ @{ print $0 @}' mail-list inventory-shipped}
@print{} Anthony      555-3412     anthony.asserturo@@hotmail.com   A
@print{} Camilla      555-2912     camilla.infusarum@@skynet.be     R
@print{} Fabius       555-1234     fabius.undevicesimus@@ucb.edu    F
@print{} Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@print{} Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@print{} Jan  21  36  64 620
@print{} Apr  21  70  74 514
@end example

@noindent
Note how the line beginning with @samp{Jean-Paul}
in @file{mail-list} was printed twice, once for each rule.

@node More Complex
@section A More Complex Example

Now that we've mastered some simple tasks, let's look at
what typical @command{awk}
programs do.  This example shows how @command{awk} can be used to
summarize, select, and rearrange the output of another utility.  It uses
features that haven't been covered yet, so don't worry if you don't
understand all the details:

@example
ls -l | awk '$6 == "Nov" @{ sum += $5 @}
             END @{ print sum @}'
@end example

@cindex @command{ls} utility
This command prints the total number of bytes in all the files in the
current directory that were last modified in November (of any year).
The @w{@samp{ls -l}} part of this example is a system command that gives
you a listing of the files in a directory, including each file's size and the date
the file was last modified. Its output looks like this:

@example
-rw-r--r--  1 arnold   user   1933 Nov  7 13:05 Makefile
-rw-r--r--  1 arnold   user  10809 Nov  7 13:03 awk.h
-rw-r--r--  1 arnold   user    983 Apr 13 12:14 awk.tab.h
-rw-r--r--  1 arnold   user  31869 Jun 15 12:20 awkgram.y
-rw-r--r--  1 arnold   user  22414 Nov  7 13:03 awk1.c
-rw-r--r--  1 arnold   user  37455 Nov  7 13:03 awk2.c
-rw-r--r--  1 arnold   user  27511 Dec  9 13:07 awk3.c
-rw-r--r--  1 arnold   user   7989 Nov  7 13:03 awk4.c
@end example

@noindent
@cindex line continuations, with C shell
The first field contains read-write permissions, the second field contains
the number of links to the file, and the third field identifies the file's owner.
The fourth field identifies the file's group.
The fifth field contains the file's size in bytes.  The
sixth, seventh, and eighth fields contain the month, day, and time,
respectively, that the file was last modified.  Finally, the ninth field
contains the @value{FN}.

@c @cindex automatic initialization
@cindex initialization, automatic
The @samp{$6 == "Nov"} in our @command{awk} program is an expression that
tests whether the sixth field of the output from @w{@samp{ls -l}}
matches the string @samp{Nov}.  Each time a line has the string
@samp{Nov} for its sixth field, @command{awk} performs the action
@samp{sum += $5}.  This adds the fifth field (the file's size) to the variable
@code{sum}.  As a result, when @command{awk} has finished reading all the
input lines, @code{sum} is the total of the sizes of the files whose
lines matched the pattern.  (This works because @command{awk} variables
are automatically initialized to zero.)

After the last line of output from @command{ls} has been processed, the
@code{END} rule executes and prints the value of @code{sum}.
In this example, the value of @code{sum} is 80600.

These more advanced @command{awk} techniques are covered in later
@value{SECTION}s
(@pxref{Action Overview}).  Before you can move on to more
advanced @command{awk} programming, you have to know how @command{awk} interprets
your input and displays your output.  By manipulating fields and using
@code{print} statements, you can produce some very useful and
impressive-looking reports.

@node Statements/Lines
@section @command{awk} Statements Versus Lines
@cindex line breaks
@cindex newlines

Most often, each line in an @command{awk} program is a separate statement or
separate rule, like this:

@example
awk '/12/  @{ print $0 @}
     /21/  @{ print $0 @}' mail-list inventory-shipped
@end example

@cindex @command{gawk}, newlines in
However, @command{gawk} ignores newlines after any of the following
symbols and keywords:

@example
,    @{    ?    :    ||    &&    do    else
@end example

@noindent
A newline at any other point is considered the end of the
statement.@footnote{The @samp{?} and @samp{:} referred to here is the
three-operand conditional expression described in
@ref{Conditional Exp}.
Splitting lines after @samp{?} and @samp{:} is a minor @command{gawk}
extension; if @option{--posix} is specified
(@pxref{Options}), then this extension is disabled.}

@cindex @code{\} (backslash), continuing lines and
@cindex backslash (@code{\}), continuing lines and
If you would like to split a single statement into two lines at a point
where a newline would terminate it, you can @dfn{continue} it by ending the
first line with a backslash character (@samp{\}).  The backslash must be
the final character on the line in order to be recognized as a continuation
character.  A backslash is allowed anywhere in the statement, even
in the middle of a string or regular expression.  For example:

@example
awk '/This regular expression is too long, so continue it\
 on the next line/ @{ print $1 @}'
@end example

@noindent
@cindex portability, backslash continuation and
We have generally not used backslash continuation in our sample programs.
@command{gawk} places no limit on the
length of a line, so backslash continuation is never strictly necessary;
it just makes programs more readable.  For this same reason, as well as
for clarity, we have kept most statements short in the programs
presented throughout the @value{DOCUMENT}.  Backslash continuation is
most useful when your @command{awk} program is in a separate source file
instead of entered from the command line.  You should also note that
many @command{awk} implementations are more particular about where you
may use backslash continuation. For example, they may not allow you to
split a string constant using backslash continuation.  Thus, for maximum
portability of your @command{awk} programs, it is best not to split your
lines in the middle of a regular expression or a string.
@c 10/2000: gawk, mawk, and current bell labs awk allow it,
@c solaris 2.7 nawk does not. Solaris /usr/xpg4/bin/awk does though!  sigh.

@cindex @command{csh} utility
@cindex backslash (@code{\}), continuing lines and, in @command{csh}
@cindex @code{\} (backslash), continuing lines and, in @command{csh}
@quotation CAUTION
@emph{Backslash continuation does not work as described
with the C shell.}  It works for @command{awk} programs in files and
for one-shot programs, @emph{provided} you are using a POSIX-compliant
shell, such as the Unix Bourne shell or Bash.  But the C shell behaves
differently!  There you must use two backslashes in a row, followed by
a newline.  Note also that when using the C shell, @emph{every} newline
in your @command{awk} program must be escaped with a backslash. To illustrate:

@example
% @kbd{awk 'BEGIN @{ \}
? @kbd{  print \\}
? @kbd{      "hello, world" \}
? @kbd{@}'}
@print{} hello, world
@end example

@noindent
Here, the @samp{%} and @samp{?} are the C shell's primary and secondary
prompts, analogous to the standard shell's @samp{$} and @samp{>}.

Compare the previous example to how it is done with a POSIX-compliant shell:

@example
$ @kbd{awk 'BEGIN @{}
>   @kbd{print \}
>       @kbd{"hello, world"}
> @kbd{@}'}
@print{} hello, world
@end example
@end quotation

@command{awk} is a line-oriented language.  Each rule's action has to
begin on the same line as the pattern.  To have the pattern and action
on separate lines, you @emph{must} use backslash continuation; there
is no other option.

@cindex backslash (@code{\}), continuing lines and, comments and
@cindex @code{\} (backslash), continuing lines and, comments and
@cindex commenting, backslash continuation and
Another thing to keep in mind is that backslash continuation and
comments do not mix. As soon as @command{awk} sees the @samp{#} that
starts a comment, it ignores @emph{everything} on the rest of the
line. For example:

@example
@group
$ @kbd{gawk 'BEGIN @{ print "dont panic" # a friendly \}
> @kbd{                                   BEGIN rule}
> @kbd{@}'}
@error{} gawk: cmd. line:2:                BEGIN rule
@error{} gawk: cmd. line:2:                ^ syntax error
@end group
@end example

@noindent
In this case, it looks like the backslash would continue the comment onto the
next line. However, the backslash-newline combination is never even
noticed because it is ``hidden'' inside the comment. Thus, the
@code{BEGIN} is noted as a syntax error.

@cindex statements, multiple
@cindex @code{;} (semicolon), separating statements in actions
@cindex semicolon (@code{;}), separating statements in actions
When @command{awk} statements within one rule are short, you might want to put
more than one of them on a line.  This is accomplished by separating the statements
with a semicolon (@samp{;}).
This also applies to the rules themselves.
Thus, the program shown at the start of this @value{SECTION}
could also be written this way:

@example
/12/ @{ print $0 @} ; /21/ @{ print $0 @}
@end example

@quotation NOTE
The requirement that states that rules on the same line must be
separated with a semicolon was not in the original @command{awk}
language; it was added for consistency with the treatment of statements
within an action.
@end quotation

@node Other Features
@section Other Features of @command{awk}

@cindex variables
The @command{awk} language provides a number of predefined, or
@dfn{built-in}, variables that your programs can use to get information
from @command{awk}.  There are other variables your program can set
as well to control how @command{awk} processes your data.

In addition, @command{awk} provides a number of built-in functions for doing
common computational and string-related operations.
@command{gawk} provides built-in functions for working with timestamps,
performing bit manipulation, for runtime string translation (internationalization),
determining the type of a variable,
and array sorting.

As we develop our presentation of the @command{awk} language, we will introduce
most of the variables and many of the functions. They are described
systematically in @ref{Built-in Variables} and in
@ref{Built-in}.

@node When
@section When to Use @command{awk}

@cindex @command{awk}, uses for
Now that you've seen some of what @command{awk} can do,
you might wonder how @command{awk} could be useful for you.  By using
utility programs, advanced patterns, field separators, arithmetic
statements, and other selection criteria, you can produce much more
complex output.  The @command{awk} language is very useful for producing
reports from large amounts of raw data, such as summarizing information
from the output of other utility programs like @command{ls}.
(@xref{More Complex}.)

Programs written with @command{awk} are usually much smaller than they would
be in other languages.  This makes @command{awk} programs easy to compose and
use.  Often, @command{awk} programs can be quickly composed at your keyboard,
used once, and thrown away.  Because @command{awk} programs are interpreted, you
can avoid the (usually lengthy) compilation part of the typical
edit-compile-test-debug cycle of software development.

@cindex Brian Kernighan's @command{awk}
Complex programs have been written in @command{awk}, including a complete
retargetable assembler for
@ifclear FOR_PRINT
eight-bit microprocessors (@pxref{Glossary}, for more information),
@end ifclear
@ifset FOR_PRINT
eight-bit microprocessors,
@end ifset
and a microcode assembler for a special-purpose Prolog
computer.
The original @command{awk}'s capabilities were strained by tasks
of such complexity, but modern versions are more capable.

@cindex @command{awk} programs, complex
If you find yourself writing @command{awk} scripts of more than, say,
a few hundred lines, you might consider using a different programming
language.  The shell is good at string and pattern matching; in addition,
it allows powerful use of the system utilities.  Python offers a nice
balance between high-level ease of programming and access to system
facilities.@footnote{Other popular scripting languages include Ruby
and Perl.}

@node Intro Summary
@section Summary

@c FIXME: Review this chapter for summary of builtin functions called.
@itemize @value{BULLET}
@item
Programs in @command{awk} consist of @var{pattern}--@var{action} pairs.

@item
An @var{action} without a @var{pattern} always runs.  The default
@var{action} for a pattern without one is @samp{@{ print $0 @}}.

@item
Use either
@samp{awk '@var{program}' @var{files}}
or
@samp{awk -f @var{program-file} @var{files}}
to run @command{awk}.

@item
You may use the special @samp{#!} header line to create @command{awk}
programs that are directly executable.

@item
Comments in @command{awk} programs start with @samp{#} and continue to
the end of the same line.

@item
Be aware of quoting issues when writing @command{awk} programs as
part of a larger shell script (or MS-Windows batch file).

@item
You may use backslash continuation to continue a source line.
Lines are automatically continued after
a comma, open brace, question mark, colon,
@samp{||}, @samp{&&}, @code{do}, and @code{else}.
@end itemize

@node Invoking Gawk
@chapter Running @command{awk} and @command{gawk}

This @value{CHAPTER} covers how to run @command{awk}, both POSIX-standard
and @command{gawk}-specific command-line options, and what
@command{awk} and
@command{gawk} do with nonoption arguments.
It then proceeds to cover how @command{gawk} searches for source files,
reading standard input along with other files, @command{gawk}'s
environment variables, @command{gawk}'s exit status, using include files,
and obsolete and undocumented options and/or features.

Many of the options and features described here are discussed in
more detail later in the @value{DOCUMENT}; feel free to skip over
things in this @value{CHAPTER} that don't interest you right now.

@menu
* Command Line::                How to run @command{awk}.
* Options::                     Command-line options and their meanings.
* Other Arguments::             Input file names and variable assignments.
* Naming Standard Input::       How to specify standard input with other
                                files.
* Environment Variables::       The environment variables @command{gawk} uses.
* Exit Status::                 @command{gawk}'s exit status.
* Include Files::               Including other files into your program.
* Loading Shared Libraries::    Loading shared libraries into your program.
* Obsolete::                    Obsolete Options and/or features.
* Undocumented::                Undocumented Options and Features.
* Invoking Summary::            Invocation summary.
@end menu

@node Command Line
@section Invoking @command{awk}
@cindex command line, invoking @command{awk} from
@cindex @command{awk}, invoking
@cindex arguments, command-line, invoking @command{awk}
@cindex options, command-line, invoking @command{awk}

There are two ways to run @command{awk}---with an explicit program or with
one or more program files.  Here are templates for both of them; items
enclosed in [@dots{}] in these templates are optional:

@display
@command{awk} [@var{options}] @option{-f} @var{progfile} [@option{--}] @var{file} @dots{}
@command{awk} [@var{options}] [@option{--}] @code{'@var{program}'} @var{file} @dots{}
@end display

@cindex GNU long options
@cindex long options
@cindex options, long
In addition to traditional one-letter POSIX-style options, @command{gawk} also
supports GNU long options.

@cindex dark corner, invoking @command{awk}
@cindex lint checking, empty programs
It is possible to invoke @command{awk} with an empty program:

@example
awk '' datafile1 datafile2
@end example

@cindex @option{--lint} option
@cindex dark corner, empty programs
@noindent
Doing so makes little sense, though; @command{awk} exits
silently when given an empty program.
@value{DARKCORNER}
If @option{--lint} has
been specified on the command line, @command{gawk} issues a
warning that the program is empty.

@node Options
@section Command-Line Options
@cindex options, command-line
@cindex command line, options
@cindex GNU long options
@cindex options, long

Options begin with a dash and consist of a single character.
GNU-style long options consist of two dashes and a keyword.
The keyword can be abbreviated, as long as the abbreviation allows the option
to be uniquely identified.  If the option takes an argument, either the
keyword is immediately followed by an equals sign (@samp{=}) and the
argument's value, or the keyword and the argument's value are separated
by whitespace.
If a particular option with a value is given more than once, it is the
last value that counts.

@cindex POSIX @command{awk}, GNU long options and
Each long option for @command{gawk} has a corresponding
POSIX-style short option.
The long and short options are
interchangeable in all contexts.
The following list describes options mandated by the POSIX standard:

@table @code
@item -F @var{fs}
@itemx --field-separator @var{fs}
@cindex @option{-F} option
@cindex @option{--field-separator} option
@cindex @code{FS} variable, @code{--field-separator} option and
Set the @code{FS} variable to @var{fs}
(@pxref{Field Separators}).

@item -f @var{source-file}
@itemx --file @var{source-file}
@cindex @option{-f} option
@cindex @option{--file} option
@cindex @command{awk} programs, location of
Read the @command{awk} program source from @var{source-file}
instead of in the first nonoption argument.
This option may be given multiple times; the @command{awk}
program consists of the concatenation of the contents of
each specified @var{source-file}.

@item -v @var{var}=@var{val}
@itemx --assign @var{var}=@var{val}
@cindex @option{-v} option
@cindex @option{--assign} option
@cindex variables, setting
Set the variable @var{var} to the value @var{val} @emph{before}
execution of the program begins.  Such variable values are available
inside the @code{BEGIN} rule
(@pxref{Other Arguments}).

The @option{-v} option can only set one variable, but it can be used
more than once, setting another variable each time, like this:
@samp{awk @w{-v foo=1} @w{-v bar=2} @dots{}}.

@cindex predefined variables, @code{-v} option@comma{} setting with
@cindex variables, predefined @code{-v} option@comma{} setting with
@quotation CAUTION
Using @option{-v} to set the values of the built-in
variables may lead to surprising results.  @command{awk} will reset the
values of those variables as it needs to, possibly ignoring any
initial value you may have given.
@end quotation

@item -W @var{gawk-opt}
@cindex @option{-W} option
Provide an implementation-specific option.
This is the POSIX convention for providing implementation-specific options.
These options
also have corresponding GNU-style long options.
Note that the long options may be abbreviated, as long as
the abbreviations remain unique.
The full list of @command{gawk}-specific options is provided next.

@item --
@cindex command line, options, end of
@cindex options, command-line, end of
Signal the end of the command-line options.  The following arguments
are not treated as options even if they begin with @samp{-}.  This
interpretation of @option{--} follows the POSIX argument parsing
conventions.

@cindex @code{-} (hyphen), filenames beginning with
@cindex hyphen (@code{-}), filenames beginning with
This is useful if you have @value{FN}s that start with @samp{-},
or in shell scripts, if you have @value{FN}s that will be specified
by the user that could start with @samp{-}.
It is also useful for passing options on to the @command{awk}
program; see @ref{Getopt Function}.
@end table

The following list describes @command{gawk}-specific options:

@c Have to use @asis here to get docbook to come out right.
@table @asis
@item @option{-b}
@itemx @option{--characters-as-bytes}
@cindex @option{-b} option
@cindex @option{--characters-as-bytes} option
Cause @command{gawk} to treat all input data as single-byte characters.
In addition, all output written with @code{print} or @code{printf}
is treated as single-byte characters.

Normally, @command{gawk} follows the POSIX standard and attempts to process
its input data according to the current locale (@pxref{Locales}). This can often involve
converting multibyte characters into wide characters (internally), and
can lead to problems or confusion if the input data does not contain valid
multibyte characters. This option is an easy way to tell @command{gawk},
``Hands off my data!''

@item @option{-c}
@itemx @option{--traditional}
@cindex @option{-c} option
@cindex @option{--traditional} option
@cindex compatibility mode (@command{gawk}), specifying
Specify @dfn{compatibility mode}, in which the GNU extensions to
the @command{awk} language are disabled, so that @command{gawk} behaves just
like BWK @command{awk}.
@xref{POSIX/GNU},
which summarizes the extensions.
@ifclear FOR_PRINT
Also see
@ref{Compatibility Mode}.
@end ifclear

@item @option{-C}
@itemx @option{--copyright}
@cindex @option{-C} option
@cindex @option{--copyright} option
@cindex GPL (General Public License), printing
Print the short version of the General Public License and then exit.

@item @option{-d}[@var{file}]
@itemx @option{--dump-variables}[@code{=}@var{file}]
@cindex @option{-d} option
@cindex @option{--dump-variables} option
@cindex dump all variables of a program
@cindex @file{awkvars.out} file
@cindex files, @file{awkvars.out}
@cindex variables, global, printing list of
Print a sorted list of global variables, their types, and final values
to @var{file}.  If no @var{file} is provided, print this
list to a file named @file{awkvars.out} in the current directory.
No space is allowed between the @option{-d} and @var{file}, if
@var{file} is supplied.

@cindex troubleshooting, typographical errors@comma{} global variables
Having a list of all global variables is a good way to look for
typographical errors in your programs.
You would also use this option if you have a large program with a lot of
functions, and you want to be sure that your functions don't
inadvertently use global variables that you meant to be local.
(This is a particularly easy mistake to make with simple variable
names like @code{i}, @code{j}, etc.)

@item @option{-D}[@var{file}]
@itemx @option{--debug}[@code{=}@var{file}]
@cindex @option{-D} option
@cindex @option{--debug} option
@cindex @command{awk} debugging, enabling
Enable debugging of @command{awk} programs
(@pxref{Debugging}).
By default, the debugger reads commands interactively from the keyboard
(standard input).
The optional @var{file} argument allows you to specify a file with a list
of commands for the debugger to execute noninteractively.
No space is allowed between the @option{-D} and @var{file}, if
@var{file} is supplied.

@item @option{-e} @var{program-text}
@itemx @option{--source} @var{program-text}
@cindex @option{-e} option
@cindex @option{--source} option
@cindex source code, mixing
Provide program source code in the @var{program-text}.
This option allows you to mix source code in files with source
code that you enter on the command line.
This is particularly useful
when you have library functions that you want to use from your command-line
programs (@pxref{AWKPATH Variable}).

Note that @command{gawk} treats each string as if it ended with
a newline character (even if it doesn't). This makes building
the total program easier.

@quotation CAUTION
At the moment, there is no requirement that each @var{program-text}
be a full syntactic unit. I.e., the following currently works:

@example
$ @kbd{gawk -e 'BEGIN @{ a = 5 ;' -e 'print a @}'}
@print{} 5
@end example

@noindent
However, this could change in the future, so it's not a
good idea to rely upon this feature.
@end quotation

@item @option{-E} @var{file}
@itemx @option{--exec} @var{file}
@cindex @option{-E} option
@cindex @option{--exec} option
@cindex @command{awk} programs, location of
@cindex CGI, @command{awk} scripts for
Similar to @option{-f}, read @command{awk} program text from @var{file}.
There are two differences from @option{-f}:

@itemize @value{BULLET}
@item
This option terminates option processing; anything
else on the command line is passed on directly to the @command{awk} program.

@item
Command-line variable assignments of the form
@samp{@var{var}=@var{value}} are disallowed.
@end itemize

This option is particularly necessary for World Wide Web CGI applications
that pass arguments through the URL; using this option prevents a malicious
(or other) user from passing in options, assignments, or @command{awk} source
code (via @option{-e}) to the CGI application.@footnote{For more detail,
please see Section 4.4 of @uref{http://www.ietf.org/rfc/rfc3875,
RFC 3875}. Also see the
@uref{https://lists.gnu.org/archive/html/bug-gawk/2014-11/msg00022.html,
explanatory note sent to the @command{gawk} bug
mailing list}.}
This option should be used
with @samp{#!} scripts (@pxref{Executable Scripts}), like so:

@example
#! /usr/local/bin/gawk -E

@var{awk program here @dots{}}
@end example

@item @option{-g}
@itemx @option{--gen-pot}
@cindex @option{-g} option
@cindex @option{--gen-pot} option
@cindex portable object files, generating
@cindex files, portable object, generating
Analyze the source program and
generate a GNU @command{gettext} portable object template file on standard
output for all string constants that have been marked for translation.
@xref{Internationalization},
for information about this option.

@item @option{-h}
@itemx @option{--help}
@cindex @option{-h} option
@cindex @option{--help} option
@cindex GNU long options, printing list of
@cindex options, printing list of
@cindex printing, list of options
Print a ``usage'' message summarizing the short- and long-style options
that @command{gawk} accepts and then exit.

@item @option{-i} @var{source-file}
@itemx @option{--include} @var{source-file}
@cindex @option{-i} option
@cindex @option{--include} option
@cindex @command{awk} programs, location of
Read an @command{awk} source library from @var{source-file}.  This option
is completely equivalent to using the @code{@@include} directive inside
your program.  It is very similar to the @option{-f} option,
but there are two important differences.  First, when @option{-i} is
used, the program source is not loaded if it has been previously
loaded, whereas with @option{-f}, @command{gawk} always loads the file.
Second, because this option is intended to be used with code libraries,
@command{gawk} does not recognize such files as constituting main program
input.  Thus, after processing an @option{-i} argument, @command{gawk}
still expects to find the main source code via the @option{-f} option
or on the command line.

@item @option{-l} @var{ext}
@itemx @option{--load} @var{ext}
@cindex @option{-l} option
@cindex @option{--load} option
@cindex loading, extensions
Load a dynamic extension named @var{ext}. Extensions
are stored as system shared libraries.
This option searches for the library using the @env{AWKLIBPATH}
environment variable.  The correct library suffix for your platform will be
supplied by default, so it need not be specified in the extension name.
The extension initialization routine should be named @code{dl_load()}.
An alternative is to use the @code{@@load} keyword inside the program to load
a shared library.  This advanced feature is described in detail in @ref{Dynamic Extensions}.

@item @option{-L}[@var{value}]
@itemx @option{--lint}[@code{=}@var{value}]
@cindex @option{-l} option
@cindex @option{--lint} option
@cindex lint checking, issuing warnings
@cindex warnings, issuing
Warn about constructs that are dubious or nonportable to
other @command{awk} implementations.
No space is allowed between the @option{-L} and @var{value}, if
@var{value} is supplied.
Some warnings are issued when @command{gawk} first reads your program.  Others
are issued at runtime, as your program executes.
With an optional argument of @samp{fatal},
lint warnings become fatal errors.
This may be drastic, but its use will certainly encourage the
development of cleaner @command{awk} programs.
With an optional argument of @samp{invalid}, only warnings about things
that are actually invalid are issued. (This is not fully implemented yet.)

Some warnings are only printed once, even if the dubious constructs they
warn about occur multiple times in your @command{awk} program.  Thus,
when eliminating problems pointed out by @option{--lint}, you should take
care to search for all occurrences of each inappropriate construct. As
@command{awk} programs are usually short, doing so is not burdensome.

@item @option{-M}
@itemx @option{--bignum}
@cindex @option{-M} option
@cindex @option{--bignum} option
Select arbitrary-precision arithmetic on numbers. This option has no effect
if @command{gawk} is not compiled to use the GNU MPFR and MP libraries
(@pxref{Arbitrary Precision Arithmetic}).

@item @option{-n}
@itemx @option{--non-decimal-data}
@cindex @option{-n} option
@cindex @option{--non-decimal-data} option
@cindex hexadecimal values@comma{} enabling interpretation of
@cindex octal values@comma{} enabling interpretation of
@cindex troubleshooting, @code{--non-decimal-data} option
Enable automatic interpretation of octal and hexadecimal
values in input data
(@pxref{Nondecimal Data}).

@quotation CAUTION
This option can severely break old programs.  Use with care.  Also note
that this option may disappear in a future version of @command{gawk}.
@end quotation

@item @option{-N}
@itemx @option{--use-lc-numeric}
@cindex @option{-N} option
@cindex @option{--use-lc-numeric} option
Force the use of the locale's decimal point character
when parsing numeric input data (@pxref{Locales}).

@item @option{-o}[@var{file}]
@itemx @option{--pretty-print}[@code{=}@var{file}]
@cindex @option{-o} option
@cindex @option{--pretty-print} option
Enable pretty-printing of @command{awk} programs.
Implies @option{--no-optimize}.
By default, the output program is created in a file named @file{awkprof.out}
(@pxref{Profiling}).
The optional @var{file} argument allows you to specify a different
@value{FN} for the output.
No space is allowed between the @option{-o} and @var{file}, if
@var{file} is supplied.

@quotation NOTE
In the past, this option would also execute your program.
This is no longer the case.
@end quotation

@item @option{-O}
@itemx @option{--optimize}
@cindex @option{--optimize} option
@cindex @option{-O} option
Enable @command{gawk}'s default optimizations on the internal
representation of the program.  At the moment, this includes simple
constant folding and tail recursion elimination in function calls.

These optimizations are enabled by default.
This option remains primarily for backwards compatibility. However, it may
be used to cancel the effect of an earlier @option{-s} option
(see later in this list).

@item @option{-p}[@var{file}]
@itemx @option{--profile}[@code{=}@var{file}]
@cindex @option{-p} option
@cindex @option{--profile} option
@cindex @command{awk} profiling, enabling
Enable profiling of @command{awk} programs
(@pxref{Profiling}).
Implies @option{--no-optimize}.
By default, profiles are created in a file named @file{awkprof.out}.
The optional @var{file} argument allows you to specify a different
@value{FN} for the profile file.
No space is allowed between the @option{-p} and @var{file}, if
@var{file} is supplied.

The profile contains execution counts for each statement in the program
in the left margin, and function call counts for each function.

@item @option{-P}
@itemx @option{--posix}
@cindex @option{-P} option
@cindex @option{--posix} option
@cindex POSIX mode
@cindex @command{gawk}, extensions@comma{} disabling
Operate in strict POSIX mode.  This disables all @command{gawk}
extensions (just like @option{--traditional}) and
disables all extensions not allowed by POSIX.
@xref{Common Extensions} for a summary of the extensions
in @command{gawk} that are disabled by this option.
Also,
the following additional
restrictions apply:

@itemize @value{BULLET}

@cindex newlines
@cindex whitespace, newlines as
@item
Newlines are not allowed after @samp{?} or @samp{:}
(@pxref{Conditional Exp}).


@cindex @code{FS} variable, as TAB character
@item
Specifying @samp{-Ft} on the command line does not set the value
of @code{FS} to be a single TAB character
(@pxref{Field Separators}).

@cindex locale decimal point character
@cindex decimal point character, locale specific
@item
The locale's decimal point character is used for parsing input
data (@pxref{Locales}).
@end itemize

@c @cindex automatic warnings
@c @cindex warnings, automatic
@cindex @option{--traditional} option, @code{--posix} option and
@cindex @option{--posix} option, @code{--traditional} option and
If you supply both @option{--traditional} and @option{--posix} on the
command line, @option{--posix} takes precedence. @command{gawk}
issues a warning if both options are supplied.

@item @option{-r}
@itemx @option{--re-interval}
@cindex @option{-r} option
@cindex @option{--re-interval} option
@cindex regular expressions, interval expressions and
Allow interval expressions
(@pxref{Regexp Operators})
in regexps.
This is now @command{gawk}'s default behavior.
Nevertheless, this option remains (both for backward compatibility
and for use in combination with @option{--traditional}).

@item @option{-s}
@itemx @option{--no-optimize}
@cindex @option{--no-optimize} option
@cindex @option{-s} option
Disable @command{gawk}'s default optimizations on the internal
representation of the program.

@item @option{-S}
@itemx @option{--sandbox}
@cindex @option{-S} option
@cindex @option{--sandbox} option
@cindex sandbox mode
Disable the @code{system()} function,
input redirections with @code{getline},
output redirections with @code{print} and @code{printf},
and dynamic extensions.
This is particularly useful when you want to run @command{awk} scripts
from questionable sources and need to make sure the scripts
can't access your system (other than the specified input @value{DF}).

@item @option{-t}
@itemx @option{--lint-old}
@cindex @option{-L} option
@cindex @option{--lint-old} option
Warn about constructs that are not available in the original version of
@command{awk} from Version 7 Unix
(@pxref{V7/SVR3.1}).

@item @option{-V}
@itemx @option{--version}
@cindex @option{-V} option
@cindex @option{--version} option
@cindex @command{gawk}, versions of, information about@comma{} printing
Print version information for this particular copy of @command{gawk}.
This allows you to determine if your copy of @command{gawk} is up to date
with respect to whatever the Free Software Foundation is currently
distributing.
It is also useful for bug reports
(@pxref{Bugs}).
@end table

As long as program text has been supplied,
any other options are flagged as invalid with a warning message but
are otherwise ignored.

@cindex @option{-F} option, @option{-Ft} sets @code{FS} to TAB
In compatibility mode, as a special case, if the value of @var{fs} supplied
to the @option{-F} option is @samp{t}, then @code{FS} is set to the TAB
character (@code{"\t"}).  This is true only for @option{--traditional} and not
for @option{--posix}
(@pxref{Field Separators}).

@cindex @option{-f} option, multiple uses
The @option{-f} option may be used more than once on the command line.
If it is, @command{awk} reads its program source from all of the named files, as
if they had been concatenated together into one big file.  This is
useful for creating libraries of @command{awk} functions.  These functions
can be written once and then retrieved from a standard place, instead
of having to be included in each individual program.
The @option{-i} option is similar in this regard.
(As mentioned in
@ref{Definition Syntax},
function names must be unique.)

With standard @command{awk}, library functions can still be used, even
if the program is entered at the keyboard,
by specifying @samp{-f /dev/tty}.  After typing your program,
type @kbd{Ctrl-d} (the end-of-file character) to terminate it.
(You may also use @samp{-f -} to read program source from the standard
input, but then you will not be able to also use the standard input as a
source of data.)

Because it is clumsy using the standard @command{awk} mechanisms to mix
source file and command-line @command{awk} programs, @command{gawk}
provides the @option{-e} option.  This does not require you to
preempt the standard input for your source code; it allows you to easily
mix command-line and library source code (@pxref{AWKPATH Variable}).
As with @option{-f}, the @option{-e} and @option{-i}
options may also be used multiple times on the command line.

@cindex @option{-e} option
If no @option{-f} or @option{-e} option is specified, then @command{gawk}
uses the first nonoption command-line argument as the text of the
program source code.

@cindex @env{POSIXLY_CORRECT} environment variable
@cindex lint checking, @env{POSIXLY_CORRECT} environment variable
@cindex POSIX mode
If the environment variable @env{POSIXLY_CORRECT} exists,
then @command{gawk} behaves in strict POSIX mode, exactly as if
you had supplied @option{--posix}.
Many GNU programs look for this environment variable to suppress
extensions that conflict with POSIX, but @command{gawk} behaves
differently: it suppresses all extensions, even those that do not
conflict with POSIX, and behaves in
strict POSIX mode. If @option{--lint} is supplied on the command line
and @command{gawk} turns on POSIX mode because of @env{POSIXLY_CORRECT},
then it issues a warning message indicating that POSIX
mode is in effect.
You would typically set this variable in your shell's startup file.
For a Bourne-compatible shell (such as Bash), you would add these
lines to the @file{.profile} file in your home directory:

@example
POSIXLY_CORRECT=true
export POSIXLY_CORRECT
@end example

@cindex @command{csh} utility, @env{POSIXLY_CORRECT} environment variable
For a C shell-compatible
shell,@footnote{Not recommended.}
you would add this line to the @file{.login} file in your home directory:

@example
setenv POSIXLY_CORRECT true
@end example

@cindex portability, @env{POSIXLY_CORRECT} environment variable
Having @env{POSIXLY_CORRECT} set is not recommended for daily use,
but it is good for testing the portability of your programs to other
environments.

@node Other Arguments
@section Other Command-Line Arguments
@cindex command line, arguments
@cindex arguments, command-line

Any additional arguments on the command line are normally treated as
input files to be processed in the order specified.   However, an
argument that has the form @code{@var{var}=@var{value}}, assigns
the value @var{value} to the variable @var{var}---it does not specify a
file at all.  (See @ref{Assignment Options}.) In the following example,
@var{count=1} is a variable assignment, not a @value{FN}:

@example
awk -f program.awk file1 count=1 file2
@end example

@cindex @command{gawk}, @code{ARGIND} variable in
@cindex @code{ARGIND} variable, command-line arguments
@cindex @code{ARGV} array, indexing into
@cindex @code{ARGC}/@code{ARGV} variables, command-line arguments
All the command-line arguments are made available to your @command{awk} program in the
@code{ARGV} array (@pxref{Built-in Variables}).  Command-line options
and the program text (if present) are omitted from @code{ARGV}.
All other arguments, including variable assignments, are
included.   As each element of @code{ARGV} is processed, @command{gawk}
sets @code{ARGIND} to the index in @code{ARGV} of the
current element.

@c FIXME: One day, move the ARGC and ARGV node closer to here.
Changing @code{ARGC} and @code{ARGV} in your @command{awk} program lets
you control how @command{awk} processes the input files; this is described
in more detail in @ref{ARGC and ARGV}.

@cindex input files, variable assignments and
@cindex variable assignments and input files
The distinction between @value{FN} arguments and variable-assignment
arguments is made when @command{awk} is about to open the next input file.
At that point in execution, it checks the @value{FN} to see whether
it is really a variable assignment; if so, @command{awk} sets the variable
instead of reading a file.

Therefore, the variables actually receive the given values after all
previously specified files have been read.  In particular, the values of
variables assigned in this fashion are @emph{not} available inside a
@code{BEGIN} rule
(@pxref{BEGIN/END}),
because such rules are run before @command{awk} begins scanning the argument list.

@cindex dark corner, escape sequences
The variable values given on the command line are processed for escape
sequences (@pxref{Escape Sequences}).
@value{DARKCORNER}

In some very early implementations of @command{awk}, when a variable assignment
occurred before any @value{FN}s, the assignment would happen @emph{before}
the @code{BEGIN} rule was executed.  @command{awk}'s behavior was thus
inconsistent; some command-line assignments were available inside the
@code{BEGIN} rule, while others were not.  Unfortunately,
some applications came to depend
upon this ``feature.''  When @command{awk} was changed to be more consistent,
the @option{-v} option was added to accommodate applications that depended
upon the old behavior.

The variable assignment feature is most useful for assigning to variables
such as @code{RS}, @code{OFS}, and @code{ORS}, which control input and
output formats, before scanning the @value{DF}s.  It is also useful for
controlling state if multiple passes are needed over a @value{DF}.  For
example:

@cindex files, multiple passes over
@example
awk 'pass == 1  @{ @var{pass 1 stuff} @}
     pass == 2  @{ @var{pass 2 stuff} @}' pass=1 mydata pass=2 mydata
@end example

Given the variable assignment feature, the @option{-F} option for setting
the value of @code{FS} is not
strictly necessary.  It remains for historical compatibility.

@node Naming Standard Input
@section Naming Standard Input

Often, you may wish to read standard input together with other files.
For example, you may wish to read one file, read standard input coming
from a pipe, and then read another file.

The way to name the standard input, with all versions of @command{awk},
is to use a single, standalone minus sign or dash, @samp{-}.  For example:

@example
@var{some_command} | awk -f myprog.awk file1 - file2
@end example

@noindent
Here, @command{awk} first reads @file{file1}, then it reads
the output of @var{some_command}, and finally it reads
@file{file2}.

You may also use @code{"-"} to name standard input when reading
files with @code{getline} (@pxref{Getline/File}).

In addition, @command{gawk} allows you to specify the special
@value{FN} @file{/dev/stdin}, both on the command line and
with @code{getline}.
Some other versions of @command{awk} also support this, but it
is not standard.
(Some operating systems provide a @file{/dev/stdin} file
in the filesystem; however, @command{gawk} always processes
this @value{FN} itself.)

@node Environment Variables
@section The Environment Variables @command{gawk} Uses
@cindex environment variables used by @command{gawk}

A number of environment variables influence how @command{gawk}
behaves.

@menu
* AWKPATH Variable::            Searching directories for @command{awk}
                                programs.
* AWKLIBPATH Variable::         Searching directories for @command{awk} shared
                                libraries.
* Other Environment Variables:: The environment variables.
@end menu

@node AWKPATH Variable
@subsection The @env{AWKPATH} Environment Variable
@cindex @env{AWKPATH} environment variable
@cindex directories, searching for source files
@cindex search paths, for source files
@cindex differences in @command{awk} and @command{gawk}, @env{AWKPATH} environment variable
@ifinfo
The previous @value{SECTION} described how @command{awk} program files can be named
on the command line with the @option{-f} option.
@end ifinfo
In most @command{awk}
implementations, you must supply a precise pathname for each program
file, unless the file is in the current directory.
But with @command{gawk}, if the @value{FN} supplied to the @option{-f}
or @option{-i} options
does not contain a directory separator @samp{/}, then @command{gawk} searches a list of
directories (called the @dfn{search path}) one by one, looking for a
file with the specified name.

The search path is a string consisting of directory names
separated by colons.@footnote{Semicolons on MS-Windows.}
@command{gawk} gets its search path from the
@env{AWKPATH} environment variable.  If that variable does not exist,
or if it has an empty value,
@command{gawk} uses a default path (described shortly).

The search path feature is particularly helpful for building libraries
of useful @command{awk} functions.  The library files can be placed in a
standard directory in the default path and then specified on
the command line with a short @value{FN}.  Otherwise, you would have to
type the full @value{FN} for each file.

By using the @option{-i} or @option{-f} options, your command-line
@command{awk} programs can use facilities in @command{awk} library files
(@pxref{Library Functions}).
Path searching is not done if @command{gawk} is in compatibility mode.
This is true for both @option{--traditional} and @option{--posix}.
@xref{Options}.

If the source code file is not found after the initial search, the path is searched
again after adding the suffix @samp{.awk} to the @value{FN}.

@command{gawk}'s path search mechanism is similar
to the shell's.
(See @uref{https://www.gnu.org/software/bash/manual/,
@cite{The Bourne-Again SHell manual}}.)
It treats a null entry in the path as indicating the current
directory.
(A null entry is indicated by starting or ending the path with a
colon or by placing two colons next to each other [@samp{::}].)

@quotation NOTE
To include the current directory in the path, either place @file{.}
as an entry in the path or write a null entry in the path.

Different past versions of @command{gawk} would also look explicitly in
the current directory, either before or after the path search.  As of
@value{PVERSION} 4.1.2, this no longer happens; if you wish to look
in the current directory, you must include @file{.} either as a separate
entry or as a null entry in the search path.
@end quotation

The default value for @env{AWKPATH} is
@samp{.:/usr/local/share/awk}.@footnote{Your version of @command{gawk}
may use a different directory; it
will depend upon how @command{gawk} was built and installed. The actual
directory is the value of @code{$(datadir)} generated when
@command{gawk} was configured.  You probably don't need to worry about this,
though.}  Since @file{.} is included at the beginning, @command{gawk}
searches first in the current directory and then in @file{/usr/local/share/awk}.
In practice, this means that you will rarely need to change the
value of @env{AWKPATH}.

@xref{Shell Startup Files}, for information on functions that help to
manipulate the @env{AWKPATH} variable.

@command{gawk} places the value of the search path that it used into
@code{ENVIRON["AWKPATH"]}. This provides access to the actual search
path value from within an @command{awk} program.

Although you can change @code{ENVIRON["AWKPATH"]} within your @command{awk}
program, this has no effect on the running program's behavior.  This makes
sense: the @env{AWKPATH} environment variable is used to find the program
source files.  Once your program is running, all the files have been
found, and @command{gawk} no longer needs to use @env{AWKPATH}.

@node AWKLIBPATH Variable
@subsection The @env{AWKLIBPATH} Environment Variable
@cindex @env{AWKLIBPATH} environment variable
@cindex directories, searching for loadable extensions
@cindex search paths, for loadable extensions
@cindex differences in @command{awk} and @command{gawk}, @code{AWKLIBPATH} environment variable

The @env{AWKLIBPATH} environment variable is similar to the @env{AWKPATH}
variable, but it is used to search for loadable extensions (stored as
system shared libraries) specified with the @option{-l} option rather
than for source files.  If the extension is not found, the path is
searched again after adding the appropriate shared library suffix for
the platform.  For example, on GNU/Linux systems, the suffix @samp{.so}
is used.  The search path specified is also used for extensions loaded
via the @code{@@load} keyword (@pxref{Loading Shared Libraries}).

If @env{AWKLIBPATH} does not exist in the environment, or if it has
an empty value, @command{gawk} uses a default path; this
is typically @samp{/usr/local/lib/gawk}, although it can vary depending
upon how @command{gawk} was built.

@xref{Shell Startup Files}, for information on functions that help to
manipulate the @env{AWKLIBPATH} variable.

@command{gawk} places the value of the search path that it used into
@code{ENVIRON["AWKLIBPATH"]}. This provides access to the actual search
path value from within an @command{awk} program.

Although you can change @code{ENVIRON["AWKLIBPATH"]} within your
@command{awk} program, this has no effect on the running program's
behavior.  This makes sense: the @env{AWKLIBPATH} environment variable
is used to find any requested extensions, and they are loaded before
the program starts to run.  Once your program is running, all the
extensions have been found, and @command{gawk} no longer needs to use
@env{AWKLIBPATH}.

@node Other Environment Variables
@subsection Other Environment Variables

A number of other environment variables affect @command{gawk}'s
behavior, but they are more specialized. Those in the following
list are meant to be used by regular users:

@table @env
@item GAWK_MSEC_SLEEP
Specifies the interval between connection retries,
in milliseconds. On systems that do not support
the @code{usleep()} system call,
the value is rounded up to an integral number of seconds.

@item GAWK_READ_TIMEOUT
Specifies the time, in milliseconds, for @command{gawk} to
wait for input before returning with an error.
@xref{Read Timeout}.

@item GAWK_SOCK_RETRIES
Controls the number of times @command{gawk} attempts to
retry a two-way TCP/IP (socket) connection before giving up.
@xref{TCP/IP Networking}.
Note that when nonfatal I/O is enabled (@pxref{Nonfatal}),
@command{gawk} only tries to open a TCP/IP socket once.

@item POSIXLY_CORRECT
Causes @command{gawk} to switch to POSIX-compatibility
mode, disabling all traditional and GNU extensions.
@xref{Options}.
@end table

The environment variables in the following list are meant
for use by the @command{gawk} developers for testing and tuning.
They are subject to change. The variables are:

@table @env
@item AWKBUFSIZE
This variable only affects @command{gawk} on POSIX-compliant systems.
With a value of @samp{exact}, @command{gawk} uses the size of each input
file as the size of the memory buffer to allocate for I/O. Otherwise,
the value should be a number, and @command{gawk} uses that number as
the size of the buffer to allocate.  (When this variable is not set,
@command{gawk} uses the smaller of the file's size and the ``default''
blocksize, which is usually the filesystem's I/O blocksize.)

@item AWK_HASH
If this variable exists with a value of @samp{gst}, @command{gawk}
switches to using the hash function from GNU Smalltalk for
managing arrays.
This function may be marginally faster than the standard function.

@item AWKREADFUNC
If this variable exists, @command{gawk} switches to reading source
files one line at a time, instead of reading in blocks. This exists
for debugging problems on filesystems on non-POSIX operating systems
where I/O is performed in records, not in blocks.

@item GAWK_MSG_SRC
If this variable exists, @command{gawk} includes the @value{FN}
and line number within the @command{gawk} source code
from which warning and/or fatal messages
are generated.  Its purpose is to help isolate the source of a
message, as there are multiple places that produce the
same warning or error message.

@item GAWK_LOCALE_DIR
Specifies the location of compiled message object files
for @command{gawk} itself. This is passed to the @code{bindtextdomain()}
function when @command{gawk} starts up.

@item GAWK_NO_DFA
If this variable exists, @command{gawk} does not use the DFA regexp matcher
for ``does it match'' kinds of tests. This can cause @command{gawk}
to be slower. Its purpose is to help isolate differences between the
two regexp matchers that @command{gawk} uses internally. (There aren't
supposed to be differences, but occasionally theory and practice don't
coordinate with each other.)

@item GAWK_STACKSIZE
This specifies the amount by which @command{gawk} should grow its
internal evaluation stack, when needed.

@item INT_CHAIN_MAX
This specifies intended maximum number of items @command{gawk} will maintain on a
hash chain for managing arrays indexed by integers.

@item STR_CHAIN_MAX
This specifies intended maximum number of items @command{gawk} will maintain on a
hash chain for managing arrays indexed by strings.

@item TIDYMEM
If this variable exists, @command{gawk} uses the @code{mtrace()} library
calls from the GNU C library to help track down possible memory leaks.
@end table

@node Exit Status
@section @command{gawk}'s Exit Status

@cindex exit status, of @command{gawk}
If the @code{exit} statement is used with a value
(@pxref{Exit Statement}), then @command{gawk} exits with
the numeric value given to it.

Otherwise, if there were no problems during execution,
@command{gawk} exits with the value of the C constant
@code{EXIT_SUCCESS}.  This is usually zero.

If an error occurs, @command{gawk} exits with the value of
the C constant @code{EXIT_FAILURE}.  This is usually one.

If @command{gawk} exits because of a fatal error, the exit
status is two.  On non-POSIX systems, this value may be mapped
to @code{EXIT_FAILURE}.

@node Include Files
@section Including Other Files into Your Program

@c Panos Papadopoulos <panos1962@gmail.com> contributed the original
@c text for this section.

This @value{SECTION} describes a feature that is specific to @command{gawk}.

@cindex @code{@@include} directive
@cindex file inclusion, @code{@@include} directive
@cindex including files, @code{@@include} directive
The @code{@@include} keyword can be used to read external @command{awk} source
files.  This gives you the ability to split large @command{awk} source files
into smaller, more manageable pieces, and also lets you reuse common @command{awk}
code from various @command{awk} scripts.  In other words, you can group
together @command{awk} functions used to carry out specific tasks
into external files. These files can be used just like function libraries,
using the @code{@@include} keyword in conjunction with the @env{AWKPATH}
environment variable.  Note that source files may also be included
using the @option{-i} option.

Let's see an example.
We'll start with two (trivial) @command{awk} scripts, namely
@file{test1} and @file{test2}. Here is the @file{test1} script:

@example
BEGIN @{
    print "This is script test1."
@}
@end example

@noindent
and here is @file{test2}:

@example
@@include "test1"
BEGIN @{
    print "This is script test2."
@}
@end example

Running @command{gawk} with @file{test2}
produces the following result:

@example
$ @kbd{gawk -f test2}
@print{} This is script test1.
@print{} This is script test2.
@end example

@command{gawk} runs the @file{test2} script, which includes @file{test1}
using the @code{@@include}
keyword.  So, to include external @command{awk} source files, you just
use @code{@@include} followed by the name of the file to be included,
enclosed in double quotes.

@quotation NOTE
Keep in mind that this is a language construct and the @value{FN} cannot
be a string variable, but rather just a literal string constant in double quotes.
@end quotation

The files to be included may be nested; e.g., given a third
script, namely @file{test3}:

@example
@group
@@include "test2"
BEGIN @{
    print "This is script test3."
@}
@end group
@end example

@noindent
Running @command{gawk} with the @file{test3} script produces the
following results:

@example
$ @kbd{gawk -f test3}
@print{} This is script test1.
@print{} This is script test2.
@print{} This is script test3.
@end example

The @value{FN} can, of course, be a pathname. For example:

@example
@@include "../io_funcs"
@end example

@noindent
and:

@example
@@include "/usr/awklib/network"
@end example

@noindent
are both valid. The @env{AWKPATH} environment variable can be of great
value when using @code{@@include}. The same rules for the use
of the @env{AWKPATH} variable in command-line file searches
(@pxref{AWKPATH Variable}) apply to
@code{@@include} also.

This is very helpful in constructing @command{gawk} function libraries.
If you have a large script with useful, general-purpose @command{awk}
functions, you can break it down into library files and put those files
in a special directory.  You can then include those ``libraries,''
either by using the full pathnames of the files, or by setting the @env{AWKPATH}
environment variable accordingly and then using @code{@@include} with
just the file part of the full pathname. Of course,
you can keep library files in more than one directory;
the more complex the working
environment is, the more directories you may need to organize the files
to be included.

Given the ability to specify multiple @option{-f} options, the
@code{@@include} mechanism is not strictly necessary.
However, the @code{@@include} keyword
can help you in constructing self-contained @command{gawk} programs,
thus reducing the need for writing complex and tedious command lines.
In particular, @code{@@include} is very useful for writing CGI scripts
to be run from web pages.

The rules for finding a source file described in @ref{AWKPATH Variable} also
apply to files loaded with @code{@@include}.

@node Loading Shared Libraries
@section Loading Dynamic Extensions into Your Program

This @value{SECTION} describes a feature that is specific to @command{gawk}.

@cindex @code{@@load} directive
@cindex loading extensions, @code{@@load} directive
@cindex extensions, loading, @code{@@load} directive
The @code{@@load} keyword can be used to read external @command{awk} extensions
(stored as system shared libraries).
This allows you to link in compiled code that may offer superior
performance and/or give you access to extended capabilities not supported
by the @command{awk} language.  The @env{AWKLIBPATH} variable is used to
search for the extension.  Using @code{@@load} is completely equivalent
to using the @option{-l} command-line option.

If the extension is not initially found in @env{AWKLIBPATH}, another
search is conducted after appending the platform's default shared library
suffix to the @value{FN}.  For example, on GNU/Linux systems, the suffix
@samp{.so} is used:

@example
$ @kbd{gawk '@@load "ordchr"; BEGIN @{print chr(65)@}'}
@print{} A
@end example

@noindent
This is equivalent to the following example:

@example
@group
$ @kbd{gawk -lordchr 'BEGIN @{print chr(65)@}'}
@print{} A
@end group
@end example

@noindent
For command-line usage, the @option{-l} option is more convenient,
but @code{@@load} is useful for embedding inside an @command{awk} source file
that requires access to an extension.

@ref{Dynamic Extensions}, describes how to write extensions (in C or C++)
that can be loaded with either @code{@@load} or the @option{-l} option.
It also describes the @code{ordchr} extension.

@node Obsolete
@section Obsolete Options and/or Features

@c update this section for each release!

@cindex options, deprecated
@cindex features, deprecated
@cindex obsolete features
This @value{SECTION} describes features and/or command-line options from
previous releases of @command{gawk} that either are not available in the
current version or are still supported but deprecated (meaning that
they will @emph{not} be in the next release).

The process-related special files @file{/dev/pid}, @file{/dev/ppid},
@file{/dev/pgrpid}, and @file{/dev/user} were deprecated in @command{gawk}
3.1, but still worked.  As of @value{PVERSION} 4.0, they are no longer
interpreted specially by @command{gawk}.  (Use @code{PROCINFO} instead;
see @ref{Auto-set}.)

@ignore
This @value{SECTION}
is thus essentially a place holder,
in case some option becomes obsolete in a future version of @command{gawk}.
@end ignore

@node Undocumented
@section Undocumented Options and Features
@cindex undocumented features
@cindex features, undocumented
@cindex Skywalker, Luke
@cindex Kenobi, Obi-Wan
@cindex Jedi knights
@cindex Knights, jedi
@quotation
@i{Use the Source, Luke!}
@author Obi-Wan
@end quotation

@cindex shells, sea
This @value{SECTION} intentionally left
blank.

@ignore
@c If these came out in the Info file or TeX document, then they wouldn't
@c be undocumented, would they?

@command{gawk} has one undocumented option:

@table @code
@item -W nostalgia
@itemx --nostalgia
Print the message @samp{awk: bailing out near line 1} and dump core.
This option was inspired by the common behavior of very early versions of
Unix @command{awk} and by a t--shirt.
The message is @emph{not} subject to translation in non-English locales.
@c so there! nyah, nyah.
@end table

Early versions of @command{awk} used to not require any separator (either
a newline or @samp{;}) between the rules in @command{awk} programs.  Thus,
it was common to see one-line programs like:

@example
awk '@{ sum += $1 @} END @{ print sum @}'
@end example

@command{gawk} actually supports this but it is purposely undocumented
because it is bad style.  The correct way to write such a program
is either:

@example
awk '@{ sum += $1 @} ; END @{ print sum @}'
@end example

@noindent
or:

@example
awk '@{ sum += $1 @}
     END @{ print sum @}' data
@end example

@noindent
@xref{Statements/Lines}, for a fuller explanation.

You can insert newlines after the @samp{;} in @code{for} loops.
This seems to have been a long-undocumented feature in Unix @command{awk}.

Similarly, you may use @code{print} or @code{printf} statements in the
@var{init} and @var{increment} parts of a @code{for} loop.  This is another
long-undocumented ``feature'' of Unix @command{awk}.

@command{gawk} lets you use the names of built-in functions that are
@command{gawk} extensions as the names of parameters in user-defined functions.
This is intended to ``future-proof'' old code that happens to use
function names added by @command{gawk} after the code was written.
Standard @command{awk} built-in functions, such as @code{sin()} or
@code{substr()} are @emph{not} shadowed in this way.

@end ignore

@node Invoking Summary
@section Summary

@itemize @value{BULLET}
@item
Use either
@samp{awk '@var{program}' @var{files}}
or
@samp{awk -f @var{program-file} @var{files}}
to run @command{awk}.

@item
The three standard options for all versions of @command{awk} are
@option{-f}, @option{-F}, and @option{-v}.  @command{gawk} supplies these
and many others, as well as corresponding GNU-style long options.

@item
Nonoption command-line arguments are usually treated as @value{FN}s,
unless they have the form @samp{@var{var}=@var{value}}, in which case
they are taken as variable assignments to be performed at that point
in processing the input.

@item
All nonoption command-line arguments, excluding the program text,
are placed in the @code{ARGV} array.  Adjusting @code{ARGC} and @code{ARGV}
affects how @command{awk} processes input.

@item
You can use a single minus sign (@samp{-}) to refer to standard input
on the command line. @command{gawk} also lets you use the special
@value{FN} @file{/dev/stdin}.

@item
@command{gawk} pays attention to a number of environment variables.
@env{AWKPATH}, @env{AWKLIBPATH}, and @env{POSIXLY_CORRECT} are the
most important ones.

@item
@command{gawk}'s exit status conveys information to the program
that invoked it. Use the @code{exit} statement from within
an @command{awk} program to set the exit status.

@item
@command{gawk} allows you to include other @command{awk} source files into
your program using the @code{@@include} statement and/or the @option{-i}
and @option{-f} command-line options.

@item
@command{gawk} allows you to load additional functions written in C
or C++ using the @code{@@load} statement and/or the @option{-l} option.
(This advanced feature is described later, in @ref{Dynamic Extensions}.)
@end itemize

@node Regexp
@chapter Regular Expressions
@cindex regexp
@cindex regular expressions

A @dfn{regular expression}, or @dfn{regexp}, is a way of describing a
set of strings.
Because regular expressions are such a fundamental part of @command{awk}
programming, their format and use deserve a separate @value{CHAPTER}.

@cindex forward slash (@code{/}) to enclose regular expressions
@cindex @code{/} (forward slash) to enclose regular expressions
A regular expression enclosed in slashes (@samp{/})
is an @command{awk} pattern that matches every input record whose text
belongs to that set.
The simplest regular expression is a sequence of letters, numbers, or
both.  Such a regexp matches any string that contains that sequence.
Thus, the regexp @samp{foo} matches any string containing @samp{foo}.
Thus, the pattern @code{/foo/} matches any input record containing
the three adjacent characters @samp{foo} @emph{anywhere} in the record.  Other
kinds of regexps let you specify more complicated classes of strings.

@ifnotinfo
Initially, the examples in this @value{CHAPTER} are simple.
As we explain more about how
regular expressions work, we present more complicated instances.
@end ifnotinfo

@menu
* Regexp Usage::                How to Use Regular Expressions.
* Escape Sequences::            How to write nonprinting characters.
* Regexp Operators::            Regular Expression Operators.
* Bracket Expressions::         What can go between @samp{[...]}.
* Leftmost Longest::            How much text matches.
* Computed Regexps::            Using Dynamic Regexps.
* GNU Regexp Operators::        Operators specific to GNU software.
* Case-sensitivity::            How to do case-insensitive matching.
* Regexp Summary::              Regular expressions summary.
@end menu

@node Regexp Usage
@section How to Use Regular Expressions

@cindex regular expressions, as patterns
A regular expression can be used as a pattern by enclosing it in
slashes.  Then the regular expression is tested against the
entire text of each record.  (Normally, it only needs
to match some part of the text in order to succeed.)  For example, the
following prints the second field of each record where the string
@samp{li} appears anywhere in the record:

@example
$ @kbd{awk '/li/ @{ print $2 @}' mail-list}
@print{} 555-5553
@print{} 555-0542
@print{} 555-6699
@print{} 555-3430
@end example

@cindex regular expressions, operators
@cindex operators, string-matching
@c @cindex operators, @code{~}
@cindex string-matching operators
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@c @cindex operators, @code{!~}
@cindex @code{if} statement, use of regexps in
@cindex @code{while} statement, use of regexps in
@cindex @code{do}-@code{while} statement, use of regexps in
@c @cindex statements, @code{if}
@c @cindex statements, @code{while}
@c @cindex statements, @code{do}
Regular expressions can also be used in matching expressions.  These
expressions allow you to specify the string to match against; it need
not be the entire current input record.  The two operators @samp{~}
and @samp{!~} perform regular expression comparisons.  Expressions
using these operators can be used as patterns, or in @code{if},
@code{while}, @code{for}, and @code{do} statements.
(@xref{Statements}.)
For example, the following is true if the expression @var{exp} (taken
as a string) matches @var{regexp}:

@example
@var{exp} ~ /@var{regexp}/
@end example

@noindent
This example matches, or selects, all input records with the uppercase
letter @samp{J} somewhere in the first field:

@example
$ @kbd{awk '$1 ~ /J/' inventory-shipped}
@print{} Jan  13  25  15 115
@print{} Jun  31  42  75 492
@print{} Jul  24  34  67 436
@print{} Jan  21  36  64 620
@end example

So does this:

@example
awk '@{ if ($1 ~ /J/) print @}' inventory-shipped
@end example

This next example is true if the expression @var{exp}
(taken as a character string)
does @emph{not} match @var{regexp}:

@example
@var{exp} !~ /@var{regexp}/
@end example

The following example matches,
or selects, all input records whose first field @emph{does not} contain
the uppercase letter @samp{J}:

@example
$ @kbd{awk '$1 !~ /J/' inventory-shipped}
@print{} Feb  15  32  24 226
@print{} Mar  15  24  34 228
@print{} Apr  31  52  63 420
@print{} May  16  34  29 208
@dots{}
@end example

@cindex regexp constants
@cindex constant regexps
@cindex regular expressions, constants, See regexp constants
When a regexp is enclosed in slashes, such as @code{/foo/}, we call it
a @dfn{regexp constant}, much like @code{5.27} is a numeric constant and
@code{"foo"} is a string constant.

@node Escape Sequences
@section Escape Sequences

@cindex escape sequences, in strings
@cindex backslash (@code{\}), in escape sequences
@cindex @code{\} (backslash), in escape sequences
Some characters cannot be included literally in string constants
(@code{"foo"}) or regexp constants (@code{/foo/}).
Instead, they should be represented with @dfn{escape sequences},
which are character sequences beginning with a backslash (@samp{\}).
One use of an escape sequence is to include a double-quote character in
a string constant.  Because a plain double quote ends the string, you
must use @samp{\"} to represent an actual double-quote character as a
part of the string.  For example:

@example
$ @kbd{awk 'BEGIN @{ print "He said \"hi!\" to her." @}'}
@print{} He said "hi!" to her.
@end example

The  backslash character itself is another character that cannot be
included normally; you must write @samp{\\} to put one backslash in the
string or regexp.  Thus, the string whose contents are the two characters
@samp{"} and @samp{\} must be written @code{"\"\\"}.

Other escape sequences represent unprintable characters
such as TAB or newline.  There is nothing to stop you from entering most
unprintable characters directly in a string constant or regexp constant,
but they may look ugly.

The following list presents
all the escape sequences used in @command{awk} and
what they represent. Unless noted otherwise, all these escape
sequences apply to both string constants and regexp constants:

@table @code
@item \\
A literal backslash, @samp{\}.

@c @cindex @command{awk} language, V.4 version
@cindex @code{\} (backslash), @code{\a} escape sequence
@cindex backslash (@code{\}), @code{\a} escape sequence
@item \a
The ``alert'' character, @kbd{Ctrl-g}, ASCII code 7 (BEL).
(This often makes some sort of audible noise.)

@cindex @code{\} (backslash), @code{\b} escape sequence
@cindex backslash (@code{\}), @code{\b} escape sequence
@item \b
Backspace, @kbd{Ctrl-h}, ASCII code 8 (BS).

@cindex @code{\} (backslash), @code{\f} escape sequence
@cindex backslash (@code{\}), @code{\f} escape sequence
@item \f
Formfeed, @kbd{Ctrl-l}, ASCII code 12 (FF).

@cindex @code{\} (backslash), @code{\n} escape sequence
@cindex backslash (@code{\}), @code{\n} escape sequence
@item \n
Newline, @kbd{Ctrl-j}, ASCII code 10 (LF).

@cindex @code{\} (backslash), @code{\r} escape sequence
@cindex backslash (@code{\}), @code{\r} escape sequence
@item \r
Carriage return, @kbd{Ctrl-m}, ASCII code 13 (CR).

@cindex @code{\} (backslash), @code{\t} escape sequence
@cindex backslash (@code{\}), @code{\t} escape sequence
@item \t
Horizontal TAB, @kbd{Ctrl-i}, ASCII code 9 (HT).

@c @cindex @command{awk} language, V.4 version
@cindex @code{\} (backslash), @code{\v} escape sequence
@cindex backslash (@code{\}), @code{\v} escape sequence
@item \v
Vertical TAB, @kbd{Ctrl-k}, ASCII code 11 (VT).

@cindex @code{\} (backslash), @code{\}@var{nnn} escape sequence
@cindex backslash (@code{\}), @code{\}@var{nnn} escape sequence
@item \@var{nnn}
The octal value @var{nnn}, where @var{nnn} stands for 1 to 3 digits
between @samp{0} and @samp{7}.  For example, the code for the ASCII ESC
(escape) character is @samp{\033}.

@c @cindex @command{awk} language, V.4 version
@c @cindex @command{awk} language, POSIX version
@cindex @code{\} (backslash), @code{\x} escape sequence
@cindex backslash (@code{\}), @code{\x} escape sequence
@cindex common extensions, @code{\x} escape sequence
@cindex extensions, common@comma{} @code{\x} escape sequence
@item \x@var{hh}@dots{}
The hexadecimal value @var{hh}, where @var{hh} stands for a sequence
of hexadecimal digits (@samp{0}--@samp{9}, and either @samp{A}--@samp{F}
or @samp{a}--@samp{f}).  A maximum of two digts are allowed after
the @samp{\x}. Any further hexadecimal digits are treated as simple
letters or numbers.  @value{COMMONEXT}
(The @samp{\x} escape sequence is not allowed in POSIX awk.)

@quotation CAUTION
In ISO C, the escape sequence continues until the first nonhexadecimal
digit is seen.
For many years, @command{gawk} would continue incorporating
hexadecimal digits into the value until a non-hexadecimal digit
or the end of the string was encountered.
However, using more than two hexadecimal digits produced
undefined results.
As of @value{PVERSION} 4.2, only two digits
are processed.
@end quotation

@cindex @code{\} (backslash), @code{\/} escape sequence
@cindex backslash (@code{\}), @code{\/} escape sequence
@item \/
A literal slash (necessary for regexp constants only).
This sequence is used when you want to write a regexp
constant that contains a slash
(such as @code{/.*:\/home\/[[:alnum:]]+:.*/}; the @samp{[[:alnum:]]}
notation is discussed in @ref{Bracket Expressions}).
Because the regexp is delimited by
slashes, you need to escape any slash that is part of the pattern,
in order to tell @command{awk} to keep processing the rest of the regexp.

@cindex @code{\} (backslash), @code{\"} escape sequence
@cindex backslash (@code{\}), @code{\"} escape sequence
@item \"
A literal double quote (necessary for string constants only).
This sequence is used when you want to write a string
constant that contains a double quote
(such as @code{"He said \"hi!\" to her."}).
Because the string is delimited by
double quotes, you need to escape any quote that is part of the string,
in order to tell @command{awk} to keep processing the rest of the string.
@end table

In @command{gawk}, a number of additional two-character sequences that begin
with a backslash have special meaning in regexps.
@xref{GNU Regexp Operators}.

In a regexp, a backslash before any character that is not in the previous list
and not listed in
@ref{GNU Regexp Operators}
means that the next character should be taken literally, even if it would
normally be a regexp operator.  For example, @code{/a\+b/} matches the three
characters @samp{a+b}.

@cindex backslash (@code{\}), in escape sequences
@cindex @code{\} (backslash), in escape sequences
@cindex portability
For complete portability, do not use a backslash before any character not
shown in the previous list or that is not an operator.

@c 11/2014: Moved so as to not stack sidebars
@sidebar Backslash Before Regular Characters
@cindex portability, backslash in escape sequences
@cindex POSIX @command{awk}, backslashes in string constants
@cindex backslash (@code{\}), in escape sequences, POSIX and
@cindex @code{\} (backslash), in escape sequences, POSIX and

@cindex troubleshooting, backslash before nonspecial character
If you place a backslash in a string constant before something that is
not one of the characters previously listed, POSIX @command{awk} purposely
leaves what happens as undefined.  There are two choices:

@c @cindex automatic warnings
@c @cindex warnings, automatic
@cindex Brian Kernighan's @command{awk}
@table @asis
@item Strip the backslash out
This is what BWK @command{awk} and @command{gawk} both do.
For example, @code{"a\qc"} is the same as @code{"aqc"}.
(Because this is such an easy bug both to introduce and to miss,
@command{gawk} warns you about it.)
Consider @samp{FS = @w{"[ \t]+\|[ \t]+"}} to use vertical bars
surrounded by whitespace as the field separator. There should be
two backslashes in the string: @samp{FS = @w{"[ \t]+\\|[ \t]+"}}.)
@c I did this!  This is why I added the warning.

@cindex @command{gawk}, escape sequences
@cindex Unix @command{awk}, backslashes in escape sequences
@cindex @command{mawk} utility
@item Leave the backslash alone
Some other @command{awk} implementations do this.
In such implementations, typing @code{"a\qc"} is the same as typing
@code{"a\\qc"}.
@end table
@end sidebar

To summarize:

@itemize @value{BULLET}
@item
The escape sequences in the preceding list are always processed first,
for both string constants and regexp constants. This happens very early,
as soon as @command{awk} reads your program.

@item
@command{gawk} processes both regexp constants and dynamic regexps
(@pxref{Computed Regexps}),
for the special operators listed in
@ref{GNU Regexp Operators}.

@item
A backslash before any other character means to treat that character
literally.
@end itemize

@sidebar Escape Sequences for Metacharacters
@cindex metacharacters, escape sequences for

Suppose you use an octal or hexadecimal
escape to represent a regexp metacharacter.
(See @ref{Regexp Operators}.)
Does @command{awk} treat the character as a literal character or as a regexp
operator?

@cindex dark corner, escape sequences, for metacharacters
Historically, such characters were taken literally.
@value{DARKCORNER}
However, the POSIX standard indicates that they should be treated
as real metacharacters, which is what @command{gawk} does.
In compatibility mode (@pxref{Options}),
@command{gawk} treats the characters represented by octal and hexadecimal
escape sequences literally when used in regexp constants. Thus,
@code{/a\52b/} is equivalent to @code{/a\*b/}.
@end sidebar

@node Regexp Operators
@section Regular Expression Operators
@cindex regular expressions, operators
@cindex metacharacters in regular expressions

You can combine regular expressions with special characters,
called @dfn{regular expression operators} or @dfn{metacharacters}, to
increase the power and versatility of regular expressions.

The escape sequences described
@ifnotinfo
earlier
@end ifnotinfo
in @ref{Escape Sequences}
are valid inside a regexp.  They are introduced by a @samp{\} and
are recognized and converted into corresponding real characters as
the very first step in processing regexps.

Here is a list of metacharacters.  All characters that are not escape
sequences and that are not listed here stand for themselves:

@c Use @asis so the docbook comes out ok. Sigh.
@table @asis
@cindex backslash (@code{\}), regexp operator
@cindex @code{\} (backslash), regexp operator
@item @code{\}
This suppresses the special meaning of a character when
matching.  For example, @samp{\$}
matches the character @samp{$}.

@cindex regular expressions, anchors in
@cindex Texinfo, chapter beginnings in files
@cindex @code{^} (caret), regexp operator
@cindex caret (@code{^}), regexp operator
@item @code{^}
This matches the beginning of a string.  @samp{^@@chapter}
matches @samp{@@chapter} at the beginning of a string,
for example, and can be used
to identify chapter beginnings in Texinfo source files.
The @samp{^} is known as an @dfn{anchor}, because it anchors the pattern to
match only at the beginning of the string.

It is important to realize that @samp{^} does not match the beginning of
a line (the point right after a @samp{\n} newline character) embedded in a string.
The condition is not true in the following example:

@example
if ("line1\nLINE 2" ~ /^L/) @dots{}
@end example

@cindex @code{$} (dollar sign), regexp operator
@cindex dollar sign (@code{$}), regexp operator
@item @code{$}
This is similar to @samp{^}, but it matches only at the end of a string.
For example, @samp{p$}
matches a record that ends with a @samp{p}.  The @samp{$} is an anchor
and does not match the end of a line
(the point right before a @samp{\n} newline character)
embedded in a string.
The condition in the following example is not true:

@example
if ("line1\nLINE 2" ~ /1$/) @dots{}
@end example

@cindex @code{.} (period), regexp operator
@cindex period (@code{.}), regexp operator
@item @code{.} (period)
This matches any single character,
@emph{including} the newline character.  For example, @samp{.P}
matches any single character followed by a @samp{P} in a string.  Using
concatenation, we can make a regular expression such as @samp{U.A}, which
matches any three-character sequence that begins with @samp{U} and ends
with @samp{A}.

@cindex POSIX @command{awk}, period (@code{.})@comma{} using
In strict POSIX mode (@pxref{Options}),
@samp{.} does not match the @sc{nul}
character, which is a character with all bits equal to zero.
Otherwise, @sc{nul} is just another character. Other versions of @command{awk}
may not be able to match the @sc{nul} character.

@cindex @code{[]} (square brackets), regexp operator
@cindex square brackets (@code{[]}), regexp operator
@cindex bracket expressions
@cindex character sets, See Also bracket expressions
@cindex character lists, See bracket expressions
@cindex character classes, See bracket expressions
@item @code{[}@dots{}@code{]}
This is called a @dfn{bracket expression}.@footnote{In other literature,
you may see a bracket expression referred to as either a
@dfn{character set}, a @dfn{character class}, or a @dfn{character list}.}
It matches any @emph{one} of the characters that are enclosed in
the square brackets.  For example, @samp{[MVX]} matches any one of
the characters @samp{M}, @samp{V}, or @samp{X} in a string.  A full
discussion of what can be inside the square brackets of a bracket expression
is given in
@ref{Bracket Expressions}.

@cindex bracket expressions, complemented
@item @code{[^}@dots{}@code{]}
This is a @dfn{complemented bracket expression}.  The first character after
the @samp{[} @emph{must} be a @samp{^}.  It matches any characters
@emph{except} those in the square brackets.  For example, @samp{[^awk]}
matches any character that is not an @samp{a}, @samp{w},
or @samp{k}.

@cindex @code{|} (vertical bar)
@cindex vertical bar (@code{|})
@item @code{|}
This is the @dfn{alternation operator} and it is used to specify
alternatives.  The @samp{|} has the lowest precedence of all the regular
expression operators.  For example, @samp{^P|[aeiouy]} matches any string
that matches either @samp{^P} or @samp{[aeiouy]}.  This means it matches
any string that starts with @samp{P} or contains (anywhere within it)
a lowercase English vowel.

The alternation applies to the largest possible regexps on either side.

@cindex @code{()} (parentheses), regexp operator
@cindex parentheses @code{()}, regexp operator
@item @code{(}@dots{}@code{)}
Parentheses are used for grouping in regular expressions, as in
arithmetic.  They can be used to concatenate regular expressions
containing the alternation operator, @samp{|}.  For example,
@samp{@@(samp|code)\@{[^@}]+\@}} matches both @samp{@@code@{foo@}} and
@samp{@@samp@{bar@}}.
(These are Texinfo formatting control sequences. The @samp{+} is
explained further on in this list.)

@cindex @code{*} (asterisk), @code{*} operator, as regexp operator
@cindex asterisk (@code{*}), @code{*} operator, as regexp operator
@item @code{*}
This symbol means that the preceding regular expression should be
repeated as many times as necessary to find a match.  For example, @samp{ph*}
applies the @samp{*} symbol to the preceding @samp{h} and looks for matches
of one @samp{p} followed by any number of @samp{h}s.  This also matches
just @samp{p} if no @samp{h}s are present.

There are two subtle points to understand about how @samp{*} works.
First, the @samp{*} applies only to the single preceding regular expression
component (e.g., in @samp{ph*}, it applies just to the @samp{h}).
To cause @samp{*} to apply to a larger subexpression, use parentheses:
@samp{(ph)*} matches @samp{ph}, @samp{phph}, @samp{phphph}, and so on.

Second, @samp{*} finds as many repetitions as possible. If the text
to be matched is @samp{phhhhhhhhhhhhhhooey}, @samp{ph*} matches all of
the @samp{h}s.

@cindex @code{+} (plus sign), regexp operator
@cindex plus sign (@code{+}), regexp operator
@item @code{+}
This symbol is similar to @samp{*}, except that the preceding expression must be
matched at least once.  This means that @samp{wh+y}
would match @samp{why} and @samp{whhy}, but not @samp{wy}, whereas
@samp{wh*y} would match all three.

@cindex @code{?} (question mark), regexp operator
@cindex question mark (@code{?}), regexp operator
@item @code{?}
This symbol is similar to @samp{*}, except that the preceding expression can be
matched either once or not at all.  For example, @samp{fe?d}
matches @samp{fed} and @samp{fd}, but nothing else.

@cindex interval expressions, regexp operator
@item @code{@{}@var{n}@code{@}}
@itemx @code{@{}@var{n}@code{,@}}
@itemx @code{@{}@var{n}@code{,}@var{m}@code{@}}
One or two numbers inside braces denote an @dfn{interval expression}.
If there is one number in the braces, the preceding regexp is repeated
@var{n} times.
If there are two numbers separated by a comma, the preceding regexp is
repeated @var{n} to @var{m} times.
If there is one number followed by a comma, then the preceding regexp
is repeated at least @var{n} times:

@table @code
@item wh@{3@}y
Matches @samp{whhhy}, but not @samp{why} or @samp{whhhhy}.

@item wh@{3,5@}y
Matches @samp{whhhy}, @samp{whhhhy}, or @samp{whhhhhy} only.

@item wh@{2,@}y
Matches @samp{whhy}, @samp{whhhy}, and so on.
@end table

@cindex POSIX @command{awk}, interval expressions in
Interval expressions were not traditionally available in @command{awk}.
They were added as part of the POSIX standard to make @command{awk}
and @command{egrep} consistent with each other.

@cindex @command{gawk}, interval expressions and
Initially, because old programs may use @samp{@{} and @samp{@}} in regexp
constants,
@command{gawk} did @emph{not} match interval expressions
in regexps.

However, beginning with @value{PVERSION} 4.0,
@command{gawk} does match interval expressions by default.
This is because compatibility with POSIX has become more
important to most @command{gawk} users than compatibility with
old programs.

For programs that use @samp{@{} and @samp{@}} in regexp constants,
it is good practice to always escape them with a backslash.  Then the
regexp constants are valid and work the way you want them to, using
any version of @command{awk}.@footnote{Use two backslashes if you're
using a string constant with a regexp operator or function.}

Finally, when @samp{@{} and @samp{@}} appear in regexp constants
in a way that cannot be interpreted as an interval expression
(such as @code{/q@{a@}/}), then they stand for themselves.
@end table

@cindex precedence, regexp operators
@cindex regular expressions, operators, precedence of
In regular expressions, the @samp{*}, @samp{+}, and @samp{?} operators,
as well as the braces @samp{@{} and @samp{@}},
have
the highest precedence, followed by concatenation, and finally by @samp{|}.
As in arithmetic, parentheses can change how operators are grouped.

@cindex POSIX @command{awk}, regular expressions and
@cindex @command{gawk}, regular expressions, precedence
In POSIX @command{awk} and @command{gawk}, the @samp{*}, @samp{+}, and
@samp{?} operators stand for themselves when there is nothing in the
regexp that precedes them.  For example, @code{/+/} matches a literal
plus sign.  However, many other versions of @command{awk} treat such a
usage as a syntax error.

If @command{gawk} is in compatibility mode (@pxref{Options}), interval
expressions are not available in regular expressions.

@node Bracket Expressions
@section Using Bracket Expressions
@cindex bracket expressions
@cindex bracket expressions, range expressions
@cindex range expressions (regexps)
@cindex character lists in regular expression

As mentioned earlier, a bracket expression matches any character among
those listed between the opening and closing square brackets.

Within a bracket expression, a @dfn{range expression} consists of two
characters separated by a hyphen.  It matches any single character that
sorts between the two characters, based upon the system's native character
set.  For example, @samp{[0-9]} is equivalent to @samp{[0123456789]}.
(See @ref{Ranges and Locales} for an explanation of how the POSIX
standard and @command{gawk} have changed over time.  This is mainly
of historical interest.)

With the increasing popularity of the
@uref{http://www.unicode.org, Unicode character standard},
there is an additional wrinkle to consider. Octal and hexadecimal
escape sequences inside bracket expressions are taken to represent
only single-byte characters (characters whose values fit within
the range 0--256).  To match a range of characters where the endpoints
of the range are larger than 256, enter the multibyte encodings of
the characters directly.

@cindex @code{\} (backslash), in bracket expressions
@cindex backslash (@code{\}), in bracket expressions
@cindex @code{^} (caret), in bracket expressions
@cindex caret (@code{^}), in bracket expressions
@cindex @code{-} (hyphen), in bracket expressions
@cindex hyphen (@code{-}), in bracket expressions
To include one of the characters @samp{\}, @samp{]}, @samp{-}, or @samp{^} in a
bracket expression, put a @samp{\} in front of it.  For example:

@example
[d\]]
@end example

@noindent
matches either @samp{d} or @samp{]}.
Additionally, if you place @samp{]} right after the opening
@samp{[}, the closing bracket is treated as one of the
characters to be matched.

@cindex POSIX @command{awk}, bracket expressions and
@cindex Extended Regular Expressions (EREs)
@cindex EREs (Extended Regular Expressions)
@cindex @command{egrep} utility
The treatment of @samp{\} in bracket expressions
is compatible with other @command{awk}
implementations and is also mandated by POSIX.
The regular expressions in @command{awk} are a superset
of the POSIX specification for Extended Regular Expressions (EREs).
POSIX EREs are based on the regular expressions accepted by the
traditional @command{egrep} utility.

@cindex bracket expressions, character classes
@cindex POSIX @command{awk}, bracket expressions and, character classes
@dfn{Character classes} are a feature introduced in the POSIX standard.
A character class is a special notation for describing
lists of characters that have a specific attribute, but the
actual characters can vary from country to country and/or
from character set to character set.  For example, the notion of what
is an alphabetic character differs between the United States and France.

A character class is only valid in a regexp @emph{inside} the
brackets of a bracket expression.  Character classes consist of @samp{[:},
a keyword denoting the class, and @samp{:]}.
@ref{table-char-classes} lists the character classes defined by the
POSIX standard.

@float Table,table-char-classes
@caption{POSIX character classes}
@multitable @columnfractions .15 .85
@headitem Class @tab Meaning
@item @code{[:alnum:]} @tab Alphanumeric characters
@item @code{[:alpha:]} @tab Alphabetic characters
@item @code{[:blank:]} @tab Space and TAB characters
@item @code{[:cntrl:]} @tab Control characters
@item @code{[:digit:]} @tab Numeric characters
@item @code{[:graph:]} @tab Characters that are both printable and visible
(a space is printable but not visible, whereas an @samp{a} is both)
@item @code{[:lower:]} @tab Lowercase alphabetic characters
@item @code{[:print:]} @tab Printable characters (characters that are not control characters)
@item @code{[:punct:]} @tab Punctuation characters (characters that are not letters, digits,
control characters, or space characters)
@item @code{[:space:]} @tab Space characters (such as space, TAB, and formfeed, to name a few)
@item @code{[:upper:]} @tab Uppercase alphabetic characters
@item @code{[:xdigit:]} @tab Characters that are hexadecimal digits
@end multitable
@end float

For example, before the POSIX standard, you had to write @code{/[A-Za-z0-9]/}
to match alphanumeric characters.  If your
character set had other alphabetic characters in it, this would not
match them.
With the POSIX character classes, you can write
@code{/[[:alnum:]]/} to match the alphabetic
and numeric characters in your character set.

@c Thanks to
@c Date: Tue, 01 Jul 2014 07:39:51 +0200
@c From: Hermann Peifer <peifer@gmx.eu>
Some utilities that match regular expressions provide a nonstandard
@samp{[:ascii:]} character class; @command{awk} does not. However, you
can simulate such a construct using @samp{[\x00-\x7F]}.  This matches
all values numerically between zero and 127, which is the defined
range of the ASCII character set.  Use a complemented character list
(@samp{[^\x00-\x7F]}) to match any single-byte characters that are not
in the ASCII range.

@cindex bracket expressions, collating elements
@cindex bracket expressions, non-ASCII
@cindex collating elements
Two additional special sequences can appear in bracket expressions.
These apply to non-ASCII character sets, which can have single symbols
(called @dfn{collating elements}) that are represented with more than one
character. They can also have several characters that are equivalent for
@dfn{collating}, or sorting, purposes.  (For example, in French, a plain ``e''
and a grave-accented ``@`e'' are equivalent.)
These sequences are:

@table @asis
@cindex bracket expressions, collating symbols
@cindex collating symbols
@item Collating symbols
Multicharacter collating elements enclosed between
@samp{[.} and @samp{.]}.  For example, if @samp{ch} is a collating element,
then @samp{[[.ch.]]} is a regexp that matches this collating element, whereas
@samp{[ch]} is a regexp that matches either @samp{c} or @samp{h}.

@cindex bracket expressions, equivalence classes
@item Equivalence classes
Locale-specific names for a list of
characters that are equal. The name is enclosed between
@samp{[=} and @samp{=]}.
For example, the name @samp{e} might be used to represent all of
``e,'' ``@^e,'' ``@`e,'' and ``@'e.'' In this case, @samp{[[=e=]]} is a regexp
that matches any of @samp{e}, @samp{@^e}, @samp{@'e}, or @samp{@`e}.
@end table

These features are very valuable in non-English-speaking locales.

@cindex internationalization, localization, character classes
@cindex @command{gawk}, character classes and
@cindex POSIX @command{awk}, bracket expressions and, character classes
@quotation CAUTION
The library functions that @command{gawk} uses for regular
expression matching currently recognize only POSIX character classes;
they do not recognize collating symbols or equivalence classes.
@end quotation
@c maybe one day ...

Inside a bracket expression, an opening bracket (@samp{[}) that does
not start a character class, collating element or equivalence class is
taken literally. This is also true of @samp{.} and @samp{*}.

@node Leftmost Longest
@section How Much Text Matches?

@cindex regular expressions, leftmost longest match
@c @cindex matching, leftmost longest
Consider the following:

@example
echo aaaabcd | awk '@{ sub(/a+/, "<A>"); print @}'
@end example

This example uses the @code{sub()} function to make a change to the input
record.  (@code{sub()} replaces the first instance of any text matched
by the first argument with the string provided as the second argument;
@pxref{String Functions}.)  Here, the regexp @code{/a+/} indicates ``one
or more @samp{a} characters,'' and the replacement text is @samp{<A>}.

The input contains four @samp{a} characters.
@command{awk} (and POSIX) regular expressions always match
the leftmost, @emph{longest} sequence of input characters that can
match.  Thus, all four @samp{a} characters are
replaced with @samp{<A>} in this example:

@example
$ @kbd{echo aaaabcd | awk '@{ sub(/a+/, "<A>"); print @}'}
@print{} <A>bcd
@end example

For simple match/no-match tests, this is not so important. But when doing
text matching and substitutions with the @code{match()}, @code{sub()}, @code{gsub()},
and @code{gensub()} functions, it is very important.
@ifinfo
@xref{String Functions},
for more information on these functions.
@end ifinfo
Understanding this principle is also important for regexp-based record
and field splitting (@pxref{Records},
and also @pxref{Field Separators}).

@node Computed Regexps
@section Using Dynamic Regexps

@cindex regular expressions, computed
@cindex regular expressions, dynamic
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@c @cindex operators, @code{~}
@c @cindex operators, @code{!~}
The righthand side of a @samp{~} or @samp{!~} operator need not be a
regexp constant (i.e., a string of characters between slashes).  It may
be any expression.  The expression is evaluated and converted to a string
if necessary; the contents of the string are then used as the
regexp.  A regexp computed in this way is called a @dfn{dynamic
regexp} or a @dfn{computed regexp}:

@example
BEGIN @{ digits_regexp = "[[:digit:]]+" @}
$0 ~ digits_regexp    @{ print @}
@end example

@noindent
This sets @code{digits_regexp} to a regexp that describes one or more digits,
and tests whether the input record matches this regexp.

@quotation NOTE
When using the @samp{~} and @samp{!~}
operators, be aware that there is a difference between a regexp constant
enclosed in slashes and a string constant enclosed in double quotes.
If you are going to use a string constant, you have to understand that
the string is, in essence, scanned @emph{twice}: the first time when
@command{awk} reads your program, and the second time when it goes to
match the string on the lefthand side of the operator with the pattern
on the right.  This is true of any string-valued expression (such as
@code{digits_regexp}, shown in the previous example), not just string constants.
@end quotation

@cindex regexp constants, slashes vs.@: quotes
@cindex @code{\} (backslash), in regexp constants
@cindex backslash (@code{\}), in regexp constants
@cindex @code{"} (double quote), in regexp constants
@cindex double quote (@code{"}), in regexp constants
What difference does it make if the string is
scanned twice? The answer has to do with escape sequences, and particularly
with backslashes.  To get a backslash into a regular expression inside a
string, you have to type two backslashes.

For example, @code{/\*/} is a regexp constant for a literal @samp{*}.
Only one backslash is needed.  To do the same thing with a string,
you have to type @code{"\\*"}.  The first backslash escapes the
second one so that the string actually contains the
two characters @samp{\} and @samp{*}.

@cindex troubleshooting, regexp constants vs.@: string constants
@cindex regexp constants, vs.@: string constants
@cindex string constants, vs.@: regexp constants
Given that you can use both regexp and string constants to describe
regular expressions, which should you use?  The answer is ``regexp
constants,'' for several reasons:

@itemize @value{BULLET}
@item
String constants are more complicated to write and
more difficult to read. Using regexp constants makes your programs
less error-prone.  Not understanding the difference between the two
kinds of constants is a common source of errors.

@item
It is more efficient to use regexp constants. @command{awk} can note
that you have supplied a regexp and store it internally in a form that
makes pattern matching more efficient.  When using a string constant,
@command{awk} must first convert the string into this internal form and
then perform the pattern matching.

@item
Using regexp constants is better form; it shows clearly that you
intend a regexp match.
@end itemize

@sidebar Using @code{\n} in Bracket Expressions of Dynamic Regexps
@cindex regular expressions, dynamic, with embedded newlines
@cindex newlines, in dynamic regexps

Some older versions of @command{awk} do not allow the newline
character to be used inside a bracket expression for a dynamic regexp:

@example
$ @kbd{awk '$0 ~ "[ \t\n]"'}
@error{} awk: newline in character class [
@error{} ]...
@error{}  source line number 1
@error{}  context is
@error{}        $0 ~ "[ >>>  \t\n]" <<<
@end example

@cindex newlines, in regexp constants
But a newline in a regexp constant works with no problem:

@example
$ @kbd{awk '$0 ~ /[ \t\n]/'}
@kbd{here is a sample line}
@print{} here is a sample line
@kbd{Ctrl-d}
@end example

@command{gawk} does not have this problem, and it isn't likely to
occur often in practice, but it's worth noting for future reference.
@end sidebar

@node GNU Regexp Operators
@section @command{gawk}-Specific Regexp Operators

@c This section adapted (long ago) from the regex-0.12 manual

@cindex regular expressions, operators, @command{gawk}
@cindex @command{gawk}, regular expressions, operators
@cindex operators, GNU-specific
@cindex regular expressions, operators, for words
@cindex word, regexp definition of
GNU software that deals with regular expressions provides a number of
additional regexp operators.  These operators are described in this
@value{SECTION} and are specific to @command{gawk};
they are not available in other @command{awk} implementations.
Most of the additional operators deal with word matching.
For our purposes, a @dfn{word} is a sequence of one or more letters, digits,
or underscores (@samp{_}):

@table @code
@c @cindex operators, @code{\s} (@command{gawk})
@cindex backslash (@code{\}), @code{\s} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\s} operator (@command{gawk})
@item \s
Matches any whitespace character.
Think of it as shorthand for
@w{@samp{[[:space:]]}}.

@c @cindex operators, @code{\S} (@command{gawk})
@cindex backslash (@code{\}), @code{\S} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\S} operator (@command{gawk})
@item \S
Matches any character that is not whitespace.
Think of it as shorthand for
@w{@samp{[^[:space:]]}}.

@c @cindex operators, @code{\w} (@command{gawk})
@cindex backslash (@code{\}), @code{\w} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\w} operator (@command{gawk})
@item \w
Matches any word-constituent character---that is, it matches any
letter, digit, or underscore. Think of it as shorthand for
@w{@samp{[[:alnum:]_]}}.

@c @cindex operators, @code{\W} (@command{gawk})
@cindex backslash (@code{\}), @code{\W} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\W} operator (@command{gawk})
@item \W
Matches any character that is not word-constituent.
Think of it as shorthand for
@w{@samp{[^[:alnum:]_]}}.

@c @cindex operators, @code{\<} (@command{gawk})
@cindex backslash (@code{\}), @code{\<} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\<} operator (@command{gawk})
@item \<
Matches the empty string at the beginning of a word.
For example, @code{/\<away/} matches @samp{away} but not
@samp{stowaway}.

@c @cindex operators, @code{\>} (@command{gawk})
@cindex backslash (@code{\}), @code{\>} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\>} operator (@command{gawk})
@item \>
Matches the empty string at the end of a word.
For example, @code{/stow\>/} matches @samp{stow} but not @samp{stowaway}.

@c @cindex operators, @code{\y} (@command{gawk})
@cindex backslash (@code{\}), @code{\y} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\y} operator (@command{gawk})
@cindex word boundaries@comma{} matching
@item \y
Matches the empty string at either the beginning or the
end of a word (i.e., the word boundar@strong{y}).  For example, @samp{\yballs?\y}
matches either @samp{ball} or @samp{balls}, as a separate word.

@c @cindex operators, @code{\B} (@command{gawk})
@cindex backslash (@code{\}), @code{\B} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\B} operator (@command{gawk})
@item \B
Matches the empty string that occurs between two
word-constituent characters. For example,
@code{/\Brat\B/} matches @samp{crate}, but it does not match @samp{dirty rat}.
@samp{\B} is essentially the opposite of @samp{\y}.
@end table

@cindex buffers, operators for
@cindex regular expressions, operators, for buffers
@cindex operators, string-matching, for buffers
There are two other operators that work on buffers.  In Emacs, a
@dfn{buffer} is, naturally, an Emacs buffer.
Other GNU programs, including @command{gawk},
consider the entire string to match as the buffer.
The operators are:

@table @code
@item \`
@c @cindex operators, @code{\`} (@command{gawk})
@cindex backslash (@code{\}), @code{\`} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\`} operator (@command{gawk})
Matches the empty string at the
beginning of a buffer (string)

@c @cindex operators, @code{\'} (@command{gawk})
@cindex backslash (@code{\}), @code{\'} operator (@command{gawk})
@cindex @code{\} (backslash), @code{\'} operator (@command{gawk})
@item \'
Matches the empty string at the
end of a buffer (string)
@end table

@cindex @code{^} (caret), regexp operator
@cindex caret (@code{^}), regexp operator
@cindex @code{?} (question mark), regexp operator
@cindex question mark (@code{?}), regexp operator
Because @samp{^} and @samp{$} always work in terms of the beginning
and end of strings, these operators don't add any new capabilities
for @command{awk}.  They are provided for compatibility with other
GNU software.

@cindex @command{gawk}, word-boundary operator
@cindex word-boundary operator (@command{gawk})
@cindex operators, word-boundary (@command{gawk})
In other GNU software, the word-boundary operator is @samp{\b}. However,
that conflicts with the @command{awk} language's definition of @samp{\b}
as backspace, so @command{gawk} uses a different letter.
An alternative method would have been to require two backslashes in the
GNU operators, but this was deemed too confusing. The current
method of using @samp{\y} for the GNU @samp{\b} appears to be the
lesser of two evils.

@cindex regular expressions, @command{gawk}, command-line options
@cindex @command{gawk}, command-line options, and regular expressions
The various command-line options
(@pxref{Options})
control how @command{gawk} interprets characters in regexps:

@table @asis
@item No options
In the default case, @command{gawk} provides all the facilities of
POSIX regexps and the
@ifnotinfo
previously described
GNU regexp operators.
@end ifnotinfo
@ifnottex
@ifnotdocbook
GNU regexp operators described
in @ref{Regexp Operators}.
@end ifnotdocbook
@end ifnottex

@item @code{--posix}
Match only POSIX regexps; the GNU operators are not special
(e.g., @samp{\w} matches a literal @samp{w}).  Interval expressions
are allowed.

@cindex Brian Kernighan's @command{awk}
@item @code{--traditional}
Match traditional Unix @command{awk} regexps. The GNU operators
are not special, and interval expressions are not available.
Because BWK @command{awk} supports them,
the POSIX character classes (@samp{[[:alnum:]]}, etc.) are available.
Characters described by octal and hexadecimal escape sequences are
treated literally, even if they represent regexp metacharacters.

@item @code{--re-interval}
Allow interval expressions in regexps, if @option{--traditional}
has been provided.
Otherwise, interval expressions are available by default.
@end table

@node Case-sensitivity
@section Case Sensitivity in Matching

@cindex regular expressions, case sensitivity
@cindex case sensitivity, regexps and
Case is normally significant in regular expressions, both when matching
ordinary characters (i.e., not metacharacters) and inside bracket
expressions.  Thus, a @samp{w} in a regular expression matches only a lowercase
@samp{w} and not an uppercase @samp{W}.

The simplest way to do a case-independent match is to use a bracket
expression---for example, @samp{[Ww]}.  However, this can be cumbersome if
you need to use it often, and it can make the regular expressions harder
to read.  There are two alternatives that you might prefer.

One way to perform a case-insensitive match at a particular point in the
program is to convert the data to a single case, using the
@code{tolower()} or @code{toupper()} built-in string functions (which we
haven't discussed yet;
@pxref{String Functions}).
For example:

@example
tolower($1) ~ /foo/  @{ @dots{} @}
@end example

@noindent
converts the first field to lowercase before matching against it.
This works in any POSIX-compliant @command{awk}.

@cindex @command{gawk}, regular expressions, case sensitivity
@cindex case sensitivity, @command{gawk}
@cindex differences in @command{awk} and @command{gawk}, regular expressions
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@cindex @code{IGNORECASE} variable, with @code{~} and @code{!~} operators
@cindex @command{gawk}, @code{IGNORECASE} variable in
@c @cindex variables, @code{IGNORECASE}
Another method, specific to @command{gawk}, is to set the variable
@code{IGNORECASE} to a nonzero value (@pxref{Built-in Variables}).
When @code{IGNORECASE} is not zero, @emph{all} regexp and string
operations ignore case.

Changing the value of @code{IGNORECASE} dynamically controls the
case sensitivity of the program as it runs.  Case is significant by
default because @code{IGNORECASE} (like most variables) is initialized
to zero:

@example
x = "aB"
if (x ~ /ab/) @dots{}   # this test will fail

IGNORECASE = 1
if (x ~ /ab/) @dots{}   # now it will succeed
@end example

In general, you cannot use @code{IGNORECASE} to make certain rules
case insensitive and other rules case sensitive, as there is no
straightforward way
to set @code{IGNORECASE} just for the pattern of
a particular rule.@footnote{Experienced C and C++ programmers will note
that it is possible, using something like
@samp{IGNORECASE = 1 && /foObAr/ @{ @dots{} @}}
and
@samp{IGNORECASE = 0 || /foobar/ @{ @dots{} @}}.
However, this is somewhat obscure and we don't recommend it.}
To do this, use either bracket expressions or @code{tolower()}.  However, one
thing you can do with @code{IGNORECASE} only is dynamically turn
case sensitivity on or off for all the rules at once.

@code{IGNORECASE} can be set on the command line or in a @code{BEGIN} rule
(@pxref{Other Arguments}; also
@pxref{Using BEGIN/END}).
Setting @code{IGNORECASE} from the command line is a way to make
a program case insensitive without having to edit it.

@c @cindex ISO 8859-1
@c @cindex ISO Latin-1
In multibyte locales,
the equivalences between upper-
and lowercase characters are tested based on the wide-character values of
the locale's character set.
Otherwise, the characters are tested based
on the ISO-8859-1 (ISO Latin-1)
character set. This character set is a superset of the traditional 128
ASCII characters, which also provides a number of characters suitable
for use with European languages.@footnote{If you don't understand this,
don't worry about it; it just means that @command{gawk} does
the right thing.}

The value of @code{IGNORECASE} has no effect if @command{gawk} is in
compatibility mode (@pxref{Options}).
Case is always significant in compatibility mode.

@node Regexp Summary
@section Summary

@itemize @value{BULLET}
@item
Regular expressions describe sets of strings to be matched.
In @command{awk}, regular expression constants are written enclosed
between slashes: @code{/}@dots{}@code{/}.

@item
Regexp constants may be used standalone in patterns and
in conditional expressions, or as part of matching expressions
using the @samp{~} and @samp{!~} operators.

@item
Escape sequences let you represent nonprintable characters and
also let you represent regexp metacharacters as literal characters
to be matched.

@item
Regexp operators provide grouping, alternation, and repetition.

@item
Bracket expressions give you a shorthand for specifying sets
of characters that can match at a particular point in a regexp.
Within bracket expressions, POSIX character classes let you specify
certain groups of characters in a locale-independent fashion.

@item
Regular expressions match the leftmost longest text in the string being
matched.  This matters for cases where you need to know the extent of
the match, such as for text substitution and when the record separator
is a regexp.

@item
Matching expressions may use dynamic regexps (i.e., string values
treated as regular expressions).

@item
@command{gawk}'s @code{IGNORECASE} variable lets you control the
case sensitivity of regexp matching.  In other @command{awk}
versions, use @code{tolower()} or @code{toupper()}.

@end itemize


@node Reading Files
@chapter Reading Input Files

@cindex reading input files
@cindex input files, reading
@cindex input files
@cindex @code{FILENAME} variable
In the typical @command{awk} program,
@command{awk} reads all input either from the
standard input (by default, this is the keyboard, but often it is a pipe from another
command) or from files whose names you specify on the @command{awk}
command line.  If you specify input files, @command{awk} reads them
in order, processing all the data from one before going on to the next.
The name of the current input file can be found in the predefined variable
@code{FILENAME}
(@pxref{Built-in Variables}).

@cindex records
@cindex fields
The input is read in units called @dfn{records}, and is processed by the
rules of your program one record at a time.
By default, each record is one line.  Each
record is automatically split into chunks called @dfn{fields}.
This makes it more convenient for programs to work on the parts of a record.

@cindex @code{getline} command
On rare occasions, you may need to use the @code{getline} command.
The  @code{getline} command is valuable both because it
can do explicit input from any number of files, and because the files
used with it do not have to be named on the @command{awk} command line
(@pxref{Getline}).

@menu
* Records::                     Controlling how data is split into records.
* Fields::                      An introduction to fields.
* Nonconstant Fields::          Nonconstant Field Numbers.
* Changing Fields::             Changing the Contents of a Field.
* Field Separators::            The field separator and how to change it.
* Constant Size::               Reading constant width data.
* Splitting By Content::        Defining Fields By Content
* Testing field creation::      Checking how @command{gawk} is splitting
                                records.
* Multiple Line::               Reading multiline records.
* Getline::                     Reading files under explicit program control
                                using the @code{getline} function.
* Read Timeout::                Reading input with a timeout.
* Retrying Input::              Retrying input after certain errors.
* Command-line directories::    What happens if you put a directory on the
                                command line.
* Input Summary::               Input summary.
* Input Exercises::             Exercises.
@end menu

@node Records
@section How Input Is Split into Records

@cindex input, splitting into records
@cindex records, splitting input into
@cindex @code{NR} variable
@cindex @code{FNR} variable
@command{awk} divides the input for your program into records and fields.
It keeps track of the number of records that have been read so far from
the current input file.  This value is stored in a predefined variable
called @code{FNR}, which is reset to zero every time a new file is started.
Another predefined variable, @code{NR}, records the total number of input
records read so far from all @value{DF}s.  It starts at zero, but is
never automatically reset to zero.

@menu
* awk split records::           How standard @command{awk} splits records.
* gawk split records::          How @command{gawk} splits records.
@end menu

@node awk split records
@subsection Record Splitting with Standard @command{awk}

@cindex separators, for records
@cindex record separators
Records are separated by a character called the @dfn{record separator}.
By default, the record separator is the newline character.
This is why records are, by default, single lines.
To use a different character for the record separator,
simply assign that character to the predefined variable @code{RS}.

@cindex newlines, as record separators
@cindex @code{RS} variable
Like any other variable,
the value of @code{RS} can be changed in the @command{awk} program
with the assignment operator, @samp{=}
(@pxref{Assignment Ops}).
The new record-separator character should be enclosed in quotation marks,
which indicate a string constant.  Often, the right time to do this is
at the beginning of execution, before any input is processed,
so that the very first record is read with the proper separator.
To do this, use the special @code{BEGIN} pattern
(@pxref{BEGIN/END}).
For example:

@example
awk 'BEGIN @{ RS = "u" @}
     @{ print $0 @}' mail-list
@end example

@noindent
changes the value of @code{RS} to @samp{u}, before reading any input.
The new value is a string whose first character is the letter ``u''; as a result, records
are separated by the letter ``u''.  Then the input file is read, and the second
rule in the @command{awk} program (the action with no pattern) prints each
record.  Because each @code{print} statement adds a newline at the end of
its output, this @command{awk} program copies the input
with each @samp{u} changed to a newline.  Here are the results of running
the program on @file{mail-list}:

@example
@group
$ @kbd{awk 'BEGIN @{ RS = "u" @}}
>      @kbd{@{ print $0 @}' mail-list}
@end group
@print{} Amelia       555-5553     amelia.zodiac
@print{} sq
@print{} e@@gmail.com    F
@print{} Anthony      555-3412     anthony.assert
@print{} ro@@hotmail.com   A
@print{} Becky        555-7685     becky.algebrar
@print{} m@@gmail.com      A
@print{} Bill         555-1675     bill.drowning@@hotmail.com       A
@print{} Broderick    555-0542     broderick.aliq
@print{} otiens@@yahoo.com R
@print{} Camilla      555-2912     camilla.inf
@print{} sar
@print{} m@@skynet.be     R
@print{} Fabi
@print{} s       555-1234     fabi
@print{} s.
@print{} ndevicesim
@print{} s@@
@print{} cb.ed
@print{}     F
@print{} J
@print{} lie        555-6699     j
@print{} lie.perscr
@print{} tabor@@skeeve.com   F
@print{} Martin       555-6480     martin.codicib
@print{} s@@hotmail.com    A
@print{} Sam
@print{} el       555-3430     sam
@print{} el.lanceolis@@sh
@print{} .ed
@print{}         A
@print{} Jean-Pa
@print{} l    555-2127     jeanpa
@print{} l.campanor
@print{} m@@ny
@print{} .ed
@print{}      R
@print{}
@end example

@noindent
Note that the entry for the name @samp{Bill} is not split.
In the original @value{DF}
(@pxref{Sample Data Files}),
the line looks like this:

@example
Bill         555-1675     bill.drowning@@hotmail.com       A
@end example

@noindent
It contains no @samp{u}, so there is no reason to split the record,
unlike the others, which each have one or more occurrences of the @samp{u}.
In fact, this record is treated as part of the previous record;
the newline separating them in the output
is the original newline in the @value{DF}, not the one added by
@command{awk} when it printed the record!

@cindex record separators, changing
@cindex separators, for records
Another way to change the record separator is on the command line,
using the variable-assignment feature
(@pxref{Other Arguments}):

@example
awk '@{ print $0 @}' RS="u" mail-list
@end example

@noindent
This sets @code{RS} to @samp{u} before processing @file{mail-list}.

Using an alphabetic character such as @samp{u} for the record separator
is highly likely to produce strange results.
Using an unusual character such as @samp{/} is more likely to
produce correct behavior in the majority of cases, but there
are no guarantees. The moral is: Know Your Data.

When using regular characters as the record separator,
there is one unusual case that occurs when @command{gawk} is
being fully POSIX-compliant (@pxref{Options}).
Then, the following (extreme) pipeline prints a surprising @samp{1}:

@example
$ @kbd{echo | gawk --posix 'BEGIN @{ RS = "a" @} ; @{ print NF @}'}
@print{} 1
@end example

There is one field, consisting of a newline.  The value of the built-in
variable @code{NF} is the number of fields in the current record.
(In the normal case, @command{gawk} treats the newline as whitespace,
printing @samp{0} as the result. Most other versions of @command{awk}
also act this way.)

@cindex dark corner, input files
Reaching the end of an input file terminates the current input record,
even if the last character in the file is not the character in @code{RS}.
@value{DARKCORNER}

@cindex empty strings
@cindex null strings
@cindex strings, empty, See null strings
The empty string @code{""} (a string without any characters)
has a special meaning
as the value of @code{RS}. It means that records are separated
by one or more blank lines and nothing else.
@xref{Multiple Line} for more details.

If you change the value of @code{RS} in the middle of an @command{awk} run,
the new value is used to delimit subsequent records, but the record
currently being processed, as well as records already processed, are not
affected.

@cindex @command{gawk}, @code{RT} variable in
@cindex @code{RT} variable
@cindex records, terminating
@cindex terminating records
@cindex differences in @command{awk} and @command{gawk}, record separators
@cindex regular expressions, as record separators
@cindex record separators, regular expressions as
@cindex separators, for records, regular expressions as
After the end of the record has been determined, @command{gawk}
sets the variable @code{RT} to the text in the input that matched
@code{RS}.

@node gawk split records
@subsection Record Splitting with @command{gawk}

@cindex common extensions, @code{RS} as a regexp
@cindex extensions, common@comma{} @code{RS} as a regexp
When using @command{gawk},
the value of @code{RS} is not limited to a one-character
string.  It can be any regular expression
(@pxref{Regexp}). @value{COMMONEXT}
In general, each record
ends at the next string that matches the regular expression; the next
record starts at the end of the matching string.  This general rule is
actually at work in the usual case, where @code{RS} contains just a
newline: a record ends at the beginning of the next matching string (the
next newline in the input), and the following record starts just after
the end of this string (at the first character of the following line).
The newline, because it matches @code{RS}, is not part of either record.

When @code{RS} is a single character, @code{RT}
contains the same single character. However, when @code{RS} is a
regular expression, @code{RT} contains
the actual input text that matched the regular expression.

If the input file ends without any text matching @code{RS},
@command{gawk} sets @code{RT} to the null string.

The following example illustrates both of these features.
It sets @code{RS} equal to a regular expression that
matches either a newline or a series of one or more uppercase letters
with optional leading and/or trailing whitespace:

@example
@group
$ @kbd{echo record 1 AAAA record 2 BBBB record 3 |}
> @kbd{gawk 'BEGIN @{ RS = "\n|( *[[:upper:]]+ *)" @}}
>             @kbd{@{ print "Record =", $0,"and RT = [" RT "]" @}'}
@end group
@print{} Record = record 1 and RT = [ AAAA ]
@print{} Record = record 2 and RT = [ BBBB ]
@print{} Record = record 3 and RT = [
@print{} ]
@end example

@noindent
The square brackets delineate the contents of @code{RT}, letting you
see the leading and trailing whitespace. The final value of
@code{RT} is a newline.
@xref{Simple Sed} for a more useful example
of @code{RS} as a regexp and @code{RT}.

If you set @code{RS} to a regular expression that allows optional
trailing text, such as @samp{RS = "abc(XYZ)?"}, it is possible, due
to implementation constraints, that @command{gawk} may match the leading
part of the regular expression, but not the trailing part, particularly
if the input text that could match the trailing part is fairly long.
@command{gawk} attempts to avoid this problem, but currently, there's
no guarantee that this will never happen.

@quotation NOTE
Remember that in @command{awk}, the @samp{^} and @samp{$} anchor
metacharacters match the beginning and end of a @emph{string}, and not
the beginning and end of a @emph{line}.  As a result, something like
@samp{RS = "^[[:upper:]]"} can only match at the beginning of a file.
This is because @command{gawk} views the input file as one long string
that happens to contain newline characters.
It is thus best to avoid anchor metacharacters in the value of @code{RS}.
@end quotation

@cindex differences in @command{awk} and @command{gawk}, @code{RS}/@code{RT} variables
The use of @code{RS} as a regular expression and the @code{RT}
variable are @command{gawk} extensions; they are not available in
compatibility mode
(@pxref{Options}).
In compatibility mode, only the first character of the value of
@code{RS} determines the end of the record.

@sidebar @code{RS = "\0"} Is Not Portable
@cindex portability, data files as single record
There are times when you might want to treat an entire @value{DF} as a
single record.  The only way to make this happen is to give @code{RS}
a value that you know doesn't occur in the input file.  This is hard
to do in a general way, such that a program always works for arbitrary
input files.

You might think that for text files, the @sc{nul} character, which
consists of a character with all bits equal to zero, is a good
value to use for @code{RS} in this case:

@example
BEGIN @{ RS = "\0" @}  # whole file becomes one record?
@end example

@cindex differences in @command{awk} and @command{gawk}, strings, storing
@command{gawk} in fact accepts this, and uses the @sc{nul}
character for the record separator.
This works for certain special files, such as @file{/proc/environ} on
GNU/Linux systems, where the @sc{nul} character is in fact the record separator.
However, this usage is @emph{not} portable
to most other @command{awk} implementations.

@cindex dark corner, strings, storing
Almost all other @command{awk} implementations@footnote{At least that we know
about.} store strings internally as C-style strings.  C strings use the
@sc{nul} character as the string terminator.  In effect, this means that
@samp{RS = "\0"} is the same as @samp{RS = ""}.
@value{DARKCORNER}

It happens that recent versions of @command{mawk} can use the @sc{nul}
character as a record separator. However, this is a special case:
@command{mawk} does not allow embedded @sc{nul} characters in strings.
(This may change in a future version of @command{mawk}.)

@cindex records, treating files as
@cindex treating files, as single records
@xref{Readfile Function} for an interesting way to read
whole files.  If you are using @command{gawk}, see @ref{Extension Sample
Readfile} for another option.
@end sidebar

@node Fields
@section Examining Fields

@cindex examining fields
@cindex fields
@cindex accessing fields
@cindex fields, examining
When @command{awk} reads an input record, the record is
automatically @dfn{parsed} or separated by the @command{awk} utility into chunks
called @dfn{fields}.  By default, fields are separated by @dfn{whitespace},
like words in a line.
Whitespace in @command{awk} means any string of one or more spaces,
TABs, or newlines; other characters
that are considered whitespace by other languages
(such as formfeed, vertical tab, etc.) are @emph{not} considered
whitespace by @command{awk}.

The purpose of fields is to make it more convenient for you to refer to
these pieces of the record.  You don't have to use them---you can
operate on the whole record if you want---but fields are what make
simple @command{awk} programs so powerful.

@cindex field operator @code{$}
@cindex @code{$} (dollar sign), @code{$} field operator
@cindex dollar sign (@code{$}), @code{$} field operator
@cindex field operators@comma{} dollar sign as
You use a dollar sign (@samp{$})
to refer to a field in an @command{awk} program,
followed by the number of the field you want.  Thus, @code{$1}
refers to the first field, @code{$2} to the second, and so on.
(Unlike in the Unix shells, the field numbers are not limited to single digits.
@code{$127} is the 127th field in the record.)
For example, suppose the following is a line of input:

@example
This seems like a pretty nice example.
@end example

@noindent
Here the first field, or @code{$1}, is @samp{This}, the second field, or
@code{$2}, is @samp{seems}, and so on.  Note that the last field,
@code{$7}, is @samp{example.}.  Because there is no space between the
@samp{e} and the @samp{.}, the period is considered part of the seventh
field.

@cindex @code{NF} variable
@cindex fields, number of
@code{NF} is a predefined variable whose value is the number of fields
in the current record.  @command{awk} automatically updates the value
of @code{NF} each time it reads a record.  No matter how many fields
there are, the last field in a record can be represented by @code{$NF}.
So, @code{$NF} is the same as @code{$7}, which is @samp{example.}.
If you try to reference a field beyond the last
one (such as @code{$8} when the record has only seven fields), you get
the empty string.  (If used in a numeric operation, you get zero.)

The use of @code{$0}, which looks like a reference to the ``zeroth'' field, is
a special case: it represents the whole input record. Use it
when you are not interested in specific fields.
Here are some more examples:

@example
$ @kbd{awk '$1 ~ /li/ @{ print $0 @}' mail-list}
@print{} Amelia       555-5553     amelia.zodiacusque@@gmail.com    F
@print{} Julie        555-6699     julie.perscrutabor@@skeeve.com   F
@end example

@noindent
This example prints each record in the file @file{mail-list} whose first
field contains the string @samp{li}.

By contrast, the following example looks for @samp{li} in @emph{the
entire record} and prints the first and last fields for each matching
input record:

@example
$ @kbd{awk '/li/ @{ print $1, $NF @}' mail-list}
@print{} Amelia F
@print{} Broderick R
@print{} Julie F
@print{} Samuel A
@end example

@node Nonconstant Fields
@section Nonconstant Field Numbers
@cindex fields, numbers
@cindex field numbers

A field number need not be a constant.  Any expression in
the @command{awk} language can be used after a @samp{$} to refer to a
field.  The value of the expression specifies the field number.  If the
value is a string, rather than a number, it is converted to a number.
Consider this example:

@example
awk '@{ print $NR @}'
@end example

@noindent
Recall that @code{NR} is the number of records read so far: one in the
first record, two in the second, and so on.  So this example prints the first
field of the first record, the second field of the second record, and so
on.  For the twentieth record, field number 20 is printed; most likely,
the record has fewer than 20 fields, so this prints a blank line.
Here is another example of using expressions as field numbers:

@example
awk '@{ print $(2*2) @}' mail-list
@end example

@command{awk} evaluates the expression @samp{(2*2)} and uses
its value as the number of the field to print.  The @samp{*}
represents multiplication, so the expression @samp{2*2} evaluates to four.
The parentheses are used so that the multiplication is done before the
@samp{$} operation; they are necessary whenever there is a binary
operator@footnote{A @dfn{binary operator}, such as @samp{*} for
multiplication, is one that takes two operands. The distinction
is required because @command{awk} also has unary (one-operand)
and ternary (three-operand) operators.}
in the field-number expression.  This example, then, prints the
type of relationship (the fourth field) for every line of the file
@file{mail-list}.  (All of the @command{awk} operators are listed, in
order of decreasing precedence, in
@ref{Precedence}.)

If the field number you compute is zero, you get the entire record.
Thus, @samp{$(2-2)} has the same value as @code{$0}.  Negative field
numbers are not allowed; trying to reference one usually terminates
the program.  (The POSIX standard does not define
what happens when you reference a negative field number.  @command{gawk}
notices this and terminates your program.  Other @command{awk}
implementations may behave differently.)

As mentioned in @ref{Fields},
@command{awk} stores the current record's number of fields in the built-in
variable @code{NF} (also @pxref{Built-in Variables}).  Thus, the expression
@code{$NF} is not a special feature---it is the direct consequence of
evaluating @code{NF} and using its value as a field number.

@node Changing Fields
@section Changing the Contents of a Field

@cindex fields, changing contents of
The contents of a field, as seen by @command{awk}, can be changed within an
@command{awk} program; this changes what @command{awk} perceives as the
current input record.  (The actual input is untouched; @command{awk} @emph{never}
modifies the input file.)
Consider the following example and its output:

@example
$ @kbd{awk '@{ nboxes = $3 ; $3 = $3 - 10}
>        @kbd{print nboxes, $3 @}' inventory-shipped}
@print{} 25 15
@print{} 32 22
@print{} 24 14
@dots{}
@end example

@noindent
The program first saves the original value of field three in the variable
@code{nboxes}.
The @samp{-} sign represents subtraction, so this program reassigns
field three, @code{$3}, as the original value of field three minus ten:
@samp{$3 - 10}.  (@xref{Arithmetic Ops}.)
Then it prints the original and new values for field three.
(Someone in the warehouse made a consistent mistake while inventorying
the red boxes.)

For this to work, the text in @code{$3} must make sense
as a number; the string of characters must be converted to a number
for the computer to do arithmetic on it.  The number resulting
from the subtraction is converted back to a string of characters that
then becomes field three.
@xref{Conversion}.

When the value of a field is changed (as perceived by @command{awk}), the
text of the input record is recalculated to contain the new field where
the old one was.  In other words, @code{$0} changes to reflect the altered
field.  Thus, this program
prints a copy of the input file, with 10 subtracted from the second
field of each line:

@example
$ @kbd{awk '@{ $2 = $2 - 10; print $0 @}' inventory-shipped}
@print{} Jan 3 25 15 115
@print{} Feb 5 32 24 226
@print{} Mar 5 24 34 228
@dots{}
@end example

It is also possible to assign contents to fields that are out
of range.  For example:

@example
$ @kbd{awk '@{ $6 = ($5 + $4 + $3 + $2)}
> @kbd{       print $6 @}' inventory-shipped}
@print{} 168
@print{} 297
@print{} 301
@dots{}
@end example

@cindex adding, fields
@cindex fields, adding
@noindent
We've just created @code{$6}, whose value is the sum of fields
@code{$2}, @code{$3}, @code{$4}, and @code{$5}.  The @samp{+} sign
represents addition.  For the file @file{inventory-shipped}, @code{$6}
represents the total number of parcels shipped for a particular month.

Creating a new field changes @command{awk}'s internal copy of the current
input record, which is the value of @code{$0}.  Thus, if you do @samp{print $0}
after adding a field, the record printed includes the new field, with
the appropriate number of field separators between it and the previously
existing fields.

@cindex @code{OFS} variable
@cindex output field separator, See @code{OFS} variable
@cindex field separators, See Also @code{OFS}
This recomputation affects and is affected by
@code{NF} (the number of fields; @pxref{Fields}).
For example, the value of @code{NF} is set to the number of the highest
field you create.
The exact format of @code{$0} is also affected by a feature that has not been discussed yet:
the @dfn{output field separator}, @code{OFS},
used to separate the fields (@pxref{Output Separators}).

Note, however, that merely @emph{referencing} an out-of-range field
does @emph{not} change the value of either @code{$0} or @code{NF}.
Referencing an out-of-range field only produces an empty string.  For
example:

@example
if ($(NF+1) != "")
    print "can't happen"
else
    print "everything is normal"
@end example

@noindent
should print @samp{everything is normal}, because @code{NF+1} is certain
to be out of range.  (@xref{If Statement}
for more information about @command{awk}'s @code{if-else} statements.
@xref{Typing and Comparison}
for more information about the @samp{!=} operator.)

It is important to note that making an assignment to an existing field
changes the
value of @code{$0} but does not change the value of @code{NF},
even when you assign the empty string to a field.  For example:

@example
$ @kbd{echo a b c d | awk '@{ OFS = ":"; $2 = ""}
>                       @kbd{print $0; print NF @}'}
@print{} a::c:d
@print{} 4
@end example

@noindent
The field is still there; it just has an empty value, delimited by
the two colons between @samp{a} and @samp{c}.
This example shows what happens if you create a new field:

@example
$ @kbd{echo a b c d | awk '@{ OFS = ":"; $2 = ""; $6 = "new"}
>                       @kbd{print $0; print NF @}'}
@print{} a::c:d::new
@print{} 6
@end example

@noindent
The intervening field, @code{$5}, is created with an empty value
(indicated by the second pair of adjacent colons),
and @code{NF} is updated with the value six.

@cindex dark corner, @code{NF} variable, decrementing
@cindex @code{NF} variable, decrementing
Decrementing @code{NF} throws away the values of the fields
after the new value of @code{NF} and recomputes @code{$0}.
@value{DARKCORNER}
Here is an example:

@example
$ @kbd{echo a b c d e f | awk '@{ print "NF =", NF;}
> @kbd{                          NF = 3; print $0 @}'}
@print{} NF = 6
@print{} a b c
@end example

@cindex portability, @code{NF} variable@comma{} decrementing
@quotation CAUTION
Some versions of @command{awk} don't
rebuild @code{$0} when @code{NF} is decremented.
@end quotation

Finally, there are times when it is convenient to force
@command{awk} to rebuild the entire record, using the current
values of the fields and @code{OFS}.  To do this, use the
seemingly innocuous assignment:

@example
@group
$1 = $1   # force record to be reconstituted
print $0  # or whatever else with $0
@end group
@end example

@noindent
This forces @command{awk} to rebuild the record.  It does help
to add a comment, as we've shown here.

There is a flip side to the relationship between @code{$0} and
the fields.  Any assignment to @code{$0} causes the record to be
reparsed into fields using the @emph{current} value of @code{FS}.
This also applies to any built-in function that updates @code{$0},
such as @code{sub()} and @code{gsub()}
(@pxref{String Functions}).

@sidebar Understanding @code{$0}

It is important to remember that @code{$0} is the @emph{full}
record, exactly as it was read from the input.  This includes
any leading or trailing whitespace, and the exact whitespace (or other
characters) that separates the fields.

It is a common error to try to change the field separators
in a record simply by setting @code{FS} and @code{OFS}, and then
expecting a plain @samp{print} or @samp{print $0} to print the
modified record.

But this does not work, because nothing was done to change the record
itself.  Instead, you must force the record to be rebuilt, typically
with a statement such as @samp{$1 = $1}, as described earlier.
@end sidebar


@node Field Separators
@section Specifying How Fields Are Separated

@menu
* Default Field Splitting::      How fields are normally separated.
* Regexp Field Splitting::       Using regexps as the field separator.
* Single Character Fields::      Making each character a separate field.
* Command Line Field Separator:: Setting @code{FS} from the command line.
* Full Line Fields::             Making the full line be a single field.
* Field Splitting Summary::      Some final points and a summary table.
@end menu

@cindex @code{FS} variable
@cindex fields, separating
@cindex field separators
@cindex fields, separating
The @dfn{field separator}, which is either a single character or a regular
expression, controls the way @command{awk} splits an input record into fields.
@command{awk} scans the input record for character sequences that
match the separator; the fields themselves are the text between the matches.

In the examples that follow, we use the bullet symbol (@bullet{}) to
represent spaces in the output.
If the field separator is @samp{oo}, then the following line:

@example
moo goo gai pan
@end example

@noindent
is split into three fields: @samp{m}, @samp{@bullet{}g}, and
@samp{@bullet{}gai@bullet{}pan}.
Note the leading spaces in the values of the second and third fields.

@cindex troubleshooting, @command{awk} uses @code{FS} not @code{IFS}
The field separator is represented by the predefined variable @code{FS}.
Shell programmers take note:  @command{awk} does @emph{not} use the
name @code{IFS} that is used by the POSIX-compliant shells (such as
the Unix Bourne shell, @command{sh}, or Bash).

@cindex @code{FS} variable, changing value of
The value of @code{FS} can be changed in the @command{awk} program with the
assignment operator, @samp{=} (@pxref{Assignment Ops}).
Often, the right time to do this is at the beginning of execution
before any input has been processed, so that the very first record
is read with the proper separator.  To do this, use the special
@code{BEGIN} pattern
(@pxref{BEGIN/END}).
For example, here we set the value of @code{FS} to the string
@code{","}:

@example
awk 'BEGIN @{ FS = "," @} ; @{ print $2 @}'
@end example

@cindex @code{BEGIN} pattern
@noindent
Given the input line:

@example
John Q. Smith, 29 Oak St., Walamazoo, MI 42139
@end example

@noindent
this @command{awk} program extracts and prints the string
@samp{@bullet{}29@bullet{}Oak@bullet{}St.}.

@cindex field separators, choice of
@cindex regular expressions as field separators
@cindex field separators, regular expressions as
Sometimes the input data contains separator characters that don't
separate fields the way you thought they would.  For instance, the
person's name in the example we just used might have a title or
suffix attached, such as:

@example
John Q. Smith, LXIX, 29 Oak St., Walamazoo, MI 42139
@end example

@noindent
The same program would extract @samp{@bullet{}LXIX} instead of
@samp{@bullet{}29@bullet{}Oak@bullet{}St.}.
If you were expecting the program to print the
address, you would be surprised.  The moral is to choose your data layout and
separator characters carefully to prevent such problems.
(If the data is not in a form that is easy to process, perhaps you
can massage it first with a separate @command{awk} program.)


@node Default Field Splitting
@subsection Whitespace Normally Separates Fields

@cindex whitespace, as field separators
Fields are normally separated by whitespace sequences
(spaces, TABs, and newlines), not by single spaces.  Two spaces in a row do not
delimit an empty field.  The default value of the field separator @code{FS}
is a string containing a single space, @w{@code{" "}}.  If @command{awk}
interpreted this value in the usual way, each space character would separate
fields, so two spaces in a row would make an empty field between them.
The reason this does not happen is that a single space as the value of
@code{FS} is a special case---it is taken to specify the default manner
of delimiting fields.

If @code{FS} is any other single character, such as @code{","}, then
each occurrence of that character separates two fields.  Two consecutive
occurrences delimit an empty field.  If the character occurs at the
beginning or the end of the line, that too delimits an empty field.  The
space character is the only single character that does not follow these
rules.

@node Regexp Field Splitting
@subsection Using Regular Expressions to Separate Fields

@cindex regular expressions, as field separators
@cindex field separators, regular expressions as
The previous @value{SUBSECTION}
discussed the use of single characters or simple strings as the
value of @code{FS}.
More generally, the value of @code{FS} may be a string containing any
regular expression.  In this case, each match in the record for the regular
expression separates fields.  For example, the assignment:

@example
FS = ", \t"
@end example

@noindent
makes every area of an input line that consists of a comma followed by a
space and a TAB into a field separator.
@ifinfo
(@samp{\t}
is an @dfn{escape sequence} that stands for a TAB;
@pxref{Escape Sequences},
for the complete list of similar escape sequences.)
@end ifinfo

For a less trivial example of a regular expression, try using
single spaces to separate fields the way single commas are used.
@code{FS} can be set to @w{@code{"[@ ]"}} (left bracket, space, right
bracket).  This regular expression matches a single space and nothing else
(@pxref{Regexp}).

There is an important difference between the two cases of @samp{FS = @w{" "}}
(a single space) and @samp{FS = @w{"[ \t\n]+"}}
(a regular expression matching one or more spaces, TABs, or newlines).
For both values of @code{FS}, fields are separated by @dfn{runs}
(multiple adjacent occurrences) of spaces, TABs,
and/or newlines.  However, when the value of @code{FS} is @w{@code{" "}},
@command{awk} first strips leading and trailing whitespace from
the record and then decides where the fields are.
For example, the following pipeline prints @samp{b}:

@example
$ @kbd{echo ' a b c d ' | awk '@{ print $2 @}'}
@print{} b
@end example

@noindent
However, this pipeline prints @samp{a} (note the extra spaces around
each letter):

@example
$ @kbd{echo ' a  b  c  d ' | awk 'BEGIN @{ FS = "[ \t\n]+" @}}
>                                  @kbd{@{ print $2 @}'}
@print{} a
@end example

@noindent
@cindex null strings
@cindex strings, null
@cindex empty strings, See null strings
In this case, the first field is null, or empty.

The stripping of leading and trailing whitespace also comes into
play whenever @code{$0} is recomputed.  For instance, study this pipeline:

@example
$ @kbd{echo '   a b c d' | awk '@{ print; $2 = $2; print @}'}
@print{}    a b c d
@print{} a b c d
@end example

@noindent
The first @code{print} statement prints the record as it was read,
with leading whitespace intact.  The assignment to @code{$2} rebuilds
@code{$0} by concatenating @code{$1} through @code{$NF} together,
separated by the value of @code{OFS} (which is a space by default).
Because the leading whitespace was ignored when finding @code{$1},
it is not part of the new @code{$0}.  Finally, the last @code{print}
statement prints the new @code{$0}.

@cindex @code{FS}, containing @code{^}
@cindex @code{^} (caret), in @code{FS}
@cindex dark corner, @code{^}, in @code{FS}
There is an additional subtlety to be aware of when using regular expressions
for field splitting.
It is not well specified in the POSIX standard, or anywhere else, what @samp{^}
means when splitting fields.  Does the @samp{^}  match only at the beginning of
the entire record? Or is each field separator a new string?  It turns out that
different @command{awk} versions answer this question differently, and you
should not rely on any specific behavior in your programs.
@value{DARKCORNER}

@cindex Brian Kernighan's @command{awk}
As a point of information, BWK @command{awk} allows @samp{^}
to match only at the beginning of the record. @command{gawk}
also works this way. For example:

@example
$ @kbd{echo 'xxAA  xxBxx  C' |}
> @kbd{gawk -F '(^x+)|( +)' '@{ for (i = 1; i <= NF; i++)}
> @kbd{                            printf "-->%s<--\n", $i @}'}
@print{} --><--
@print{} -->AA<--
@print{} -->xxBxx<--
@print{} -->C<--
@end example

@node Single Character Fields
@subsection Making Each Character a Separate Field

@cindex common extensions, single character fields
@cindex extensions, common@comma{} single character fields
@cindex differences in @command{awk} and @command{gawk}, single-character fields
@cindex single-character fields
@cindex fields, single-character
There are times when you may want to examine each character
of a record separately.  This can be done in @command{gawk} by
simply assigning the null string (@code{""}) to @code{FS}. @value{COMMONEXT}
In this case,
each individual character in the record becomes a separate field.
For example:

@example
$ @kbd{echo a b | gawk 'BEGIN @{ FS = "" @}}
>                  @kbd{@{}
>                      @kbd{for (i = 1; i <= NF; i = i + 1)}
>                          @kbd{print "Field", i, "is", $i}
>                  @kbd{@}'}
@print{} Field 1 is a
@print{} Field 2 is
@print{} Field 3 is b
@end example

@cindex dark corner, @code{FS} as null string
@cindex @code{FS} variable, as null string
Traditionally, the behavior of @code{FS} equal to @code{""} was not defined.
In this case, most versions of Unix @command{awk} simply treat the entire record
as only having one field.
@value{DARKCORNER}
In compatibility mode
(@pxref{Options}),
if @code{FS} is the null string, then @command{gawk} also
behaves this way.

@node Command Line Field Separator
@subsection Setting @code{FS} from the Command Line
@cindex @option{-F} option, command-line
@cindex field separator, on command line
@cindex command line, @code{FS} on@comma{} setting
@cindex @code{FS} variable, setting from command line

@code{FS} can be set on the command line.  Use the @option{-F} option to
do so.  For example:

@example
awk -F, '@var{program}' @var{input-files}
@end example

@noindent
sets @code{FS} to the @samp{,} character.  Notice that the option uses
an uppercase @samp{F} instead of a lowercase @samp{f}. The latter
option (@option{-f}) specifies a file containing an @command{awk} program.

The value used for the argument to @option{-F} is processed in exactly the
same way as assignments to the predefined variable @code{FS}.
Any special characters in the field separator must be escaped
appropriately.  For example, to use a @samp{\} as the field separator
on the command line, you would have to type:

@example
# same as FS = "\\"
awk -F\\\\ '@dots{}' files @dots{}
@end example

@noindent
@cindex @code{\} (backslash), as field separator
@cindex backslash (@code{\}), as field separator
Because @samp{\} is used for quoting in the shell, @command{awk} sees
@samp{-F\\}.  Then @command{awk} processes the @samp{\\} for escape
characters (@pxref{Escape Sequences}), finally yielding
a single @samp{\} to use for the field separator.

@c @cindex historical features
As a special case, in compatibility mode
(@pxref{Options}),
if the argument to @option{-F} is @samp{t}, then @code{FS} is set to
the TAB character.  If you type @samp{-F\t} at the
shell, without any quotes, the @samp{\} gets deleted, so @command{awk}
figures that you really want your fields to be separated with TABs and
not @samp{t}s.  Use @samp{-v FS="t"} or @samp{-F"[t]"} on the command line
if you really do want to separate your fields with @samp{t}s.
Use @samp{-F '\t'} when not in compatibility mode to specify that TABs
separate fields.

As an example, let's use an @command{awk} program file called @file{edu.awk}
that contains the pattern @code{/edu/} and the action @samp{print $1}:

@example
/edu/   @{ print $1 @}
@end example

Let's also set @code{FS} to be the @samp{-} character and run the
program on the file @file{mail-list}.  The following command prints a
list of the names of the people that work at or attend a university, and
the first three digits of their phone numbers:

@example
$ @kbd{awk -F- -f edu.awk mail-list}
@print{} Fabius       555
@print{} Samuel       555
@print{} Jean
@end example

@noindent
Note the third line of output.  The third line
in the original file looked like this:

@example
Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@end example

The @samp{-} as part of the person's name was used as the field
separator, instead of the @samp{-} in the phone number that was
originally intended.  This demonstrates why you have to be careful in
choosing your field and record separators.

@cindex Unix @command{awk}, password files@comma{} field separators and
Perhaps the most common use of a single character as the field separator
occurs when processing the Unix system password file.  On many Unix
systems, each user has a separate entry in the system password file, with one
line per user.  The information in these lines is separated by colons.
The first field is the user's login name and the second is the user's
encrypted or shadow password.  (A shadow password is indicated by the
presence of a single @samp{x} in the second field.)  A password file
entry might look like this:

@cindex Robbins, Arnold
@example
arnold:x:2076:10:Arnold Robbins:/home/arnold:/bin/bash
@end example

The following program searches the system password file and prints
the entries for users whose full name is not indicated:

@example
awk -F: '$5 == ""' /etc/passwd
@end example

@node Full Line Fields
@subsection Making the Full Line Be a Single Field

Occasionally, it's useful to treat the whole input line as a
single field.  This can be done easily and portably simply by
setting @code{FS} to @code{"\n"} (a newline):@footnote{Thanks to
Andrew Schorr for this tip.}

@example
awk -F'\n' '@var{program}' @var{files @dots{}}
@end example

@noindent
When you do this, @code{$1} is the same as @code{$0}.

@sidebar Changing @code{FS} Does Not Affect the Fields

@cindex POSIX @command{awk}, field separators and
@cindex field separator, POSIX and
According to the POSIX standard, @command{awk} is supposed to behave
as if each record is split into fields at the time it is read.
In particular, this means that if you change the value of @code{FS}
after a record is read, the values of the fields (i.e., how they were split)
should reflect the old value of @code{FS}, not the new one.

@cindex dark corner, field separators
@cindex @command{sed} utility
@cindex stream editors
However, many older implementations of @command{awk} do not work this way.  Instead,
they defer splitting the fields until a field is actually
referenced.  The fields are split
using the @emph{current} value of @code{FS}!
@value{DARKCORNER}
This behavior can be difficult
to diagnose. The following example illustrates the difference
between the two methods:

@example
sed 1q /etc/passwd | awk '@{ FS = ":" ; print $1 @}'
@end example

@noindent
which usually prints:

@example
root
@end example

@noindent
on an incorrect implementation of @command{awk}, while @command{gawk}
prints the full first line of the file, something like:

@example
root:x:0:0:Root:/:
@end example

(The @command{sed}@footnote{The @command{sed} utility is a ``stream editor.''
Its behavior is also defined by the POSIX standard.}
command prints just the first line of @file{/etc/passwd}.)
@end sidebar

@node Field Splitting Summary
@subsection Field-Splitting Summary

It is important to remember that when you assign a string constant
as the value of @code{FS}, it undergoes normal @command{awk} string
processing.  For example, with Unix @command{awk} and @command{gawk},
the assignment @samp{FS = "\.."} assigns the character string @code{".."}
to @code{FS} (the backslash is stripped).  This creates a regexp meaning
``fields are separated by occurrences of any two characters.''
If instead you want fields to be separated by a literal period followed
by any single character, use @samp{FS = "\\.."}.

The following list summarizes how fields are split, based on the value
of @code{FS} (@samp{==} means ``is equal to''):

@table @code
@item FS == " "
Fields are separated by runs of whitespace.  Leading and trailing
whitespace are ignored.  This is the default.

@item FS == @var{any other single character}
Fields are separated by each occurrence of the character.  Multiple
successive occurrences delimit empty fields, as do leading and
trailing occurrences.
The character can even be a regexp metacharacter; it does not need
to be escaped.

@item FS == @var{regexp}
Fields are separated by occurrences of characters that match @var{regexp}.
Leading and trailing matches of @var{regexp} delimit empty fields.

@item FS == ""
Each individual character in the record becomes a separate field.
(This is a common extension; it is not specified by the POSIX standard.)
@end table

@sidebar @code{FS} and @code{IGNORECASE}

The @code{IGNORECASE} variable
(@pxref{User-modified})
affects field splitting @emph{only} when the value of @code{FS} is a regexp.
It has no effect when @code{FS} is a single character, even if
that character is a letter.  Thus, in the following code:

@example
FS = "c"
IGNORECASE = 1
$0 = "aCa"
print $1
@end example

@noindent
The output is @samp{aCa}.  If you really want to split fields on an
alphabetic character while ignoring case, use a regexp that will
do it for you (e.g., @samp{FS = "[c]"}).  In this case, @code{IGNORECASE}
will take effect.
@end sidebar


@node Constant Size
@section Reading Fixed-Width Data

@cindex data, fixed-width
@cindex fixed-width data
@cindex advanced features, fixed-width data

@c O'Reilly doesn't like it as a note the first thing in the section.
This @value{SECTION} discusses an advanced
feature of @command{gawk}.  If you are a novice @command{awk} user,
you might want to skip it on the first reading.

@command{gawk} provides a facility for dealing with fixed-width fields
with no distinctive field separator. We discuss this feature in
the following @value{SUBSECTION}s.

@menu
* Fixed width data::            Processing fixed-width data.
* Skipping intervening::        Skipping intervening fields.
* Allowing trailing data::      Capturing optional trailing data.
* Fields with fixed data::      Field values with fixed-width data.
@end menu

@node Fixed width data
@subsection Processing Fixed-Width Data

An example of fixed-width data would be the input for old Fortran programs
where numbers are run together, or the output of programs that did not
anticipate the use of their output as input for other programs.

An example of the latter is a table where all the columns are lined up
by the use of a variable number of spaces and @emph{empty fields are
just spaces}.  Clearly, @command{awk}'s normal field splitting based
on @code{FS} does not work well in this case.  Although a portable
@command{awk} program can use a series of @code{substr()} calls on
@code{$0} (@pxref{String Functions}), this is awkward and inefficient
for a large number of fields.

@cindex troubleshooting, fatal errors, field widths@comma{} specifying
@cindex @command{w} utility
@cindex @code{FIELDWIDTHS} variable
@cindex @command{gawk}, @code{FIELDWIDTHS} variable in
The splitting of an input record into fixed-width fields is specified by
assigning a string containing space-separated numbers to the built-in
variable @code{FIELDWIDTHS}.  Each number specifies the width of the
field, @emph{including} columns between fields.  If you want to ignore
the columns between fields, you can specify the width as a separate
field that is subsequently ignored.  It is a fatal error to supply a
field width that has a negative value.

The following data is the output of the Unix @command{w} utility.  It is useful
to illustrate the use of @code{FIELDWIDTHS}:

@example
@group
 10:06pm  up 21 days, 14:04,  23 users
User     tty       login@  idle   JCPU   PCPU  what
hzuo     ttyV0     8:58pm            9      5  vi p24.tex
hzang    ttyV3     6:37pm    50                -csh
eklye    ttyV5     9:53pm            7      1  em thes.tex
dportein ttyV6     8:17pm  1:47                -csh
gierd    ttyD3    10:00pm     1                elm
dave     ttyD4     9:47pm            4      4  w
brent    ttyp0    26Jun91  4:46  26:46   4:41  bash
dave     ttyq4    26Jun9115days     46     46  wnewmail
@end group
@end example

The following program takes this input, converts the idle time to
number of seconds, and prints out the first two fields and the calculated
idle time:

@example
BEGIN  @{ FIELDWIDTHS = "9 6 10 6 7 7 35" @}
NR > 2 @{
    idle = $4
    sub(/^ +/, "", idle)   # strip leading spaces
    if (idle == "")
        idle = 0
    if (idle ~ /:/) @{      # hh:mm
        split(idle, t, ":")
        idle = t[1] * 60 + t[2]
    @}
    if (idle ~ /days/)
        idle *= 24 * 60 * 60

    print $1, $2, idle
@}
@end example

@quotation NOTE
The preceding program uses a number of @command{awk} features that
haven't been introduced yet.
@end quotation

Running the program on the data produces the following results:

@example
hzuo      ttyV0  0
hzang     ttyV3  50
eklye     ttyV5  0
dportein  ttyV6  107
gierd     ttyD3  1
dave      ttyD4  0
brent     ttyp0  286
dave      ttyq4  1296000
@end example

Another (possibly more practical) example of fixed-width input data
is the input from a deck of balloting cards.  In some parts of
the United States, voters mark their choices by punching holes in computer
cards.  These cards are then processed to count the votes for any particular
candidate or on any particular issue.  Because a voter may choose not to
vote on some issue, any column on the card may be empty.  An @command{awk}
program for processing such data could use the @code{FIELDWIDTHS} feature
to simplify reading the data.  (Of course, getting @command{gawk} to run on
a system with card readers is another story!)

@node Skipping intervening
@subsection Skipping Intervening Fields

Starting in @value{PVERSION} 4.2, each field width may optionally be
preceded by a colon-separated value specifying the number of characters
to skip before the field starts.  Thus, the preceding program could be
rewritten to specify @code{FIELDWIDTHS} like so:

@example
BEGIN  @{ FIELDWIDTHS = "8 1:5 4:7 6 1:6 1:6 2:33" @}
@end example

This strips away some of the white space separating the fields. With such
a change, the program produces the following results:

@example
hzang    ttyV3 50
eklye    ttyV5 0
dportein ttyV6 107
gierd    ttyD3 1
dave     ttyD4 0
brent    ttyp0 286
dave     ttyq4 1296000
@end example

@node Allowing trailing data
@subsection Capturing Optional Trailing Data

There are times when fixed-width data may be followed by additional data
that has no fixed length.  Such data may or may not be present, but if
it is, it should be possible to get at it from an @command{awk} program.

Starting with @value{PVERSION} 4.2, in order to provide a way to say ``anything
else in the record after the defined fields,'' @command{gawk}
allows you to add a final @samp{*} character to the value of
@code{FIELDWIDTHS}. There can only be one such character, and it must
be the final non-whitespace character in @code{FIELDWIDTHS}.
For example:

@example
$ @kbd{cat fw.awk}                         @ii{Show the program}
@print{} BEGIN @{ FIELDWIDTHS = "2 2 *" @}
@print{} @{ print NF, $1, $2, $3 @}
$ @kbd{cat fw.in}                          @ii{Show sample input}
@print{} 1234abcdefghi
$ @kbd{gawk -f fw.awk fw.in}               @ii{Run the program}
@print{} 3 12 34 abcdefghi
@end example

@node Fields with fixed data
@subsection Field Values With Fixed-Width Data

So far, so good.  But what happens if there isn't as much data as there
should be based on the contents of @code{FIELDWIDTHS}? Or, what happens
if there is more data than expected?

For many years, what happens in these cases was not well defined. Starting
with @value{PVERSION} 4.2, the rules are as follows:

@table @asis
@item Enough data for some fields
For example, if @code{FIELDWIDTHS} is set to @code{"2 3 4"} and the
input record is @samp{aabbb}.  In this case, @code{NF} is set to two.

@item Not enough data for a field
For example, if @code{FIELDWIDTHS} is set to @code{"2 3 4"} and the
input record is @samp{aab}.  In this case, @code{NF} is set to two and
@code{$2} has the value @code{"b"}. The idea is that even though there
aren't as many characters as were expected, there are some, so the data
should be made available to the program.

@item Too much data
For example, if @code{FIELDWIDTHS} is set to @code{"2 3 4"} and the
input record is @samp{aabbbccccddd}.  In this case, @code{NF} is set to
three and the extra characters (@samp{ddd}) are ignored.  If you want
@command{gawk} to capture the extra characters, supply a final @samp{*}
in the value of @code{FIELDWIDTHS}.

@item Too much data, but with @samp{*} supplied
For example, if @code{FIELDWIDTHS} is set to @code{"2 3 4 *"} and the
input record is @samp{aabbbccccddd}.  In this case, @code{NF} is set to
four, and @code{$4} has the value @code{"ddd"}.

@end table

@node Splitting By Content
@section Defining Fields by Content

@c O'Reilly doesn't like it as a note the first thing in the section.
This @value{SECTION} discusses an advanced
feature of @command{gawk}.  If you are a novice @command{awk} user,
you might want to skip it on the first reading.

@cindex advanced features, specifying field content
Normally, when using @code{FS}, @command{gawk} defines the fields as the
parts of the record that occur in between each field separator. In other
words, @code{FS} defines what a field @emph{is not}, instead of what a field
@emph{is}.
However, there are times when you really want to define the fields by
what they are, and not by what they are not.

The most notorious such case
is so-called @dfn{comma-separated values} (CSV) data. Many spreadsheet programs,
for example, can export their data into text files, where each record is
terminated with a newline, and fields are separated by commas. If
commas only separated the data, there wouldn't be an issue. The problem comes when
one of the fields contains an @emph{embedded} comma.
In such cases, most programs embed the field in double quotes.@footnote{The
CSV format lacked a formal standard definition for many years.
@uref{http://www.ietf.org/rfc/rfc4180.txt, RFC 4180}
standardizes the most common practices.}
So, we might have data like this:

@example
@c file eg/misc/addresses.csv
Robbins,Arnold,"1234 A Pretty Street, NE",MyTown,MyState,12345-6789,USA
@c endfile
@end example

@cindex @command{gawk}, @code{FPAT} variable in
@cindex @code{FPAT} variable
The @code{FPAT} variable offers a solution for cases like this.
The value of @code{FPAT} should be a string that provides a regular expression.
This regular expression describes the contents of each field.

In the case of CSV data as presented here, each field is either ``anything that
is not a comma,'' or ``a double quote, anything that is not a double quote, and a
closing double quote.''  If written as a regular expression constant
(@pxref{Regexp}),
we would have @code{/([^,]+)|("[^"]+")/}.
Writing this as a string requires us to escape the double quotes, leading to:

@example
FPAT = "([^,]+)|(\"[^\"]+\")"
@end example

Putting this to use, here is a simple program to parse the data:

@example
@c file eg/misc/simple-csv.awk
@group
BEGIN @{
    FPAT = "([^,]+)|(\"[^\"]+\")"
@}
@end group

@group
@{
    print "NF = ", NF
    for (i = 1; i <= NF; i++) @{
        printf("$%d = <%s>\n", i, $i)
    @}
@}
@end group
@c endfile
@end example

When run, we get the following:

@example
$ @kbd{gawk -f simple-csv.awk addresses.csv}
NF =  7
$1 = <Robbins>
$2 = <Arnold>
$3 = <"1234 A Pretty Street, NE">
$4 = <MyTown>
$5 = <MyState>
$6 = <12345-6789>
$7 = <USA>
@end example

Note the embedded comma in the value of @code{$3}.

A straightforward improvement when processing CSV data of this sort
would be to remove the quotes when they occur, with something like this:

@example
if (substr($i, 1, 1) == "\"") @{
    len = length($i)
    $i = substr($i, 2, len - 2)    # Get text within the two quotes
@}
@end example

As with @code{FS}, the @code{IGNORECASE} variable (@pxref{User-modified})
affects field splitting with @code{FPAT}.

Assigning a value to @code{FPAT} overrides field splitting
with @code{FS} and with @code{FIELDWIDTHS}.

@quotation NOTE
Some programs export CSV data that contains embedded newlines between
the double quotes.  @command{gawk} provides no way to deal with this.
Even though a formal specification for CSV data exists, there isn't much
more to be done;
the @code{FPAT} mechanism provides an elegant solution for the majority
of cases, and the @command{gawk} developers are satisfied with that.
@end quotation

As written, the regexp used for @code{FPAT} requires that each field
contain at least one character.  A straightforward modification
(changing the first @samp{+} to @samp{*}) allows fields to be empty:

@example
FPAT = "([^,]*)|(\"[^\"]+\")"
@end example

@c FIXME: 4/2015
@c Consider use of FPAT = "([^,]*)|(\"[^\"]*\")"
@c (star in latter part of value) to allow quoted strings to be empty.
@c Per email from Ed Morton <mortoneccc@comcast.net>

Finally, the @code{patsplit()} function makes the same functionality
available for splitting regular strings (@pxref{String Functions}).


@node Testing field creation
@section Checking How @command{gawk} Is Splitting Records

@cindex @command{gawk}, splitting fields and
As we've seen, @command{gawk} provides three independent methods to split
input records into fields.  The mechanism used is based on which of the
three variables---@code{FS}, @code{FIELDWIDTHS}, or @code{FPAT}---was
last assigned to. In addition, an API input parser may choose to override
the record parsing mechanism; please refer to @ref{Input Parsers} for
further information about this feature.

To restore normal field splitting after using @code{FIELDWIDTHS}
and/or @code{FPAT}, simply assign a value to @code{FS}.
You can use @samp{FS = FS} to do this,
without having to know the current value of @code{FS}.

In order to tell which kind of field splitting is in effect,
use @code{PROCINFO["FS"]} (@pxref{Auto-set}).
The value is @code{"FS"} if regular field splitting is being used,
@code{"FIELDWIDTHS"} if fixed-width field splitting is being used,
or @code{"FPAT"} if content-based field splitting is being used:

@example
if (PROCINFO["FS"] == "FS")
    @var{regular field splitting} @dots{}
else if (PROCINFO["FS"] == "FIELDWIDTHS")
    @var{fixed-width field splitting} @dots{}
else if (PROCINFO["FS"] == "FPAT")
    @var{content-based field splitting} @dots{}
else
    @var{API input parser field splitting} @dots{} @ii{(advanced feature)}
@end example

This information is useful when writing a function that needs to
temporarily change @code{FS} or @code{FIELDWIDTHS}, read some records,
and then restore the original settings (@pxref{Passwd Functions} for an
example of such a function).

@node Multiple Line
@section Multiple-Line Records

@cindex multiple-line records
@cindex records, multiline
@cindex input, multiline records
@cindex files, reading, multiline records
@cindex input, files, See input files
In some databases, a single line cannot conveniently hold all the
information in one entry.  In such cases, you can use multiline
records.  The first step in doing this is to choose your data format.

@cindex record separators, with multiline records
One technique is to use an unusual character or string to separate
records.  For example, you could use the formfeed character (written
@samp{\f} in @command{awk}, as in C) to separate them, making each record
a page of the file.  To do this, just set the variable @code{RS} to
@code{"\f"} (a string containing the formfeed character).  Any
other character could equally well be used, as long as it won't be part
of the data in a record.

@cindex @code{RS} variable, multiline records and
Another technique is to have blank lines separate records.  By a special
dispensation, an empty string as the value of @code{RS} indicates that
records are separated by one or more blank lines.  When @code{RS} is set
to the empty string, each record always ends at the first blank line
encountered.  The next record doesn't start until the first nonblank
line that follows.  No matter how many blank lines appear in a row, they
all act as one record separator.
(Blank lines must be completely empty; lines that contain only
whitespace do not count.)

@cindex leftmost longest match
@cindex matching, leftmost longest
You can achieve the same effect as @samp{RS = ""} by assigning the
string @code{"\n\n+"} to @code{RS}. This regexp matches the newline
at the end of the record and one or more blank lines after the record.
In addition, a regular expression always matches the longest possible
sequence when there is a choice
(@pxref{Leftmost Longest}).
So, the next record doesn't start until
the first nonblank line that follows---no matter how many blank lines
appear in a row, they are considered one record separator.

@cindex dark corner, multiline records
However, there is an important difference between @samp{RS = ""} and
@samp{RS = "\n\n+"}. In the first case, leading newlines in the input
@value{DF} are ignored, and if a file ends without extra blank lines
after the last record, the final newline is removed from the record.
In the second case, this special processing is not done.
@value{DARKCORNER}

@cindex field separator, in multiline records
@cindex @code{FS}, in multiline records
Now that the input is separated into records, the second step is to
separate the fields in the records.  One way to do this is to divide each
of the lines into fields in the normal manner.  This happens by default
as the result of a special feature.  When @code{RS} is set to the empty
string @emph{and} @code{FS} is set to a single character,
the newline character @emph{always} acts as a field separator.
This is in addition to whatever field separations result from
@code{FS}.@footnote{When @code{FS} is the null string (@code{""})
or a regexp, this special feature of @code{RS} does not apply.
It does apply to the default field separator of a single space:
@samp{FS = @w{" "}}.}

The original motivation for this special exception was probably to provide
useful behavior in the default case (i.e., @code{FS} is equal
to @w{@code{" "}}).  This feature can be a problem if you really don't
want the newline character to separate fields, because there is no way to
prevent it.  However, you can work around this by using the @code{split()}
function to break up the record manually
(@pxref{String Functions}).
If you have a single-character field separator, you can work around
the special feature in a different way, by making @code{FS} into a
regexp for that single character.  For example, if the field
separator is a percent character, instead of
@samp{FS = "%"}, use @samp{FS = "[%]"}.

Another way to separate fields is to
put each field on a separate line: to do this, just set the
variable @code{FS} to the string @code{"\n"}.
(This single-character separator matches a single newline.)
A practical example of a @value{DF} organized this way might be a mailing
list, where blank lines separate the entries.  Consider a mailing
list in a file named @file{addresses}, which looks like this:

@example
Jane Doe
123 Main Street
Anywhere, SE 12345-6789

John Smith
456 Tree-lined Avenue
Smallville, MW 98765-4321
@dots{}
@end example

@noindent
A simple program to process this file is as follows:

@example
# addrs.awk --- simple mailing list program

# Records are separated by blank lines.
# Each line is one field.
BEGIN @{ RS = "" ; FS = "\n" @}

@{
      print "Name is:", $1
      print "Address is:", $2
      print "City and State are:", $3
      print ""
@}
@end example

Running the program produces the following output:

@example
$ @kbd{awk -f addrs.awk addresses}
@print{} Name is: Jane Doe
@print{} Address is: 123 Main Street
@print{} City and State are: Anywhere, SE 12345-6789
@print{}
@print{} Name is: John Smith
@print{} Address is: 456 Tree-lined Avenue
@print{} City and State are: Smallville, MW 98765-4321
@print{}
@dots{}
@end example

@xref{Labels Program} for a more realistic program dealing with
address lists.  The following list summarizes how records are split,
based on the value of
@ifinfo
@code{RS}.
(@samp{==} means ``is equal to.'')
@end ifinfo
@ifnotinfo
@code{RS}:
@end ifnotinfo

@table @code
@item RS == "\n"
Records are separated by the newline character (@samp{\n}).  In effect,
every line in the @value{DF} is a separate record, including blank lines.
This is the default.

@item RS == @var{any single character}
Records are separated by each occurrence of the character.  Multiple
successive occurrences delimit empty records.

@item RS == ""
Records are separated by runs of blank lines.
When @code{FS} is a single character, then
the newline character
always serves as a field separator, in addition to whatever value
@code{FS} may have. Leading and trailing newlines in a file are ignored.

@item RS == @var{regexp}
Records are separated by occurrences of characters that match @var{regexp}.
Leading and trailing matches of @var{regexp} delimit empty records.
(This is a @command{gawk} extension; it is not specified by the
POSIX standard.)
@end table

@cindex @command{gawk}, @code{RT} variable in
@cindex @code{RT} variable
If not in compatibility mode (@pxref{Options}), @command{gawk} sets
@code{RT} to the input text that matched the value specified by @code{RS}.
But if the input file ended without any text that matches @code{RS},
then @command{gawk} sets @code{RT} to the null string.

@node Getline
@section Explicit Input with @code{getline}

@cindex @code{getline} command, explicit input with
@cindex input, explicit
So far we have been getting our input data from @command{awk}'s main
input stream---either the standard input (usually your keyboard, sometimes
the output from another program) or the
files specified on the command line.  The @command{awk} language has a
special built-in command called @code{getline} that
can be used to read input under your explicit control.

The @code{getline} command is used in several different ways and should
@emph{not} be used by beginners.
The examples that follow the explanation of the @code{getline} command
include material that has not been covered yet.  Therefore, come back
and study the @code{getline} command @emph{after} you have reviewed the
rest of
@ifinfo
this @value{DOCUMENT}
@end ifinfo
@ifhtml
this @value{DOCUMENT}
@end ifhtml
@ifnotinfo
@ifnothtml
Parts I and II
@end ifnothtml
@end ifnotinfo
and have a good knowledge of how @command{awk} works.

@cindex @command{gawk}, @code{ERRNO} variable in
@cindex @code{ERRNO} variable, with @command{getline} command
@cindex differences in @command{awk} and @command{gawk}, @code{getline} command
@cindex @code{getline} command, return values
@cindex @option{--sandbox} option, input redirection with @code{getline}

The @code{getline} command returns 1 if it finds a record and 0 if
it encounters the end of the file.  If there is some error in getting
a record, such as a file that cannot be opened, then @code{getline}
returns @minus{}1.  In this case, @command{gawk} sets the variable
@code{ERRNO} to a string describing the error that occurred.

If @code{ERRNO} indicates that the I/O operation may be
retried, and @code{PROCINFO["@var{input}", "RETRY"]} is set,
then @code{getline} returns @minus{}2
instead of @minus{}1, and further calls to @code{getline}
may be attempted.  @xref{Retrying Input} for further information about
this feature.

In the following examples, @var{command} stands for a string value that
represents a shell command.

@quotation NOTE
When @option{--sandbox} is specified (@pxref{Options}),
reading lines from files, pipes, and coprocesses is disabled.
@end quotation

@menu
* Plain Getline::               Using @code{getline} with no arguments.
* Getline/Variable::            Using @code{getline} into a variable.
* Getline/File::                Using @code{getline} from a file.
* Getline/Variable/File::       Using @code{getline} into a variable from a
                                file.
* Getline/Pipe::                Using @code{getline} from a pipe.
* Getline/Variable/Pipe::       Using @code{getline} into a variable from a
                                pipe.
* Getline/Coprocess::           Using @code{getline} from a coprocess.
* Getline/Variable/Coprocess::  Using @code{getline} into a variable from a
                                coprocess.
* Getline Notes::               Important things to know about @code{getline}.
* Getline Summary::             Summary of @code{getline} Variants.
@end menu

@node Plain Getline
@subsection Using @code{getline} with No Arguments

The @code{getline} command can be used without arguments to read input
from the current input file.  All it does in this case is read the next
input record and split it up into fields.  This is useful if you've
finished processing the current record, but want to do some special
processing on the next record @emph{right now}.  For example:

@example
# Remove text between /* and */, inclusive
@{
    if ((i = index($0, "/*")) != 0) @{
        out = substr($0, 1, i - 1)  # leading part of the string
        rest = substr($0, i + 2)    # ... */ ...
        j = index(rest, "*/")       # is */ in trailing part?
        if (j > 0) @{
            rest = substr(rest, j + 2)  # remove comment
        @} else @{
            while (j == 0) @{
                # get more text
                if (getline <= 0) @{
                    print("unexpected EOF or error:", ERRNO) > "/dev/stderr"
                    exit
                @}
                # build up the line using string concatenation
                rest = rest $0
                j = index(rest, "*/")   # is */ in trailing part?
                if (j != 0) @{
                    rest = substr(rest, j + 2)
                    break
                @}
            @}
        @}
        # build up the output line using string concatenation
        $0 = out rest
    @}
    print $0
@}
@end example

@c 8/2014: Here is some sample input:
@ignore
mon/*comment*/key
rab/*commen
t*/bit
horse /*comment*/more text
part 1 /*comment*/part 2 /*comment*/part 3
no comment
@end ignore

This @command{awk} program deletes C-style comments (@samp{/* @dots{}
*/}) from the input.
It uses a number of features we haven't covered yet, including
string concatenation
(@pxref{Concatenation})
and the @code{index()} and @code{substr()} built-in
functions
(@pxref{String Functions}).
By replacing the @samp{print $0} with other
statements, you could perform more complicated processing on the
decommented input, such as searching for matches of a regular
expression.  (This program has a subtle problem---it does not work if one
comment ends and another begins on the same line.)

This form of the @code{getline} command sets @code{NF},
@code{NR}, @code{FNR}, @code{RT}, and the value of @code{$0}.

@quotation NOTE
The new value of @code{$0} is used to test
the patterns of any subsequent rules.  The original value
of @code{$0} that triggered the rule that executed @code{getline}
is lost.
By contrast, the @code{next} statement reads a new record
but immediately begins processing it normally, starting with the first
rule in the program.  @xref{Next Statement}.
@end quotation

@node Getline/Variable
@subsection Using @code{getline} into a Variable
@cindex @code{getline} into a variable
@cindex variables, @code{getline} command into@comma{} using

You can use @samp{getline @var{var}} to read the next record from
@command{awk}'s input into the variable @var{var}.  No other processing is
done.
For example, suppose the next line is a comment or a special string,
and you want to read it without triggering
any rules.  This form of @code{getline} allows you to read that line
and store it in a variable so that the main
read-a-line-and-check-each-rule loop of @command{awk} never sees it.
The following example swaps every two lines of input:

@example
@group
@{
     if ((getline tmp) > 0) @{
          print tmp
          print $0
     @} else
          print $0
@}
@end group
@end example

@noindent
It takes the following list:

@example
wan
tew
free
phore
@end example

@noindent
and produces these results:

@example
tew
wan
phore
free
@end example

The @code{getline} command used in this way sets only the variables
@code{NR}, @code{FNR}, and @code{RT} (and, of course, @var{var}).
The record is not
split into fields, so the values of the fields (including @code{$0}) and
the value of @code{NF} do not change.

@node Getline/File
@subsection Using @code{getline} from a File

@cindex @code{getline} from a file
@cindex input redirection
@cindex redirection of input
@cindex @code{<} (left angle bracket), @code{<} operator (I/O)
@cindex left angle bracket (@code{<}), @code{<} operator (I/O)
@cindex operators, input/output
Use @samp{getline < @var{file}} to read the next record from @var{file}.
Here, @var{file} is a string-valued expression that
specifies the @value{FN}.  @samp{< @var{file}} is called a @dfn{redirection}
because it directs input to come from a different place.
For example, the following
program reads its input record from the file @file{secondary.input} when it
encounters a first field with a value equal to 10 in the current input
file:

@example
@{
    if ($1 == 10) @{
         getline < "secondary.input"
         print
    @} else
         print
@}
@end example

Because the main input stream is not used, the values of @code{NR} and
@code{FNR} are not changed. However, the record it reads is split into fields in
the normal manner, so the values of @code{$0} and the other fields are
changed, resulting in a new value of @code{NF}.
@code{RT} is also set.

@cindex POSIX @command{awk}, @code{<} operator and
@c Thanks to Paul Eggert for initial wording here
According to POSIX, @samp{getline < @var{expression}} is ambiguous if
@var{expression} contains unparenthesized operators other than
@samp{$}; for example, @samp{getline < dir "/" file} is ambiguous
because the concatenation operator (not discussed yet; @pxref{Concatenation})
is not parenthesized.  You should write it as @samp{getline < (dir "/" file)} if
you want your program to be portable to all @command{awk} implementations.

@node Getline/Variable/File
@subsection Using @code{getline} into a Variable from a File
@cindex variables, @code{getline} command into@comma{} using

Use @samp{getline @var{var} < @var{file}} to read input
from the file
@var{file}, and put it in the variable @var{var}.  As earlier, @var{file}
is a string-valued expression that specifies the file from which to read.

In this version of @code{getline}, none of the predefined variables are
changed and the record is not split into fields.  The only variable
changed is @var{var}.@footnote{This is not quite true. @code{RT} could
be changed if @code{RS} is a regular expression.}
For example, the following program copies all the input files to the
output, except for records that say @w{@samp{@@include @var{filename}}}.
Such a record is replaced by the contents of the file
@var{filename}:

@example
@{
     if (NF == 2 && $1 == "@@include") @{
          while ((getline line < $2) > 0)
               print line
          close($2)
     @} else
          print
@}
@end example

Note here how the name of the extra input file is not built into
the program; it is taken directly from the data, specifically from the second field on
the @code{@@include} line.

The @code{close()} function is called to ensure that if two identical
@code{@@include} lines appear in the input, the entire specified file is
included twice.
@xref{Close Files And Pipes}.

One deficiency of this program is that it does not process nested
@code{@@include} statements
(i.e., @code{@@include} statements in included files)
the way a true macro preprocessor would.
@xref{Igawk Program} for a program
that does handle nested @code{@@include} statements.

@node Getline/Pipe
@subsection Using @code{getline} from a Pipe

@c From private email, dated October 2, 1988. Used by permission, March 2013.
@cindex Kernighan, Brian
@quotation
@i{Omniscience has much to recommend it.
Failing that, attention to details would be useful.}
@author Brian Kernighan
@end quotation

@cindex @code{|} (vertical bar), @code{|} operator (I/O)
@cindex vertical bar (@code{|}), @code{|} operator (I/O)
@cindex input pipeline
@cindex pipe, input
@cindex operators, input/output
The output of a command can also be piped into @code{getline}, using
@samp{@var{command} | getline}.  In
this case, the string @var{command} is run as a shell command and its output
is piped into @command{awk} to be used as input.  This form of @code{getline}
reads one record at a time from the pipe.
For example, the following program copies its input to its output, except for
lines that begin with @samp{@@execute}, which are replaced by the output
produced by running the rest of the line as a shell command:

@example
@group
@{
     if ($1 == "@@execute") @{
          tmp = substr($0, 10)        # Remove "@@execute"
          while ((tmp | getline) > 0)
               print
          close(tmp)
     @} else
          print
@}
@end group
@end example

@noindent
The @code{close()} function is called to ensure that if two identical
@samp{@@execute} lines appear in the input, the command is run for
each one.
@ifnottex
@ifnotdocbook
@xref{Close Files And Pipes}.
@end ifnotdocbook
@end ifnottex
@c This example is unrealistic, since you could just use system
Given the input:

@example
foo
bar
baz
@@execute who
bletch
@end example

@noindent
the program might produce:

@cindex Robbins, Bill
@cindex Robbins, Miriam
@cindex Robbins, Arnold
@example
foo
bar
baz
arnold     ttyv0   Jul 13 14:22
miriam     ttyp0   Jul 13 14:23     (murphy:0)
bill       ttyp1   Jul 13 14:23     (murphy:0)
bletch
@end example

@noindent
Notice that this program ran the command @command{who} and printed the result.
(If you try this program yourself, you will of course get different results,
depending upon who is logged in on your system.)

This variation of @code{getline} splits the record into fields, sets the
value of @code{NF}, and recomputes the value of @code{$0}.  The values of
@code{NR} and @code{FNR} are not changed.
@code{RT} is set.

@cindex POSIX @command{awk}, @code{|} I/O operator and
@c Thanks to Paul Eggert for initial wording here
According to POSIX, @samp{@var{expression} | getline} is ambiguous if
@var{expression} contains unparenthesized operators other than
@samp{$}---for example, @samp{@w{"echo "} "date" | getline} is ambiguous
because the concatenation operator is not parenthesized.  You should
write it as @samp{(@w{"echo "} "date") | getline} if you want your program
to be portable to all @command{awk} implementations.

@cindex Brian Kernighan's @command{awk}
@cindex @command{mawk} utility
@quotation NOTE
Unfortunately, @command{gawk} has not been consistent in its treatment
of a construct like @samp{@w{"echo "} "date" | getline}.
Most versions, including the current version, treat it at as
@samp{@w{("echo "} "date") | getline}.
(This is also how BWK @command{awk} behaves.)
Some versions instead treat it as
@samp{@w{"echo "} ("date" | getline)}.
(This is how @command{mawk} behaves.)
In short, @emph{always} use explicit parentheses, and then you won't
have to worry.
@end quotation

@node Getline/Variable/Pipe
@subsection Using @code{getline} into a Variable from a Pipe
@cindex variables, @code{getline} command into@comma{} using

When you use @samp{@var{command} | getline @var{var}}, the
output of @var{command} is sent through a pipe to
@code{getline} and into the variable @var{var}.  For example, the
following program reads the current date and time into the variable
@code{current_time}, using the @command{date} utility, and then
prints it:

@example
BEGIN @{
     "date" | getline current_time
     close("date")
     print "Report printed on " current_time
@}
@end example

In this version of @code{getline}, none of the predefined variables are
changed and the record is not split into fields. However, @code{RT} is set.

@ifinfo
@c Thanks to Paul Eggert for initial wording here
According to POSIX, @samp{@var{expression} | getline @var{var}} is ambiguous if
@var{expression} contains unparenthesized operators other than
@samp{$}; for example, @samp{@w{"echo "} "date" | getline @var{var}} is ambiguous
because the concatenation operator is not parenthesized. You should
write it as @samp{(@w{"echo "} "date") | getline @var{var}} if you want your
program to be portable to other @command{awk} implementations.
@end ifinfo

@node Getline/Coprocess
@subsection Using @code{getline} from a Coprocess
@cindex coprocesses, @code{getline} from
@cindex @code{getline} command, coprocesses@comma{} using from
@cindex @code{|} (vertical bar), @code{|&} operator (I/O)
@cindex vertical bar (@code{|}), @code{|&} operator (I/O)
@cindex operators, input/output
@cindex differences in @command{awk} and @command{gawk}, input/output operators

Reading input into @code{getline} from a pipe is a one-way operation.
The command that is started with @samp{@var{command} | getline} only
sends data @emph{to} your @command{awk} program.

On occasion, you might want to send data to another program
for processing and then read the results back.
@command{gawk} allows you to start a @dfn{coprocess}, with which two-way
communications are possible.  This is done with the @samp{|&}
operator.
Typically, you write data to the coprocess first and then
read the results back, as shown in the following:

@example
print "@var{some query}" |& "db_server"
"db_server" |& getline
@end example

@noindent
which sends a query to @command{db_server} and then reads the results.

The values of @code{NR} and
@code{FNR} are not changed,
because the main input stream is not used.
However, the record is split into fields in
the normal manner, thus changing the values of @code{$0}, of the other fields,
and of @code{NF} and @code{RT}.

Coprocesses are an advanced feature. They are discussed here only because
this is the @value{SECTION} on @code{getline}.
@xref{Two-way I/O},
where coprocesses are discussed in more detail.

@node Getline/Variable/Coprocess
@subsection Using @code{getline} into a Variable from a Coprocess
@cindex variables, @code{getline} command into@comma{} using

When you use @samp{@var{command} |& getline @var{var}}, the output from
the coprocess @var{command} is sent through a two-way pipe to @code{getline}
and into the variable @var{var}.

In this version of @code{getline}, none of the predefined variables are
changed and the record is not split into fields.  The only variable
changed is @var{var}.
However, @code{RT} is set.

@ifinfo
Coprocesses are an advanced feature. They are discussed here only because
this is the @value{SECTION} on @code{getline}.
@xref{Two-way I/O},
where coprocesses are discussed in more detail.
@end ifinfo

@node Getline Notes
@subsection Points to Remember About @code{getline}
Here are some miscellaneous points about @code{getline} that
you should bear in mind:

@itemize @value{BULLET}
@item
When @code{getline} changes the value of @code{$0} and @code{NF},
@command{awk} does @emph{not} automatically jump to the start of the
program and start testing the new record against every pattern.
However, the new record is tested against any subsequent rules.

@cindex differences in @command{awk} and @command{gawk}, implementation limitations
@cindex implementation issues, @command{gawk}, limits
@cindex @command{awk}, implementations, limits
@cindex @command{gawk}, implementation issues, limits
@item
Some very old @command{awk} implementations limit the number of pipelines that an @command{awk}
program may have open to just one.  In @command{gawk}, there is no such limit.
You can open as many pipelines (and coprocesses) as the underlying operating
system permits.

@cindex side effects, @code{FILENAME} variable
@cindex @code{FILENAME} variable, @code{getline}@comma{} setting with
@cindex dark corner, @code{FILENAME} variable
@cindex @code{getline} command, @code{FILENAME} variable and
@cindex @code{BEGIN} pattern, @code{getline} and
@item
An interesting side effect occurs if you use @code{getline} without a
redirection inside a @code{BEGIN} rule. Because an unredirected @code{getline}
reads from the command-line @value{DF}s, the first @code{getline} command
causes @command{awk} to set the value of @code{FILENAME}. Normally,
@code{FILENAME} does not have a value inside @code{BEGIN} rules, because you
have not yet started to process the command-line @value{DF}s.
@value{DARKCORNER}
(See @ref{BEGIN/END};
also @pxref{Auto-set}.)

@item
Using @code{FILENAME} with @code{getline}
(@samp{getline < FILENAME})
is likely to be a source of
confusion.  @command{awk} opens a separate input stream from the
current input file.  However, by not using a variable, @code{$0}
and @code{NF} are still updated.  If you're doing this, it's
probably by accident, and you should reconsider what it is you're
trying to accomplish.

@item
@ifdocbook
The next @value{SECTION}
@end ifdocbook
@ifnotdocbook
@ref{Getline Summary},
@end ifnotdocbook
presents a table summarizing the
@code{getline} variants and which variables they can affect.
It is worth noting that those variants that do not use redirection
can cause @code{FILENAME} to be updated if they cause
@command{awk} to start reading a new input file.

@item
@cindex Moore, Duncan
If the variable being assigned is an expression with side effects,
different versions of @command{awk} behave differently upon encountering
end-of-file.  Some versions don't evaluate the expression; many versions
(including @command{gawk}) do.  Here is an example, courtesy of Duncan Moore:

@ignore
Date: Sun, 01 Apr 2012 11:49:33 +0100
From: Duncan Moore <duncan.moore@@gmx.com>
@end ignore

@example
BEGIN @{
    system("echo 1 > f")
    while ((getline a[++c] < "f") > 0) @{ @}
    print c
@}
@end example

@noindent
Here, the side effect is the @samp{++c}.  Is @code{c} incremented if
end-of-file is encountered before the element in @code{a} is assigned?

@command{gawk} treats @code{getline} like a function call, and evaluates
the expression @samp{a[++c]} before attempting to read from @file{f}.
However, some versions of @command{awk} only evaluate the expression once they
know that there is a string value to be assigned.
@end itemize

@node Getline Summary
@subsection Summary of @code{getline} Variants
@cindex @code{getline} command, variants

@ref{table-getline-variants}
summarizes the eight variants of @code{getline},
listing which predefined variables are set by each one,
and whether the variant is standard or a @command{gawk} extension.
Note: for each variant, @command{gawk} sets the @code{RT} predefined variable.

@float Table,table-getline-variants
@caption{@code{getline} variants and what they set}
@multitable @columnfractions .33 .38 .27
@headitem Variant @tab Effect @tab @command{awk} / @command{gawk}
@item @code{getline} @tab Sets @code{$0}, @code{NF}, @code{FNR}, @code{NR}, and @code{RT} @tab @command{awk}
@item @code{getline} @var{var} @tab Sets @var{var}, @code{FNR}, @code{NR}, and @code{RT} @tab @command{awk}
@item @code{getline <} @var{file} @tab Sets @code{$0}, @code{NF}, and @code{RT} @tab @command{awk}
@item @code{getline @var{var} < @var{file}} @tab Sets @var{var} and @code{RT} @tab @command{awk}
@item @var{command} @code{| getline} @tab Sets @code{$0}, @code{NF}, and @code{RT} @tab @command{awk}
@item @var{command} @code{| getline} @var{var} @tab Sets @var{var} and @code{RT} @tab @command{awk}
@item @var{command} @code{|& getline} @tab Sets @code{$0}, @code{NF}, and @code{RT} @tab @command{gawk}
@item @var{command} @code{|& getline} @var{var} @tab Sets @var{var} and @code{RT} @tab @command{gawk}
@end multitable
@end float

@node Read Timeout
@section Reading Input with a Timeout
@cindex timeout, reading input

@cindex differences in @command{awk} and @command{gawk}, read timeouts
This @value{SECTION} describes a feature that is specific to @command{gawk}.

You may specify a timeout in milliseconds for reading input from the keyboard,
a pipe, or two-way communication, including TCP/IP sockets. This can be done
on a per-input, per-command, or per-connection basis, by setting a special
element in the @code{PROCINFO} array (@pxref{Auto-set}):

@example
PROCINFO["input_name", "READ_TIMEOUT"] = @var{timeout in milliseconds}
@end example

When set, this causes @command{gawk} to time out and return failure
if no data is available to read within the specified timeout period.
For example, a TCP client can decide to give up on receiving
any response from the server after a certain amount of time:

@example
@group
Service = "/inet/tcp/0/localhost/daytime"
PROCINFO[Service, "READ_TIMEOUT"] = 100
if ((Service |& getline) > 0)
    print $0
else if (ERRNO != "")
    print ERRNO
@end group
@end example

Here is how to read interactively from the user@footnote{This assumes
that standard input is the keyboard.} without waiting
for more than five seconds:

@example
PROCINFO["/dev/stdin", "READ_TIMEOUT"] = 5000
while ((getline < "/dev/stdin") > 0)
    print $0
@end example

@command{gawk} terminates the read operation if input does not
arrive after waiting for the timeout period, returns failure,
and sets @code{ERRNO} to an appropriate string value.
A negative or zero value for the timeout is the same as specifying
no timeout at all.

A timeout can also be set for reading from the keyboard in the implicit
loop that reads input records and matches them against patterns,
like so:

@example
$ @kbd{gawk 'BEGIN @{ PROCINFO["-", "READ_TIMEOUT"] = 5000 @}}
> @kbd{@{ print "You entered: " $0 @}'}
@kbd{gawk}
@print{} You entered: gawk
@end example

In this case, failure to respond within five seconds results in the following
error message:

@example
@error{} gawk: cmd. line:2: (FILENAME=- FNR=1) fatal: error reading input file `-': Connection timed out
@end example

The timeout can be set or changed at any time, and will take effect on the
next attempt to read from the input device. In the following example,
we start with a timeout value of one second, and progressively
reduce it by one-tenth of a second until we wait indefinitely
for the input to arrive:

@example
PROCINFO[Service, "READ_TIMEOUT"] = 1000
while ((Service |& getline) > 0) @{
    print $0
    PROCINFO[Service, "READ_TIMEOUT"] -= 100
@}
@end example

@quotation NOTE
You should not assume that the read operation will block
exactly after the tenth record has been printed. It is possible that
@command{gawk} will read and buffer more than one record's
worth of data the first time. Because of this, changing the value
of timeout like in the preceding example is not very useful.
@end quotation

If the @code{PROCINFO} element is not present and the
@env{GAWK_READ_TIMEOUT} environment variable exists,
@command{gawk} uses its value to initialize the timeout value.
The exclusive use of the environment variable to specify timeout
has the disadvantage of not being able to control it
on a per-command or per-connection basis.

@command{gawk} considers a timeout event to be an error even though
the attempt to read from the underlying device may
succeed in a later attempt. This is a limitation, and it also
means that you cannot use this to multiplex input from
two or more sources.  @xref{Retrying Input} for a way to enable 
later I/O attempts to succeed.

Assigning a timeout value prevents read operations from
blocking indefinitely. But bear in mind that there are other ways
@command{gawk} can stall waiting for an input device to be ready.
A network client can sometimes take a long time to establish
a connection before it can start reading any data,
or the attempt to open a FIFO special file for reading can block
indefinitely until some other process opens it for writing.

@node Retrying Input
@section Retrying Reads After Certain Input Errors
@cindex retrying input

@cindex differences in @command{awk} and @command{gawk}, retrying input
This @value{SECTION} describes a feature that is specific to @command{gawk}.

When @command{gawk} encounters an error while reading input, by
default @code{getline} returns @minus{}1, and subsequent attempts to
read from that file result in an end-of-file indication.  However, you
may optionally instruct @command{gawk} to allow I/O to be retried when
certain errors are encountered by setting a special element in
the @code{PROCINFO} array (@pxref{Auto-set}):

@example
PROCINFO["@var{input_name}", "RETRY"] = 1
@end example

When this element exists, @command{gawk} checks the value of the system
(C language)
@code{errno} variable when an I/O error occurs.  If @code{errno} indicates
a subsequent I/O attempt may succeed, @code{getline} instead returns
@minus{}2 and
further calls to @code{getline} may succeed.  This applies to the @code{errno}
values @code{EAGAIN}, @code{EWOULDBLOCK}, @code{EINTR}, or @code{ETIMEDOUT}.

This feature is useful in conjunction with
@code{PROCINFO["@var{input_name}", "READ_TIMEOUT"]} or situations where a file
descriptor has been configured to behave in a non-blocking fashion.

@node Command-line directories
@section Directories on the Command Line
@cindex differences in @command{awk} and @command{gawk}, command-line directories
@cindex directories, command-line
@cindex command line, directories on

According to the POSIX standard, files named on the @command{awk}
command line must be text files; it is a fatal error if they are not.
Most versions of @command{awk} treat a directory on the command line as
a fatal error.

By default, @command{gawk} produces a warning for a directory on the
command line, but otherwise ignores it.  This makes it easier to use
shell wildcards with your @command{awk} program:

@example
$ @kbd{gawk -f whizprog.awk *}        @ii{Directories could kill this program}
@end example

If either of the @option{--posix}
or @option{--traditional} options is given, then @command{gawk} reverts
to treating a directory on the command line as a fatal error.

@xref{Extension Sample Readdir} for a way to treat directories
as usable data from an @command{awk} program.

@node Input Summary
@section Summary

@itemize @value{BULLET}
@item
Input is split into records based on the value of @code{RS}.
The possibilities are as follows:

@multitable @columnfractions .25 .35 .40
@headitem Value of @code{RS} @tab Records are split on @dots{} @tab @command{awk} / @command{gawk}
@item Any single character @tab That character @tab @command{awk}
@item The empty string (@code{""}) @tab Runs of two or more newlines @tab @command{awk}
@item A regexp @tab Text that matches the regexp @tab @command{gawk}
@end multitable

@item
@code{FNR} indicates how many records have been read from the current input file;
@code{NR} indicates how many records have been read in total.

@item
@command{gawk} sets @code{RT} to the text matched by @code{RS}.

@item
After splitting the input into records, @command{awk} further splits
the records into individual fields, named @code{$1}, @code{$2}, and so
on. @code{$0} is the whole record, and @code{NF} indicates how many
fields there are.  The default way to split fields is between whitespace
characters.

@item
Fields may be referenced using a variable, as in @code{$NF}.  Fields
may also be assigned values, which causes the value of @code{$0} to be
recomputed when it is later referenced. Assigning to a field with a number
greater than @code{NF} creates the field and rebuilds the record, using
@code{OFS} to separate the fields.  Incrementing @code{NF} does the same
thing. Decrementing @code{NF} throws away fields and rebuilds the record.

@item
Field splitting is more complicated than record splitting:

@multitable @columnfractions .40 .40 .20
@headitem Field separator value @tab Fields are split @dots{} @tab @command{awk} / @command{gawk}
@item @code{FS == " "} @tab On runs of whitespace @tab @command{awk}
@item @code{FS == @var{any single character}} @tab On that character @tab @command{awk}
@item @code{FS == @var{regexp}} @tab On text matching the regexp @tab @command{awk}
@item @code{FS == ""}  @tab Such that each individual character is a separate field @tab @command{gawk}
@item @code{FIELDWIDTHS == @var{list of columns}} @tab Based on character position @tab @command{gawk}
@item @code{FPAT == @var{regexp}} @tab On the text surrounding text matching the regexp @tab @command{gawk}
@end multitable

@item
Using @samp{FS = "\n"} causes the entire record to be a single field
(assuming that newlines separate records).

@item
@code{FS} may be set from the command line using the @option{-F} option.
This can also be done using command-line variable assignment.

@item
Use @code{PROCINFO["FS"]} to see how fields are being split.

@item
Use @code{getline} in its various forms to read additional records
from the default input stream, from a file, or from a pipe or coprocess.

@item
Use @code{PROCINFO[@var{file}, "READ_TIMEOUT"]} to cause reads to time out
for @var{file}.

@item
Directories on the command line are fatal for standard @command{awk};
@command{gawk} ignores them if not in POSIX mode.

@end itemize

@c EXCLUDE START
@node Input Exercises
@section Exercises

@enumerate
@item
Using the @code{FIELDWIDTHS} variable (@pxref{Constant Size}),
write a program to read election data, where each record represents
one voter's votes.  Come up with a way to define which columns are
associated with each ballot item, and print the total votes,
including abstentions, for each item.

@item
@ref{Plain Getline}, presented a program to remove C-style
comments (@samp{/* @dots{} */}) from the input.  That program
does not work if one comment ends on one line and another one
starts later on the same line.
That can be fixed by making one simple change. What is it?

@end enumerate
@c EXCLUDE END

@node Printing
@chapter Printing Output

@cindex printing
@cindex output, printing, See printing
One of the most common programming actions is to @dfn{print}, or output,
some or all of the input.  Use the @code{print} statement
for simple output, and the @code{printf} statement
for fancier formatting.
The @code{print} statement is not limited when
computing @emph{which} values to print. However, with two exceptions,
you cannot specify @emph{how} to print them---how many
columns, whether to use exponential notation or not, and so on.
(For the exceptions, @pxref{Output Separators} and
@ref{OFMT}.)
For printing with specifications, you need the @code{printf} statement
(@pxref{Printf}).

@cindex @code{print} statement
@cindex @code{printf} statement
Besides basic and formatted printing, this @value{CHAPTER}
also covers I/O redirections to files and pipes, introduces
the special @value{FN}s that @command{gawk} processes internally,
and discusses the @code{close()} built-in function.

@menu
* Print::                       The @code{print} statement.
* Print Examples::              Simple examples of @code{print} statements.
* Output Separators::           The output separators and how to change them.
* OFMT::                        Controlling Numeric Output With @code{print}.
* Printf::                      The @code{printf} statement.
* Redirection::                 How to redirect output to multiple files and
                                pipes.
* Special FD::                  Special files for I/O.
* Special Files::               File name interpretation in @command{gawk}.
                                @command{gawk} allows access to inherited file
                                descriptors.
* Close Files And Pipes::       Closing Input and Output Files and Pipes.
* Nonfatal::                    Enabling Nonfatal Output.
* Output Summary::              Output summary.
* Output Exercises::            Exercises.
@end menu

@node Print
@section The @code{print} Statement

Use the @code{print} statement to produce output with simple, standardized
formatting.  You specify only the strings or numbers to print, in a
list separated by commas.  They are output, separated by single spaces,
followed by a newline.  The statement looks like this:

@example
print @var{item1}, @var{item2}, @dots{}
@end example

@noindent
The entire list of items may be optionally enclosed in parentheses.  The
parentheses are necessary if any of the item expressions uses the @samp{>}
relational operator; otherwise it could be confused with an output redirection
(@pxref{Redirection}).

The items to print can be constant strings or numbers, fields of the
current record (such as @code{$1}), variables, or any @command{awk}
expression.  Numeric values are converted to strings and then printed.

@cindex records, printing
@cindex lines, blank, printing
@cindex text, printing
The simple statement @samp{print} with no items is equivalent to
@samp{print $0}: it prints the entire current record.  To print a blank
line, use @samp{print ""}.
To print a fixed piece of text, use a string constant, such as
@w{@code{"Don't Panic"}}, as one item.  If you forget to use the
double-quote characters, your text is taken as an @command{awk}
expression, and you will probably get an error.  Keep in mind that a
space is printed between any two items.

Note that the @code{print} statement is a statement and not an
expression---you can't use it in the pattern part of a
pattern--action statement, for example.

@node Print Examples
@section @code{print} Statement Examples

Each @code{print} statement makes at least one line of output.  However, it
isn't limited to only one line.  If an item value is a string containing a
newline, the newline is output along with the rest of the string.  A
single @code{print} statement can make any number of lines this way.

@cindex newlines, printing
The following is an example of printing a string that contains embedded
@ifinfo
newlines
(the @samp{\n} is an escape sequence, used to represent the newline
character; @pxref{Escape Sequences}):
@end ifinfo
@ifhtml
newlines
(the @samp{\n} is an escape sequence, used to represent the newline
character; @pxref{Escape Sequences}):
@end ifhtml
@ifnotinfo
@ifnothtml
newlines:
@end ifnothtml
@end ifnotinfo

@example
@group
$ @kbd{awk 'BEGIN @{ print "line one\nline two\nline three" @}'}
@print{} line one
@print{} line two
@print{} line three
@end group
@end example

@cindex fields, printing
The next example, which is run on the @file{inventory-shipped} file,
prints the first two fields of each input record, with a space between
them:

@example
$ @kbd{awk '@{ print $1, $2 @}' inventory-shipped}
@print{} Jan 13
@print{} Feb 15
@print{} Mar 15
@dots{}
@end example

@cindex @code{print} statement, commas, omitting
@cindex troubleshooting, @code{print} statement@comma{} omitting commas
A common mistake in using the @code{print} statement is to omit the comma
between two items.  This often has the effect of making the items run
together in the output, with no space.  The reason for this is that
juxtaposing two string expressions in @command{awk} means to concatenate
them.  Here is the same program, without the comma:

@example
$ @kbd{awk '@{ print $1 $2 @}' inventory-shipped}
@print{} Jan13
@print{} Feb15
@print{} Mar15
@dots{}
@end example

@cindex @code{BEGIN} pattern, headings@comma{} adding
To someone unfamiliar with the @file{inventory-shipped} file, neither
example's output makes much sense.  A heading line at the beginning
would make it clearer.  Let's add some headings to our table of months
(@code{$1}) and green crates shipped (@code{$2}).  We do this using
a @code{BEGIN} rule (@pxref{BEGIN/END}) so that the headings are only
printed once:

@example
awk 'BEGIN @{  print "Month Crates"
              print "----- ------" @}
           @{  print $1, $2 @}' inventory-shipped
@end example

@noindent
When run, the program prints the following:

@example
Month Crates
----- ------
Jan 13
Feb 15
Mar 15
@dots{}
@end example

@noindent
The only problem, however, is that the headings and the table data
don't line up!  We can fix this by printing some spaces between the
two fields:

@example
@group
awk 'BEGIN @{ print "Month Crates"
             print "----- ------" @}
           @{ print $1, "     ", $2 @}' inventory-shipped
@end group
@end example

@cindex @code{printf} statement, columns@comma{} aligning
@cindex columns, aligning
Lining up columns this way can get pretty
complicated when there are many columns to fix.  Counting spaces for two
or three columns is simple, but any more than this can take up
a lot of time. This is why the @code{printf} statement was
created (@pxref{Printf});
one of its specialties is lining up columns of data.

@cindex line continuations, in @code{print} statement
@cindex @code{print} statement, line continuations and
@quotation NOTE
You can continue either a @code{print} or
@code{printf} statement simply by putting a newline after any comma
(@pxref{Statements/Lines}).
@end quotation

@node Output Separators
@section Output Separators

@cindex @code{OFS} variable
As mentioned previously, a @code{print} statement contains a list
of items separated by commas.  In the output, the items are normally
separated by single spaces.  However, this doesn't need to be the case;
a single space is simply the default.  Any string of
characters may be used as the @dfn{output field separator} by setting the
predefined variable @code{OFS}.  The initial value of this variable
is the string @w{@code{" "}} (i.e., a single space).

The output from an entire @code{print} statement is called an @dfn{output
record}.  Each @code{print} statement outputs one output record, and
then outputs a string called the @dfn{output record separator} (or
@code{ORS}).  The initial value of @code{ORS} is the string @code{"\n"}
(i.e., a newline character).  Thus, each @code{print} statement normally
makes a separate line.

@cindex output, records
@cindex output record separator, See @code{ORS} variable
@cindex @code{ORS} variable
@cindex @code{BEGIN} pattern, @code{OFS}/@code{ORS} variables, assigning values to
In order to change how output fields and records are separated, assign
new values to the variables @code{OFS} and @code{ORS}.  The usual
place to do this is in the @code{BEGIN} rule
(@pxref{BEGIN/END}), so
that it happens before any input is processed.  It can also be done
with assignments on the command line, before the names of the input
files, or using the @option{-v} command-line option
(@pxref{Options}).
The following example prints the first and second fields of each input
record, separated by a semicolon, with a blank line added after each
newline:


@example
$ @kbd{awk 'BEGIN @{ OFS = ";"; ORS = "\n\n" @}}
>            @kbd{@{ print $1, $2 @}' mail-list}
@print{} Amelia;555-5553
@print{}
@print{} Anthony;555-3412
@print{}
@print{} Becky;555-7685
@print{}
@print{} Bill;555-1675
@print{}
@print{} Broderick;555-0542
@print{}
@print{} Camilla;555-2912
@print{}
@print{} Fabius;555-1234
@print{}
@print{} Julie;555-6699
@print{}
@print{} Martin;555-6480
@print{}
@print{} Samuel;555-3430
@print{}
@print{} Jean-Paul;555-2127
@print{}
@end example

If the value of @code{ORS} does not contain a newline, the program's output
runs together on a single line.

@node OFMT
@section Controlling Numeric Output with @code{print}
@cindex numeric, output format
@cindex formats@comma{} numeric output
When printing numeric values with the @code{print} statement,
@command{awk} internally converts each number to a string of characters
and prints that string.  @command{awk} uses the @code{sprintf()} function
to do this conversion
(@pxref{String Functions}).
For now, it suffices to say that the @code{sprintf()}
function accepts a @dfn{format specification} that tells it how to format
numbers (or strings), and that there are a number of different ways in which
numbers can be formatted.  The different format specifications are discussed
more fully in
@ref{Control Letters}.

@cindexawkfunc{sprintf}
@cindex @code{OFMT} variable
@cindex output, format specifier@comma{} @code{OFMT}
The predefined variable @code{OFMT} contains the format specification
that @code{print} uses with @code{sprintf()} when it wants to convert a
number to a string for printing.
The default value of @code{OFMT} is @code{"%.6g"}.
The way @code{print} prints numbers can be changed
by supplying a different format specification
for the value of @code{OFMT}, as shown in the following example:

@example
$ @kbd{awk 'BEGIN @{}
>   @kbd{OFMT = "%.0f"  # print numbers as integers (rounds)}
>   @kbd{print 17.23, 17.54 @}'}
@print{} 17 18
@end example

@noindent
@cindex dark corner, @code{OFMT} variable
@cindex POSIX @command{awk}, @code{OFMT} variable and
@cindex @code{OFMT} variable, POSIX @command{awk} and
According to the POSIX standard, @command{awk}'s behavior is undefined
if @code{OFMT} contains anything but a floating-point conversion specification.
@value{DARKCORNER}

@node Printf
@section Using @code{printf} Statements for Fancier Printing

@cindex @code{printf} statement
@cindex output, formatted
@cindex formatting output
For more precise control over the output format than what is
provided by @code{print}, use @code{printf}.
With @code{printf} you can
specify the width to use for each item, as well as various
formatting choices for numbers (such as what output base to use, whether to
print an exponent, whether to print a sign, and how many digits to print
after the decimal point).

@menu
* Basic Printf::                Syntax of the @code{printf} statement.
* Control Letters::             Format-control letters.
* Format Modifiers::            Format-specification modifiers.
* Printf Examples::             Several examples.
@end menu

@node Basic Printf
@subsection Introduction to the @code{printf} Statement

@cindex @code{printf} statement, syntax of
A simple @code{printf} statement looks like this:

@example
printf @var{format}, @var{item1}, @var{item2}, @dots{}
@end example

@noindent
As for @code{print}, the entire list of arguments may optionally be
enclosed in parentheses. Here too, the parentheses are necessary if any
of the item expressions uses the @samp{>} relational operator; otherwise,
it can be confused with an output redirection (@pxref{Redirection}).

@cindex format specifiers
The difference between @code{printf} and @code{print} is the @var{format}
argument.  This is an expression whose value is taken as a string; it
specifies how to output each of the other arguments.  It is called the
@dfn{format string}.

The format string is very similar to that in the ISO C library function
@code{printf()}.  Most of @var{format} is text to output verbatim.
Scattered among this text are @dfn{format specifiers}---one per item.
Each format specifier says to output the next item in the argument list
at that place in the format.

The @code{printf} statement does not automatically append a newline
to its output.  It outputs only what the format string specifies.
So if a newline is needed, you must include one in the format string.
The output separator variables @code{OFS} and @code{ORS} have no effect
on @code{printf} statements. For example:

@example
@group
$ @kbd{awk 'BEGIN @{}
>    @kbd{ORS = "\nOUCH!\n"; OFS = "+"}
>    @kbd{msg = "Don\47t Panic!"}
>    @kbd{printf "%s\n", msg}
> @kbd{@}'}
@print{} Don't Panic!
@end group
@end example

@noindent
Here, neither the @samp{+} nor the @samp{OUCH!} appears in
the output message.

@node Control Letters
@subsection Format-Control Letters
@cindex @code{printf} statement, format-control characters
@cindex format specifiers, @code{printf} statement

A format specifier starts with the character @samp{%} and ends with
a @dfn{format-control letter}---it tells the @code{printf} statement
how to output one item.  The format-control letter specifies what @emph{kind}
of value to print.  The rest of the format specifier is made up of
optional @dfn{modifiers} that control @emph{how} to print the value, such as
the field width.  Here is a list of the format-control letters:

@c @asis for docbook to come out right
@table @asis
@item @code{%c}
Print a number as a character; thus, @samp{printf "%c",
65} outputs the letter @samp{A}. The output for a string value is
the first character of the string.

@cindex dark corner, format-control characters
@cindex @command{gawk}, format-control characters
@quotation NOTE
The POSIX standard says the first character of a string is printed.
In locales with multibyte characters, @command{gawk} attempts to
convert the leading bytes of the string into a valid wide character
and then to print the multibyte encoding of that character.
Similarly, when printing a numeric value, @command{gawk} allows the
value to be within the numeric range of values that can be held
in a wide character.
If the conversion to multibyte encoding fails, @command{gawk}
uses the low eight bits of the value as the character to print.

Other @command{awk} versions generally restrict themselves to printing
the first byte of a string or to numeric values within the range of
a single byte (0--255).
@value{DARKCORNER}
@end quotation


@item @code{%d}, @code{%i}
Print a decimal integer.
The two control letters are equivalent.
(The @samp{%i} specification is for compatibility with ISO C.)

@item @code{%e}, @code{%E}
Print a number in scientific (exponential) notation.
For example:

@example
printf "%4.3e\n", 1950
@end example

@noindent
prints @samp{1.950e+03}, with a total of four significant figures, three of
which follow the decimal point.
(The @samp{4.3} represents two modifiers,
discussed in the next @value{SUBSECTION}.)
@samp{%E} uses @samp{E} instead of @samp{e} in the output.

@item @code{%f}
Print a number in floating-point notation.
For example:

@example
printf "%4.3f", 1950
@end example

@noindent
prints @samp{1950.000}, with a total of four significant figures, three of
which follow the decimal point.
(The @samp{4.3} represents two modifiers,
discussed in the next @value{SUBSECTION}.)

On systems supporting IEEE 754 floating-point format, values
representing negative
infinity are formatted as
@samp{-inf} or @samp{-infinity},
and positive infinity as
@samp{inf} or @samp{infinity}.
The special ``not a number'' value formats as @samp{-nan} or @samp{nan}
(@pxref{Math Definitions}).

@item @code{%F}
Like @samp{%f}, but the infinity and ``not a number'' values are spelled
using uppercase letters.

The @samp{%F} format is a POSIX extension to ISO C; not all systems
support it.  On those that don't, @command{gawk} uses @samp{%f} instead.

@item @code{%g}, @code{%G}
Print a number in either scientific notation or in floating-point
notation, whichever uses fewer characters; if the result is printed in
scientific notation, @samp{%G} uses @samp{E} instead of @samp{e}.

@item @code{%o}
Print an unsigned octal integer
(@pxref{Nondecimal-numbers}).

@item @code{%s}
Print a string.

@item @code{%u}
Print an unsigned decimal integer.
(This format is of marginal use, because all numbers in @command{awk}
are floating point; it is provided primarily for compatibility with C.)

@item @code{%x}, @code{%X}
Print an unsigned hexadecimal integer;
@samp{%X} uses the letters @samp{A} through @samp{F}
instead of @samp{a} through @samp{f}
(@pxref{Nondecimal-numbers}).

@item @code{%%}
Print a single @samp{%}.
This does not consume an
argument and it ignores any modifiers.
@end table

@cindex dark corner, format-control characters
@cindex @command{gawk}, format-control characters
@quotation NOTE
When using the integer format-control letters for values that are
outside the range of the widest C integer type, @command{gawk} switches to
the @samp{%g} format specifier. If @option{--lint} is provided on the
command line (@pxref{Options}), @command{gawk}
warns about this.  Other versions of @command{awk} may print invalid
values or do something else entirely.
@value{DARKCORNER}
@end quotation

@node Format Modifiers
@subsection Modifiers for @code{printf} Formats

@cindex @code{printf} statement, modifiers
@cindex modifiers@comma{} in format specifiers
A format specification can also include @dfn{modifiers} that can control
how much of the item's value is printed, as well as how much space it gets.
The modifiers come between the @samp{%} and the format-control letter.
We use the bullet symbol ``@bullet{}'' in the following examples to
represent
spaces in the output. Here are the possible modifiers, in the order in
which they may appear:

@table @asis
@cindex differences in @command{awk} and @command{gawk}, @code{print}/@code{printf} statements
@cindex @code{printf} statement, positional specifiers
@c the code{} does NOT start a secondary
@cindex positional specifiers, @code{printf} statement
@item @code{@var{N}$}
An integer constant followed by a @samp{$} is a @dfn{positional specifier}.
Normally, format specifications are applied to arguments in the order
given in the format string.  With a positional specifier, the format
specification is applied to a specific argument, instead of what
would be the next argument in the list.  Positional specifiers begin
counting with one. Thus:

@example
printf "%s %s\n", "don't", "panic"
printf "%2$s %1$s\n", "panic", "don't"
@end example

@noindent
prints the famous friendly message twice.

At first glance, this feature doesn't seem to be of much use.
It is in fact a @command{gawk} extension, intended for use in translating
messages at runtime.
@xref{Printf Ordering},
which describes how and why to use positional specifiers.
For now, we ignore them.

@item @code{-} (Minus)
The minus sign, used before the width modifier (see later on in
this list),
says to left-justify
the argument within its specified width.  Normally, the argument
is printed right-justified in the specified width.  Thus:

@example
printf "%-4s", "foo"
@end example

@noindent
prints @samp{foo@bullet{}}.

@item @var{space}
For numeric conversions, prefix positive values with a space and
negative values with a minus sign.

@item @code{+}
The plus sign, used before the width modifier (see later on in
this list),
says to always supply a sign for numeric conversions, even if the data
to format is positive. The @samp{+} overrides the space modifier.

@item @code{#}
Use an ``alternative form'' for certain control letters.
For @samp{%o}, supply a leading zero.
For @samp{%x} and @samp{%X}, supply a leading @samp{0x} or @samp{0X} for
a nonzero result.
For @samp{%e}, @samp{%E}, @samp{%f}, and @samp{%F}, the result always
contains a decimal point.
For @samp{%g} and @samp{%G}, trailing zeros are not removed from the result.

@item @code{0}
A leading @samp{0} (zero) acts as a flag indicating that output should be
padded with zeros instead of spaces.
This applies only to the numeric output formats.
This flag only has an effect when the field width is wider than the
value to print.

@item @code{'}
A single quote or apostrophe character is a POSIX extension to ISO C.
It indicates that the integer part of a floating-point value, or the
entire part of an integer decimal value, should have a thousands-separator
character in it.  This only works in locales that support such characters.
For example:

@example
$ @kbd{cat thousands.awk}          @ii{Show source program}
@print{} BEGIN @{ printf "%'d\n", 1234567 @}
$ @kbd{LC_ALL=C gawk -f thousands.awk}
@print{} 1234567                   @ii{Results in} "C" @ii{locale}
$ @kbd{LC_ALL=en_US.UTF-8 gawk -f thousands.awk}
@print{} 1,234,567                 @ii{Results in US English UTF locale}
@end example

@noindent
For more information about locales and internationalization issues,
see @ref{Locales}.

@quotation NOTE
The @samp{'} flag is a nice feature, but its use complicates things: it
becomes difficult to use it in command-line programs.  For information
on appropriate quoting tricks, see @ref{Quoting}.
@end quotation

@item @var{width}
This is a number specifying the desired minimum width of a field.  Inserting any
number between the @samp{%} sign and the format-control character forces the
field to expand to this width.  The default way to do this is to
pad with spaces on the left.  For example:

@example
printf "%4s", "foo"
@end example

@noindent
prints @samp{@bullet{}foo}.

The value of @var{width} is a minimum width, not a maximum.  If the item
value requires more than @var{width} characters, it can be as wide as
necessary.  Thus, the following:

@example
printf "%4s", "foobar"
@end example

@noindent
prints @samp{foobar}.

Preceding the @var{width} with a minus sign causes the output to be
padded with spaces on the right, instead of on the left.

@item @code{.@var{prec}}
A period followed by an integer constant
specifies the precision to use when printing.
The meaning of the precision varies by control letter:

@table @asis
@item @code{%d}, @code{%i}, @code{%o}, @code{%u}, @code{%x}, @code{%X}
Minimum number of digits to print.

@item @code{%e}, @code{%E}, @code{%f}, @code{%F}
Number of digits to the right of the decimal point.

@item @code{%g}, @code{%G}
Maximum number of significant digits.

@item @code{%s}
Maximum number of characters from the string that should print.
@end table

Thus, the following:

@example
printf "%.4s", "foobar"
@end example

@noindent
prints @samp{foob}.
@end table

The C library @code{printf}'s dynamic @var{width} and @var{prec}
capability (e.g., @code{"%*.*s"}) is supported.  Instead of
supplying explicit @var{width} and/or @var{prec} values in the format
string, they are passed in the argument list.  For example:

@example
w = 5
p = 3
s = "abcdefg"
printf "%*.*s\n", w, p, s
@end example

@noindent
is exactly equivalent to:

@example
s = "abcdefg"
printf "%5.3s\n", s
@end example

@noindent
Both programs output @samp{@w{@bullet{}@bullet{}abc}}.
Earlier versions of @command{awk} did not support this capability.
If you must use such a version, you may simulate this feature by using
concatenation to build up the format string, like so:

@example
w = 5
p = 3
s = "abcdefg"
printf "%" w "." p "s\n", s
@end example

@noindent
This is not particularly easy to read, but it does work.

@c @cindex lint checks
@cindex troubleshooting, fatal errors, @code{printf} format strings
@cindex POSIX @command{awk}, @code{printf} format strings and
C programmers may be used to supplying additional modifiers (@samp{h},
@samp{j}, @samp{l}, @samp{L}, @samp{t}, and @samp{z}) in @code{printf}
format strings. These are not valid in @command{awk}.  Most @command{awk}
implementations silently ignore them.  If @option{--lint} is provided
on the command line (@pxref{Options}), @command{gawk} warns about their
use. If @option{--posix} is supplied, their use is a fatal error.

@node Printf Examples
@subsection Examples Using @code{printf}

The following simple example shows
how to use @code{printf} to make an aligned table:

@example
awk '@{ printf "%-10s %s\n", $1, $2 @}' mail-list
@end example

@noindent
This command
prints the names of the people (@code{$1}) in the file
@file{mail-list} as a string of 10 characters that are left-justified.  It also
prints the phone numbers (@code{$2}) next on the line.  This
produces an aligned two-column table of names and phone numbers,
as shown here:

@example
$ @kbd{awk '@{ printf "%-10s %s\n", $1, $2 @}' mail-list}
@print{} Amelia     555-5553
@print{} Anthony    555-3412
@print{} Becky      555-7685
@print{} Bill       555-1675
@print{} Broderick  555-0542
@print{} Camilla    555-2912
@print{} Fabius     555-1234
@print{} Julie      555-6699
@print{} Martin     555-6480
@print{} Samuel     555-3430
@print{} Jean-Paul  555-2127
@end example

In this case, the phone numbers had to be printed as strings because
the numbers are separated by dashes.  Printing the phone numbers as
numbers would have produced just the first three digits: @samp{555}.
This would have been pretty confusing.

It wasn't necessary to specify a width for the phone numbers because
they are last on their lines.  They don't need to have spaces
after them.

The table could be made to look even nicer by adding headings to the
tops of the columns.  This is done using a @code{BEGIN} rule
(@pxref{BEGIN/END})
so that the headers are only printed once, at the beginning of
the @command{awk} program:

@example
awk 'BEGIN @{ print "Name      Number"
             print "----      ------" @}
           @{ printf "%-10s %s\n", $1, $2 @}' mail-list
@end example

The preceding example mixes @code{print} and @code{printf} statements in
the same program.  Using just @code{printf} statements can produce the
same results:

@example
awk 'BEGIN @{ printf "%-10s %s\n", "Name", "Number"
             printf "%-10s %s\n", "----", "------" @}
           @{ printf "%-10s %s\n", $1, $2 @}' mail-list
@end example

@noindent
Printing each column heading with the same format specification
used for the column elements ensures that the headings
are aligned just like the columns.

The fact that the same format specification is used three times can be
emphasized by storing it in a variable, like this:

@example
awk 'BEGIN @{ format = "%-10s %s\n"
             printf format, "Name", "Number"
             printf format, "----", "------" @}
           @{ printf format, $1, $2 @}' mail-list
@end example


@node Redirection
@section Redirecting Output of @code{print} and @code{printf}

@cindex output redirection
@cindex redirection of output
@cindex @option{--sandbox} option, output redirection with @code{print}, @code{printf}
So far, the output from @code{print} and @code{printf} has gone
to the standard
output, usually the screen.  Both @code{print} and @code{printf} can
also send their output to other places.
This is called @dfn{redirection}.

@quotation NOTE
When @option{--sandbox} is specified (@pxref{Options}),
redirecting output to files, pipes, and coprocesses is disabled.
@end quotation

A redirection appears after the @code{print} or @code{printf} statement.
Redirections in @command{awk} are written just like redirections in shell
commands, except that they are written inside the @command{awk} program.

@c the commas here are part of the see also
@cindex @code{print} statement, See Also redirection@comma{} of output
@cindex @code{printf} statement, See Also redirection@comma{} of output
There are four forms of output redirection: output to a file, output
appended to a file, output through a pipe to another command, and output
to a coprocess.  We show them all for the @code{print} statement,
but they work identically for @code{printf}:

@table @code
@cindex @code{>} (right angle bracket), @code{>} operator (I/O)
@cindex right angle bracket (@code{>}), @code{>} operator (I/O)
@cindex operators, input/output
@item print @var{items} > @var{output-file}
This redirection prints the items into the output file named
@var{output-file}.  The @value{FN} @var{output-file} can be any
expression.  Its value is changed to a string and then used as a
@value{FN} (@pxref{Expressions}).

When this type of redirection is used, the @var{output-file} is erased
before the first output is written to it.  Subsequent writes to the same
@var{output-file} do not erase @var{output-file}, but append to it.
(This is different from how you use redirections in shell scripts.)
If @var{output-file} does not exist, it is created.  For example, here
is how an @command{awk} program can write a list of peoples' names to one
file named @file{name-list}, and a list of phone numbers to another file
named @file{phone-list}:

@example
$ @kbd{awk '@{ print $2 > "phone-list"}
>        @kbd{print $1 > "name-list" @}' mail-list}
$ @kbd{cat phone-list}
@print{} 555-5553
@print{} 555-3412
@dots{}
$ @kbd{cat name-list}
@print{} Amelia
@print{} Anthony
@dots{}
@end example

@noindent
Each output file contains one name or number per line.

@cindex @code{>} (right angle bracket), @code{>>} operator (I/O)
@cindex right angle bracket (@code{>}), @code{>>} operator (I/O)
@item print @var{items} >> @var{output-file}
This redirection prints the items into the preexisting output file
named @var{output-file}.  The difference between this and the
single-@samp{>} redirection is that the old contents (if any) of
@var{output-file} are not erased.  Instead, the @command{awk} output is
appended to the file.
If @var{output-file} does not exist, then it is created.

@cindex @code{|} (vertical bar), @code{|} operator (I/O)
@cindex pipe, output
@cindex output, pipes
@item print @var{items} | @var{command}
It is possible to send output to another program through a pipe
instead of into a file.   This redirection opens a pipe to
@var{command}, and writes the values of @var{items} through this pipe
to another process created to execute @var{command}.

The redirection argument @var{command} is actually an @command{awk}
expression.  Its value is converted to a string whose contents give
the shell command to be run.  For example, the following produces two
files, one unsorted list of peoples' names, and one list sorted in reverse
alphabetical order:

@ignore
10/2000:
This isn't the best style, since COMMAND is assigned for each
record.  It's done to avoid overfull hboxes in TeX.  Leave it
alone for now and let's hope no-one notices.
@end ignore

@example
@group
awk '@{ print $1 > "names.unsorted"
       command = "sort -r > names.sorted"
       print $1 | command @}' mail-list
@end group
@end example

The unsorted list is written with an ordinary redirection, while
the sorted list is written by piping through the @command{sort} utility.

The next example uses redirection to mail a message to the mailing
list @code{bug-system}.  This might be useful when trouble is encountered
in an @command{awk} script run periodically for system maintenance:

@example
report = "mail bug-system"
print("Awk script failed:", $0) | report
print("at record number", FNR, "of", FILENAME) | report
close(report)
@end example

The @code{close()} function is called here because it's a good idea to close
the pipe as soon as all the intended output has been sent to it.
@xref{Close Files And Pipes}
for more information.

This example also illustrates the use of a variable to represent
a @var{file} or @var{command}---it is not necessary to always
use a string constant.  Using a variable is generally a good idea,
because (if you mean to refer to that same file or command)
@command{awk} requires that the string value be written identically
every time.

@cindex coprocesses
@cindex @code{|} (vertical bar), @code{|&} operator (I/O)
@cindex operators, input/output
@cindex differences in @command{awk} and @command{gawk}, input/output operators
@item print @var{items} |& @var{command}
This redirection prints the items to the input of @var{command}.
The difference between this and the
single-@samp{|} redirection is that the output from @var{command}
can be read with @code{getline}.
Thus, @var{command} is a @dfn{coprocess}, which works together with
but is subsidiary to the @command{awk} program.

This feature is a @command{gawk} extension, and is not available in
POSIX @command{awk}.
@ifnotdocbook
@xref{Getline/Coprocess},
for a brief discussion.
@xref{Two-way I/O},
for a more complete discussion.
@end ifnotdocbook
@ifdocbook
@xref{Getline/Coprocess}
for a brief discussion and
@ref{Two-way I/O}
for a more complete discussion.
@end ifdocbook
@end table

Redirecting output using @samp{>}, @samp{>>}, @samp{|}, or @samp{|&}
asks the system to open a file, pipe, or coprocess only if the particular
@var{file} or @var{command} you specify has not already been written
to by your program or if it has been closed since it was last written to.

@cindex troubleshooting, printing
It is a common error to use @samp{>} redirection for the first @code{print}
to a file, and then to use @samp{>>} for subsequent output:

@example
# clear the file
print "Don't panic" > "guide.txt"
@dots{}
# append
print "Avoid improbability generators" >> "guide.txt"
@end example

@noindent
This is indeed how redirections must be used from the shell.  But in
@command{awk}, it isn't necessary.  In this kind of case, a program should
use @samp{>} for all the @code{print} statements, because the output file
is only opened once. (It happens that if you mix @samp{>} and @samp{>>}
output is produced in the expected order. However, mixing the operators
for the same file is definitely poor style, and is confusing to readers
of your program.)

@cindex differences in @command{awk} and @command{gawk}, implementation limitations
@cindex implementation issues, @command{gawk}, limits
@cindex @command{awk}, implementation issues, pipes
@cindex @command{gawk}, implementation issues, pipes
@ifnotinfo
As mentioned earlier
(@pxref{Getline Notes}),
many
@end ifnotinfo
@ifnottex
@ifnotdocbook
Many
@end ifnotdocbook
@end ifnottex
older
@command{awk} implementations limit the number of pipelines that an @command{awk}
program may have open to just one!  In @command{gawk}, there is no such limit.
@command{gawk} allows a program to
open as many pipelines as the underlying operating system permits.

@sidebar Piping into @command{sh}
@cindex shells, piping commands into

A particularly powerful way to use redirection is to build command lines
and pipe them into the shell, @command{sh}.  For example, suppose you
have a list of files brought over from a system where all the @value{FN}s
are stored in uppercase, and you wish to rename them to have names in
all lowercase.  The following program is both simple and efficient:

@c @cindex @command{mv} utility
@example
@{ printf("mv %s %s\n", $0, tolower($0)) | "sh" @}

END @{ close("sh") @}
@end example

The @code{tolower()} function returns its argument string with all
uppercase characters converted to lowercase
(@pxref{String Functions}).
The program builds up a list of command lines,
using the @command{mv} utility to rename the files.
It then sends the list to the shell for execution.

@xref{Shell Quoting} for a function that can help in generating
command lines to be fed to the shell.
@end sidebar

@node Special FD
@section Special Files for Standard Preopened Data Streams
@cindex standard input
@cindex input, standard
@cindex standard output
@cindex output, standard
@cindex error output
@cindex standard error
@cindex file descriptors
@cindex files, descriptors, See file descriptors

Running programs conventionally have three input and output streams
already available to them for reading and writing.  These are known
as the @dfn{standard input}, @dfn{standard output}, and @dfn{standard
error output}.  These open streams (and any other open files or pipes)
are often referred to by the technical term @dfn{file descriptors}.

These streams are, by default, connected to your keyboard and screen, but
they are often redirected with the shell, via the @samp{<}, @samp{<<},
@samp{>}, @samp{>>}, @samp{>&}, and @samp{|} operators.  Standard error
is typically used for writing error messages; the reason there are two separate
streams, standard output and standard error, is so that they can be
redirected separately.

@cindex differences in @command{awk} and @command{gawk}, error messages
@cindex error handling
In traditional implementations of @command{awk}, the only way to write an error
message to standard error in an @command{awk} program is as follows:

@example
print "Serious error detected!" | "cat 1>&2"
@end example

@noindent
This works by opening a pipeline to a shell command that can access the
standard error stream that it inherits from the @command{awk} process.
@c 8/2014: Mike Brennan says not to cite this as inefficient. So, fixed.
This is far from elegant, and it also requires a
separate process.  So people writing @command{awk} programs often
don't do this.  Instead, they send the error messages to the
screen, like this:

@example
print "Serious error detected!" > "/dev/tty"
@end example

@noindent
(@file{/dev/tty} is a special file supplied by the operating system
that is connected to your keyboard and screen. It represents the
``terminal,''@footnote{The ``tty'' in @file{/dev/tty} stands for
``Teletype,'' a serial terminal.} which on modern systems is a keyboard
and screen, not a serial console.)
This generally has the same effect, but not always: although the
standard error stream is usually the screen, it can be redirected; when
that happens, writing to the screen is not correct.  In fact, if
@command{awk} is run from a background job, it may not have a
terminal at all.
Then opening @file{/dev/tty} fails.

@command{gawk}, BWK @command{awk}, and @command{mawk} provide
special @value{FN}s for accessing the three standard streams.
If the @value{FN} matches one of these special names when @command{gawk}
(or one of the others) redirects input or output, then it directly uses
the descriptor that the @value{FN} stands for.  These special
@value{FN}s work for all operating systems that @command{gawk}
has been ported to, not just those that are POSIX-compliant:

@cindex common extensions, @code{/dev/stdin} special file
@cindex common extensions, @code{/dev/stdout} special file
@cindex common extensions, @code{/dev/stderr} special file
@cindex extensions, common@comma{} @code{/dev/stdin} special file
@cindex extensions, common@comma{} @code{/dev/stdout} special file
@cindex extensions, common@comma{} @code{/dev/stderr} special file
@cindex file names, standard streams in @command{gawk}
@cindex @code{/dev/@dots{}} special files
@cindex files, @code{/dev/@dots{}} special files
@cindex @code{/dev/fd/@var{N}} special files (@command{gawk})
@table @file
@item /dev/stdin
The standard input (file descriptor 0).

@item /dev/stdout
The standard output (file descriptor 1).

@item /dev/stderr
The standard error output (file descriptor 2).
@end table

With these facilities,
the proper way to write an error message then becomes:

@example
print "Serious error detected!" > "/dev/stderr"
@end example

@cindex troubleshooting, quotes with file names
Note the use of quotes around the @value{FN}.
Like with any other redirection, the value must be a string.
It is a common error to omit the quotes, which leads
to confusing results.

@command{gawk} does not treat these @value{FN}s as special when
in POSIX-compatibility mode. However, because BWK @command{awk}
supports them, @command{gawk} does support them even when
invoked with the @option{--traditional} option (@pxref{Options}).

@node Special Files
@section Special @value{FFN}s in @command{gawk}
@cindex @command{gawk}, file names in

Besides access to standard input, standard output, and standard error,
@command{gawk} provides access to any open file descriptor.
Additionally, there are special @value{FN}s reserved for
TCP/IP networking.

@menu
* Other Inherited Files::       Accessing other open files with
                                @command{gawk}.
* Special Network::             Special files for network communications.
* Special Caveats::             Things to watch out for.
@end menu

@node Other Inherited Files
@subsection Accessing Other Open Files with @command{gawk}

Besides the @code{/dev/stdin}, @code{/dev/stdout}, and @code{/dev/stderr}
special @value{FN}s mentioned earlier, @command{gawk} provides syntax
for accessing any other inherited open file:

@table @file
@item /dev/fd/@var{N}
The file associated with file descriptor @var{N}.  Such a file must
be opened by the program initiating the @command{awk} execution (typically
the shell).  Unless special pains are taken in the shell from which
@command{gawk} is invoked, only descriptors 0, 1, and 2 are available.
@end table

The @value{FN}s @file{/dev/stdin}, @file{/dev/stdout}, and @file{/dev/stderr}
are essentially aliases for @file{/dev/fd/0}, @file{/dev/fd/1}, and
@file{/dev/fd/2}, respectively. However, those names are more self-explanatory.

Note that using @code{close()} on a @value{FN} of the
form @code{"/dev/fd/@var{N}"}, for file descriptor numbers
above two, does actually close the given file descriptor.

@node Special Network
@subsection Special Files for Network Communications
@cindex networks, support for
@cindex TCP/IP, support for

@command{gawk} programs
can open a two-way
TCP/IP connection, acting as either a client or a server.
This is done using a special @value{FN} of the form:

@example
@file{/@var{net-type}/@var{protocol}/@var{local-port}/@var{remote-host}/@var{remote-port}}
@end example

The @var{net-type} is one of @samp{inet}, @samp{inet4}, or @samp{inet6}.
The @var{protocol} is one of @samp{tcp} or @samp{udp},
and the other fields represent the other essential pieces of information
for making a networking connection.
These @value{FN}s are used with the @samp{|&} operator for communicating
with @w{a coprocess}
(@pxref{Two-way I/O}).
This is an advanced feature, mentioned here only for completeness.
Full discussion is delayed until
@ref{TCP/IP Networking}.

@node Special Caveats
@subsection Special @value{FFN} Caveats

Here are some things to bear in mind when using the
special @value{FN}s that @command{gawk} provides:

@itemize @value{BULLET}
@cindex compatibility mode (@command{gawk}), file names
@cindex file names, in compatibility mode
@item
Recognition of the @value{FN}s for the three standard preopened
files is disabled only in POSIX mode.

@item
Recognition of the other special @value{FN}s is disabled if @command{gawk} is in
compatibility mode (either @option{--traditional} or @option{--posix};
@pxref{Options}).

@item
@command{gawk} @emph{always}
interprets these special @value{FN}s.
For example, using @samp{/dev/fd/4}
for output actually writes on file descriptor 4, and not on a new
file descriptor that is @code{dup()}ed from file descriptor 4.  Most of
the time this does not matter; however, it is important to @emph{not}
close any of the files related to file descriptors 0, 1, and 2.
Doing so results in unpredictable behavior.
@end itemize

@node Close Files And Pipes
@section Closing Input and Output Redirections
@cindex files, output, See output files
@cindex input files, closing
@cindex output, files@comma{} closing
@cindex pipe, closing
@cindex coprocesses, closing
@cindex @code{getline} command, coprocesses@comma{} using from

If the same @value{FN} or the same shell command is used with @code{getline}
more than once during the execution of an @command{awk} program
(@pxref{Getline}),
the file is opened (or the command is executed) the first time only.
At that time, the first record of input is read from that file or command.
The next time the same file or command is used with @code{getline},
another record is read from it, and so on.

Similarly, when a file or pipe is opened for output, @command{awk} remembers
the @value{FN} or command associated with it, and subsequent
writes to the same file or command are appended to the previous writes.
The file or pipe stays open until @command{awk} exits.

@cindexawkfunc{close}
This implies that special steps are necessary in order to read the same
file again from the beginning, or to rerun a shell command (rather than
reading more output from the same command).  The @code{close()} function
makes these things possible:

@example
close(@var{filename})
@end example

@noindent
or:

@example
close(@var{command})
@end example

The argument @var{filename} or @var{command} can be any expression.  Its
value must @emph{exactly} match the string that was used to open the file or
start the command (spaces and other ``irrelevant'' characters
included). For example, if you open a pipe with this:

@example
"sort -r names" | getline foo
@end example

@noindent
then you must close it with this:

@example
close("sort -r names")
@end example

Once this function call is executed, the next @code{getline} from that
file or command, or the next @code{print} or @code{printf} to that
file or command, reopens the file or reruns the command.
Because the expression that you use to close a file or pipeline must
exactly match the expression used to open the file or run the command,
it is good practice to use a variable to store the @value{FN} or command.
The previous example becomes the following:

@example
@group
sortcom = "sort -r names"
sortcom | getline foo
@end group
@group
@dots{}
close(sortcom)
@end group
@end example

@noindent
This helps avoid hard-to-find typographical errors in your @command{awk}
programs.  Here are some of the reasons for closing an output file:

@itemize @value{BULLET}
@item
To write a file and read it back later on in the same @command{awk}
program.  Close the file after writing it, then
begin reading it with @code{getline}.

@item
To write numerous files, successively, in the same @command{awk}
program.  If the files aren't closed, eventually @command{awk} may exceed a
system limit on the number of open files in one process.  It is best to
close each one when the program has finished writing it.

@item
To make a command finish.  When output is redirected through a pipe,
the command reading the pipe normally continues to try to read input
as long as the pipe is open.  Often this means the command cannot
really do its work until the pipe is closed.  For example, if
output is redirected to the @command{mail} program, the message is not
actually sent until the pipe is closed.

@item
To run the same program a second time, with the same arguments.
This is not the same thing as giving more input to the first run!

For example, suppose a program pipes output to the @command{mail} program.
If it outputs several lines redirected to this pipe without closing
it, they make a single message of several lines.  By contrast, if the
program closes the pipe after each line of output, then each line makes
a separate message.
@end itemize

@cindex differences in @command{awk} and @command{gawk}, @code{close()} function
@cindex portability, @code{close()} function and
@cindex @code{close()} function, portability
If you use more files than the system allows you to have open,
@command{gawk} attempts to multiplex the available open files among
your @value{DF}s.  @command{gawk}'s ability to do this depends upon the
facilities of your operating system, so it may not always work.  It is
therefore both good practice and good portability advice to always
use @code{close()} on your files when you are done with them.
In fact, if you are using a lot of pipes, it is essential that
you close commands when done. For example, consider something like this:

@example
@{
    @dots{}
    command = ("grep " $1 " /some/file | my_prog -q " $3)
    while ((command | getline) > 0) @{
        @var{process output of} command
    @}
    # need close(command) here
@}
@end example

This example creates a new pipeline based on data in @emph{each} record.
Without the call to @code{close()} indicated in the comment, @command{awk}
creates child processes to run the commands, until it eventually
runs out of file descriptors for more pipelines.

Even though each command has finished (as indicated by the end-of-file
return status from @code{getline}), the child process is not
terminated;@footnote{The technical terminology is rather morbid.
The finished child is called a ``zombie,'' and cleaning up after
it is referred to as ``reaping.''}
@c Good old UNIX: give the marketing guys fits, that's the ticket
more importantly, the file descriptor for the pipe
is not closed and released until @code{close()} is called or
@command{awk} exits.

@code{close()} silently does nothing if given an argument that
does not represent a file, pipe, or coprocess that was opened with
a redirection.  In such a case, it returns a negative value,
indicating an error. In addition, @command{gawk} sets @code{ERRNO}
to a string indicating the error.

Note also that @samp{close(FILENAME)} has no ``magic'' effects on the
implicit loop that reads through the files named on the command line.
It is, more likely, a close of a file that was never opened with a
redirection, so @command{awk} silently does nothing, except return
a negative value.

@cindex @code{|} (vertical bar), @code{|&} operator (I/O), pipes@comma{} closing
When using the @samp{|&} operator to communicate with a coprocess,
it is occasionally useful to be able to close one end of the two-way
pipe without closing the other.
This is done by supplying a second argument to @code{close()}.
As in any other call to @code{close()},
the first argument is the name of the command or special file used
to start the coprocess.
The second argument should be a string, with either of the values
@code{"to"} or @code{"from"}.  Case does not matter.
As this is an advanced feature, discussion is
delayed until
@ref{Two-way I/O},
which describes it in more detail and gives an example.

@sidebar Using @code{close()}'s Return Value
@cindex dark corner, @code{close()} function
@cindex @code{close()} function, return value
@cindex return value@comma{} @code{close()} function
@cindex differences in @command{awk} and @command{gawk}, @code{close()} function
@cindex Unix @command{awk}, @code{close()} function and

In many older versions of Unix @command{awk}, the @code{close()} function
is actually a statement.
@value{DARKCORNER}
It is a syntax error to try and use the return
value from @code{close()}:

@example
command = "@dots{}"
command | getline info
retval = close(command)  # syntax error in many Unix awks
@end example

@cindex @command{gawk}, @code{ERRNO} variable in
@cindex @code{ERRNO} variable, with @command{close()} function
@command{gawk} treats @code{close()} as a function.
The return value is @minus{}1 if the argument names something
that was never opened with a redirection, or if there is
a system problem closing the file or process.
In these cases, @command{gawk} sets the predefined variable
@code{ERRNO} to a string describing the problem.

In @command{gawk}, starting with @value{PVERSION} 4.2, when closing a pipe or
coprocess (input or output), the return value is the exit status of the
command, as described in @ref{table-close-pipe-return-values}.@footnote{Prior
to @value{PVERSION} 4.2, the return value from closing a pipe or co-process
was the full 16-bit exit value as defined by the @code{wait()} system
call.} Otherwise, it is the return value from the system's @code{close()}
or @code{fclose()} C functions when closing input or output files,
respectively.  This value is zero if the close succeeds, or @minus{}1
if it fails.

@float Table,table-close-pipe-return-values
@caption{Return values from @code{close()} of a pipe}
@multitable @columnfractions .50 .50
@headitem Situation @tab Return value from @code{close()}
@item Normal exit of command @tab Command's exit status
@item Death by signal of command @tab 256 + number of murderous signal
@item Death by signal of command with core dump @tab 512 + number of murderous signal
@item Some kind of error @tab @minus{}1
@end multitable
@end float

The POSIX standard is very vague; it says that @code{close()}
returns zero on success and a nonzero value otherwise.  In general,
different implementations vary in what they report when closing
pipes; thus, the return value cannot be used portably.
@value{DARKCORNER}
In POSIX mode (@pxref{Options}), @command{gawk} just returns zero
when closing a pipe.
@end sidebar

@node Nonfatal
@section Enabling Nonfatal Output

This @value{SECTION} describes a @command{gawk}-specific feature.

In standard @command{awk}, output with @code{print} or @code{printf}
to a nonexistent file, or some other I/O error (such as filling up the
disk) is a fatal error.

@example
$ @kbd{gawk 'BEGIN @{ print "hi" > "/no/such/file" @}'}
@error{} gawk: cmd. line:1: fatal: can't redirect to `/no/such/file' (No
@error{} such file or directory)
@end example

@command{gawk} makes it possible to detect that an error has
occurred, allowing you to possibly recover from the error, or
at least print an error message of your choosing before exiting.
You can do this in one of two ways:

@itemize @bullet
@item
For all output files, by assigning any value to @code{PROCINFO["NONFATAL"]}.

@item
On a per-file basis, by assigning any value to
@code{PROCINFO[@var{filename}, "NONFATAL"]}.
Here, @var{filename} is the name of the file to which
you wish output to be nonfatal.
@end itemize

Once you have enabled nonfatal output, you must check @code{ERRNO}
after every relevant @code{print} or @code{printf} statement to
see if something went wrong.  It is also a good idea to initialize
@code{ERRNO} to zero before attempting the output. For example:

@example
$ @kbd{gawk '}
> @kbd{BEGIN @{}
> @kbd{    PROCINFO["NONFATAL"] = 1}
> @kbd{    ERRNO = 0}
> @kbd{    print "hi" > "/no/such/file"}
> @kbd{    if (ERRNO) @{}
> @kbd{        print("Output failed:", ERRNO) > "/dev/stderr"}
> @kbd{        exit 1}
> @kbd{    @}}
> @kbd{@}'}
@error{} Output failed: No such file or directory
@end example

Here, @command{gawk} did not produce a fatal error; instead
it let the @command{awk} program code detect the problem and handle it.

This mechanism works also for standard output and standard error.
For standard output, you may use @code{PROCINFO["-", "NONFATAL"]}
or @code{PROCINFO["/dev/stdout", "NONFATAL"]}.  For standard error, use
@code{PROCINFO["/dev/stderr", "NONFATAL"]}.

When attempting to open a TCP/IP socket (@pxref{TCP/IP Networking}),
@command{gawk} tries multiple times. The @env{GAWK_SOCK_RETRIES}
environment variable (@pxref{Other Environment Variables}) allows you to
override @command{gawk}'s builtin default number of attempts.  However,
once nonfatal I/O is enabled for a given socket, @command{gawk} only
retries once, relying on @command{awk}-level code to notice that there
was a problem.

@node Output Summary
@section Summary

@itemize @value{BULLET}
@item
The @code{print} statement prints comma-separated expressions. Each
expression is separated by the value of @code{OFS} and terminated by
the value of @code{ORS}.  @code{OFMT} provides the conversion format
for numeric values for the @code{print} statement.

@item
The @code{printf} statement provides finer-grained control over output,
with format-control letters for different data types and various flags
that modify the behavior of the format-control letters.

@item
Output from both @code{print} and @code{printf} may be redirected to
files, pipes, and coprocesses.

@item
@command{gawk} provides special @value{FN}s for access to standard input,
output, and error, and for network communications.

@item
Use @code{close()} to close open file, pipe, and coprocess redirections.
For coprocesses, it is possible to close only one direction of the
communications.

@item
Normally errors with @code{print} or @code{printf} are fatal.
@command{gawk} lets you make output errors be nonfatal either for
all files or on a per-file basis. You must then check for errors
after every relevant output statement.

@end itemize

@c EXCLUDE START
@node Output Exercises
@section Exercises

@enumerate
@item
Rewrite the program:

@example
awk 'BEGIN @{ print "Month Crates"
             print "----- ------" @}
           @{ print $1, "     ", $2 @}' inventory-shipped
@end example

@noindent
from @ref{Output Separators}, by using a new value of @code{OFS}.

@item
Use the @code{printf} statement to line up the headings and table data
for the @file{inventory-shipped} example that was covered in @ref{Print}.

@item
What happens if you forget the double quotes when redirecting
output, as follows:

@example
BEGIN @{ print "Serious error detected!" > /dev/stderr @}
@end example

@end enumerate
@c EXCLUDE END


@node Expressions
@chapter Expressions
@cindex expressions

Expressions are the basic building blocks of @command{awk} patterns
and actions.  An expression evaluates to a value that you can print, test,
or pass to a function.  Additionally, an expression
can assign a new value to a variable or a field by using an assignment operator.

An expression can serve as a pattern or action statement on its own.
Most other kinds of
statements contain one or more expressions that specify the data on which to
operate.  As in other languages, expressions in @command{awk} can include
variables, array references, constants, and function calls, as well as
combinations of these with various operators.

@menu
* Values::                      Constants, Variables, and Regular Expressions.
* All Operators::               @command{gawk}'s operators.
* Truth Values and Conditions:: Testing for true and false.
* Function Calls::              A function call is an expression.
* Precedence::                  How various operators nest.
* Locales::                     How the locale affects things.
* Expressions Summary::         Expressions summary.
@end menu

@node Values
@section Constants, Variables, and Conversions

Expressions are built up from values and the operations performed
upon them. This @value{SECTION} describes the elementary objects
that provide the values used in expressions.

@menu
* Constants::                   String, numeric and regexp constants.
* Using Constant Regexps::      When and how to use a regexp constant.
* Variables::                   Variables give names to values for later use.
* Conversion::                  The conversion of strings to numbers and vice
                                versa.
@end menu

@node Constants
@subsection Constant Expressions

@cindex constants, types of

The simplest type of expression is the @dfn{constant}, which always has
the same value.  There are three types of constants: numeric,
string, and regular expression.

Each is used in the appropriate context when you need a data
value that isn't going to change.  Numeric constants can
have different forms, but are internally stored in an identical manner.

@menu
* Scalar Constants::            Numeric and string constants.
* Nondecimal-numbers::          What are octal and hex numbers.
* Regexp Constants::            Regular Expression constants.
@end menu

@node Scalar Constants
@subsubsection Numeric and String Constants

@cindex constants, numeric
@cindex numeric constants
A @dfn{numeric constant} stands for a number.  This number can be an
integer, a decimal fraction, or a number in scientific (exponential)
notation.@footnote{The internal representation of all numbers,
including integers, uses double-precision floating-point numbers.
On most modern systems, these are in IEEE 754 standard format.
@xref{Arbitrary Precision Arithmetic}, for much more information.}
Here are some examples of numeric constants that all
have the same value:

@example
105
1.05e+2
1050e-1
@end example

@cindex string constants
A @dfn{string constant} consists of a sequence of characters enclosed in
double quotation marks.  For example:

@example
"parrot"
@end example

@noindent
@cindex differences in @command{awk} and @command{gawk}, strings
@cindex strings, length limitations
represents the string whose contents are @samp{parrot}.  Strings in
@command{gawk} can be of any length, and they can contain any of the possible
eight-bit ASCII characters, including ASCII @sc{nul} (character code zero).
Other @command{awk}
implementations may have difficulty with some character codes.

@node Nondecimal-numbers
@subsubsection Octal and Hexadecimal Numbers
@cindex octal numbers
@cindex hexadecimal numbers
@cindex numbers, octal
@cindex numbers, hexadecimal

In @command{awk}, all numbers are in decimal (i.e., base 10).  Many other
programming languages allow you to specify numbers in other bases, often
octal (base 8) and hexadecimal (base 16).
In octal, the numbers go 0, 1, 2, 3, 4, 5, 6, 7, 10, 11, 12, and so on.
Just as @samp{11} in decimal is 1 times 10 plus 1, so
@samp{11} in octal is 1 times 8 plus 1. This equals 9 in decimal.
In hexadecimal, there are 16 digits. Because the everyday decimal
number system only has ten digits (@samp{0}--@samp{9}), the letters
@samp{a} through @samp{f} represent the rest.
(Case in the letters is usually irrelevant; hexadecimal @samp{a} and @samp{A}
have the same value.)
Thus, @samp{11} in
hexadecimal is 1 times 16 plus 1, which equals 17 in decimal.

Just by looking at plain @samp{11}, you can't tell what base it's in.
So, in C, C++, and other languages derived from C,
@c such as PERL, but we won't mention that....
there is a special notation to signify the base.
Octal numbers start with a leading @samp{0},
and hexadecimal numbers start with a leading @samp{0x} or @samp{0X}:

@table @code
@item 11
Decimal value 11

@item 011
Octal 11, decimal value 9

@item 0x11
Hexadecimal 11, decimal value 17
@end table

This example shows the difference:

@example
$ @kbd{gawk 'BEGIN @{ printf "%d, %d, %d\n", 011, 11, 0x11 @}'}
@print{} 9, 11, 17
@end example

Being able to use octal and hexadecimal constants in your programs is most
useful when working with data that cannot be represented conveniently as
characters or as regular numbers, such as binary data of various sorts.

@cindex @command{gawk}, octal numbers and
@cindex @command{gawk}, hexadecimal numbers and
@command{gawk} allows the use of octal and hexadecimal
constants in your program text.  However, such numbers in the input data
are not treated differently; doing so by default would break old
programs.
(If you really need to do this, use the @option{--non-decimal-data}
command-line option;
@pxref{Nondecimal Data}.)
If you have octal or hexadecimal data,
you can use the @code{strtonum()} function
(@pxref{String Functions})
to convert the data into a number.
Most of the time, you will want to use octal or hexadecimal constants
when working with the built-in bit-manipulation functions;
see @ref{Bitwise Functions}
for more information.

Unlike in some early C implementations, @samp{8} and @samp{9} are not
valid in octal constants.  For example, @command{gawk} treats @samp{018}
as decimal 18:

@example
$ @kbd{gawk 'BEGIN @{ print "021 is", 021 ; print 018 @}'}
@print{} 021 is 17
@print{} 18
@end example

@cindex compatibility mode (@command{gawk}), octal numbers
@cindex compatibility mode (@command{gawk}), hexadecimal numbers
Octal and hexadecimal source code constants are a @command{gawk} extension.
If @command{gawk} is in compatibility mode
(@pxref{Options}),
they are not available.

@sidebar A Constant's Base Does Not Affect Its Value

Once a numeric constant has
been converted internally into a number,
@command{gawk} no longer remembers
what the original form of the constant was; the internal value is
always used.  This has particular consequences for conversion of
numbers to strings:

@example
$ @kbd{gawk 'BEGIN @{ printf "0x11 is <%s>\n", 0x11 @}'}
@print{} 0x11 is <17>
@end example
@end sidebar

@node Regexp Constants
@subsubsection Regular Expression Constants

@cindex regexp constants
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
A @dfn{regexp constant} is a regular expression description enclosed in
slashes, such as @code{@w{/^beginning and end$/}}.  Most regexps used in
@command{awk} programs are constant, but the @samp{~} and @samp{!~}
matching operators can also match computed or dynamic regexps
(which are typically just ordinary strings or variables that contain a regexp,
but could be more complex expressions).

@node Using Constant Regexps
@subsection Using Regular Expression Constants

Regular expression constants consist of text describing
a regular expression enclosed in slashes (such as @code{/the +answer/}).
This @value{SECTION} describes how such constants work in
POSIX @command{awk} and @command{gawk}, and then goes on to describe
@dfn{strongly typed regexp constants}, which are a @command{gawk} extension.

@menu
* Standard Regexp Constants::   Regexp constants in standard @command{awk}.
* Strong Regexp Constants::     Strongly typed regexp constants.
@end menu

@node Standard Regexp Constants
@subsubsection Standard Regular Expression Constants

@cindex dark corner, regexp constants
When used on the righthand side of the @samp{~} or @samp{!~}
operators, a regexp constant merely stands for the regexp that is to be
matched.
However, regexp constants (such as @code{/foo/}) may be used like simple expressions.
When a
regexp constant appears by itself, it has the same meaning as if it appeared
in a pattern (i.e., @samp{($0 ~ /foo/)}).
@value{DARKCORNER}
@xref{Expression Patterns}.
This means that the following two code segments:

@example
if ($0 ~ /barfly/ || $0 ~ /camelot/)
    print "found"
@end example

@noindent
and:

@example
if (/barfly/ || /camelot/)
    print "found"
@end example

@noindent
are exactly equivalent.
One rather bizarre consequence of this rule is that the following
Boolean expression is valid, but does not do what its author probably
intended:

@example
# Note that /foo/ is on the left of the ~
if (/foo/ ~ $1) print "found foo"
@end example

@c @cindex automatic warnings
@c @cindex warnings, automatic
@cindex @command{gawk}, regexp constants and
@cindex regexp constants, in @command{gawk}
@noindent
This code is ``obviously'' testing @code{$1} for a match against the regexp
@code{/foo/}.  But in fact, the expression @samp{/foo/ ~ $1} really means
@samp{($0 ~ /foo/) ~ $1}.  In other words, first match the input record
against the regexp @code{/foo/}.  The result is either zero or one,
depending upon the success or failure of the match.  That result
is then matched against the first field in the record.
Because it is unlikely that you would ever really want to make this kind of
test, @command{gawk} issues a warning when it sees this construct in
a program.
Another consequence of this rule is that the assignment statement:

@example
matches = /foo/
@end example

@noindent
assigns either zero or one to the variable @code{matches}, depending
upon the contents of the current input record.

@cindex differences in @command{awk} and @command{gawk}, regexp constants
@cindex dark corner, regexp constants, as arguments to user-defined functions
@cindexgawkfunc{gensub}
@cindexawkfunc{sub}
@cindexawkfunc{gsub}
Constant regular expressions are also used as the first argument for
the @code{gensub()}, @code{sub()}, and @code{gsub()} functions, as the
second argument of the @code{match()} function,
and as the third argument of the @code{split()} and @code{patsplit()} functions
(@pxref{String Functions}).
Modern implementations of @command{awk}, including @command{gawk}, allow
the third argument of @code{split()} to be a regexp constant, but some
older implementations do not.
@value{DARKCORNER}
Because some built-in functions accept regexp constants as arguments,
confusion can arise when attempting to use regexp constants as arguments
to user-defined functions (@pxref{User-defined}).  For example:

@example
@group
function mysub(pat, repl, str, global)
@{
    if (global)
        gsub(pat, repl, str)
    else
        sub(pat, repl, str)
    return str
@}
@end group

@group
@{
    @dots{}
    text = "hi! hi yourself!"
    mysub(/hi/, "howdy", text, 1)
    @dots{}
@}
@end group
@end example

@c @cindex automatic warnings
@c @cindex warnings, automatic
In this example, the programmer wants to pass a regexp constant to the
user-defined function @code{mysub()}, which in turn passes it on to
either @code{sub()} or @code{gsub()}.  However, what really happens is that
the @code{pat} parameter is assigned a value of either one or zero, depending upon whether
or not @code{$0} matches @code{/hi/}.
@command{gawk} issues a warning when it sees a regexp constant used as
a parameter to a user-defined function, because passing a truth value in
this way is probably not what was intended.

@node Strong Regexp Constants
@subsubsection Strongly Typed Regexp Constants

This @value{SECTION} describes a @command{gawk}-specific feature.

As we saw in the previous @value{SECTION},
regexp constants (@code{/@dots{}/}) hold a strange position in the
@command{awk} language. In most contexts, they act like an expression:
@samp{$0 ~ /@dots{}/}. In other contexts, they denote only a regexp to
be matched. In no case are they really a ``first class citizen'' of the
language. That is, you cannot define a scalar variable whose type is
``regexp'' in the same sense that you can define a variable to be a
number or a string:

@example
num = 42        @ii{Numeric variable}
str = "hi"      @ii{String variable}
re = /foo/      @ii{Wrong!} re @ii{is the result of} $0 ~ /foo/
@end example

For a number of more advanced use cases,
it would be nice to have regexp constants that
are @dfn{strongly typed}; in other words, that denote a regexp useful
for matching, and not an expression.

@command{gawk} provides this feature.  A strongly typed regexp constant
looks almost like a regular regexp constant, except that it is preceded
by an @samp{@@} sign:

@example
re = @@/foo/     @ii{Regexp variable}
@end example

Strongly typed regexp constants @emph{cannot} be used everywhere that a
regular regexp constant can, because this would make the language even more
confusing.  Instead, you may use them only in certain contexts:

@itemize @bullet
@item
On the righthand side of the @samp{~} and @samp{!~} operators: @samp{some_var ~ @@/foo/}
(@pxref{Regexp Usage}).

@item
In the @code{case} part of a @code{switch} statement
(@pxref{Switch Statement}).

@item
As an argument to one of the built-in functions that accept regexp constants:
@code{gensub()},
@code{gsub()},
@code{match()},
@code{patsplit()},
@code{split()},
and
@code{sub()}
(@pxref{String Functions}).

@item
As a parameter in a call to a user-defined function
(@pxref{User-defined}).

@item
On the righthand side of an assignment to a variable: @samp{some_var = @@/foo/}.
In this case, the type of @code{some_var} is regexp. Additionally, @code{some_var}
can be used with @samp{~} and @samp{!~}, passed to one of the built-in functions
listed above, or passed as a parameter to a user-defined function.
@end itemize

You may use the @code{typeof()} built-in function
(@pxref{Type Functions})
to determine if a variable or function parameter is
a regexp variable.

The true power of this feature comes from the ability to create variables that
have regexp type. Such variables can be passed on to user-defined functions,
without the confusing aspects of computed regular expressions created from
strings or string constants. They may also be passed through indirect function
calls (@pxref{Indirect Calls})
and on to the built-in functions that accept regexp constants.

When used in numeric conversions, strongly typed regexp variables convert
to zero. When used in string conversions, they convert to the string
value of the original regexp text.

@node Variables
@subsection Variables

@cindex variables, user-defined
@cindex user-defined, variables
@dfn{Variables} are ways of storing values at one point in your program for
use later in another part of your program.  They can be manipulated
entirely within the program text, and they can also be assigned values
on the @command{awk} command line.

@menu
* Using Variables::             Using variables in your programs.
* Assignment Options::          Setting variables on the command line and a
                                summary of command-line syntax. This is an
                                advanced method of input.
@end menu

@node Using Variables
@subsubsection Using Variables in a Program

Variables let you give names to values and refer to them later.  Variables
have already been used in many of the examples.  The name of a variable
must be a sequence of letters, digits, or underscores, and it may not begin
with a digit.
Here, a @dfn{letter} is any one of the 52 upper- and lowercase
English letters.  Other characters that may be defined as letters
in non-English locales are not valid in variable names.
Case is significant in variable names; @code{a} and @code{A}
are distinct variables.

A variable name is a valid expression by itself; it represents the
variable's current value.  Variables are given new values with
@dfn{assignment operators}, @dfn{increment operators}, and
@dfn{decrement operators}
(@pxref{Assignment Ops}).
In addition, the @code{sub()} and @code{gsub()} functions can
change a variable's value, and the @code{match()}, @code{split()},
and @code{patsplit()} functions can change the contents of their
array parameters (@pxref{String Functions}).

@cindex variables, built-in
@cindex variables, initializing
A few variables have special built-in meanings, such as @code{FS} (the
field separator) and @code{NF} (the number of fields in the current input
record).  @xref{Built-in Variables} for a list of the predefined variables.
These predefined variables can be used and assigned just like all other
variables, but their values are also used or changed automatically by
@command{awk}.  All predefined variables' names are entirely uppercase.

Variables in @command{awk} can be assigned either numeric or string values.
The kind of value a variable holds can change over the life of a program.
By default, variables are initialized to the empty string, which
is zero if converted to a number.  There is no need to explicitly
initialize a variable in @command{awk},
which is what you would do in C and in most other traditional languages.

@node Assignment Options
@subsubsection Assigning Variables on the Command Line
@cindex variables, assigning on command line
@cindex command line, variables@comma{} assigning on

Any @command{awk} variable can be set by including a @dfn{variable assignment}
among the arguments on the command line when @command{awk} is invoked
(@pxref{Other Arguments}).
Such an assignment has the following form:

@example
@var{variable}=@var{text}
@end example

@cindex @option{-v} option
@noindent
With it, a variable is set either at the beginning of the
@command{awk} run or in between input files.
When the assignment is preceded with the @option{-v} option,
as in the following:

@example
-v @var{variable}=@var{text}
@end example

@noindent
the variable is set at the very beginning, even before the
@code{BEGIN} rules execute.  The @option{-v} option and its assignment
must precede all the @value{FN} arguments, as well as the program text.
(@xref{Options} for more information about
the @option{-v} option.)
Otherwise, the variable assignment is performed at a time determined by
its position among the input file arguments---after the processing of the
preceding input file argument.  For example:

@example
awk '@{ print $n @}' n=4 inventory-shipped n=2 mail-list
@end example

@noindent
prints the value of field number @code{n} for all input records.  Before
the first file is read, the command line sets the variable @code{n}
equal to four.  This causes the fourth field to be printed in lines from
@file{inventory-shipped}.  After the first file has finished,
but before the second file is started, @code{n} is set to two, so that the
second field is printed in lines from @file{mail-list}:

@example
$ @kbd{awk '@{ print $n @}' n=4 inventory-shipped n=2 mail-list}
@print{} 15
@print{} 24
@dots{}
@print{} 555-5553
@print{} 555-3412
@dots{}
@end example

@cindex dark corner, command-line arguments
Command-line arguments are made available for explicit examination by
the @command{awk} program in the @code{ARGV} array
(@pxref{ARGC and ARGV}).
@command{awk} processes the values of command-line assignments for escape
sequences
(@pxref{Escape Sequences}).
@value{DARKCORNER}

@node Conversion
@subsection Conversion of Strings and Numbers

Number-to-string and string-to-number conversion are generally
straightforward.  There can be subtleties to be aware of;
this @value{SECTION} discusses this important facet of @command{awk}.

@menu
* Strings And Numbers::         How @command{awk} Converts Between Strings And
                                Numbers.
* Locale influences conversions:: How the locale may affect conversions.
@end menu

@node Strings And Numbers
@subsubsection How @command{awk} Converts Between Strings and Numbers

@cindex converting, strings to numbers
@cindex strings, converting
@cindex numbers, converting
@cindex converting, numbers to strings
Strings are converted to numbers and numbers are converted to strings, if the context
of the @command{awk} program demands it.  For example, if the value of
either @code{foo} or @code{bar} in the expression @samp{foo + bar}
happens to be a string, it is converted to a number before the addition
is performed.  If numeric values appear in string concatenation, they
are converted to strings.  Consider the following:

@example
@group
two = 2; three = 3
print (two three) + 4
@end group
@end example

@noindent
This prints the (numeric) value 27.  The numeric values of
the variables @code{two} and @code{three} are converted to strings and
concatenated together.  The resulting string is converted back to the
number 23, to which 4 is then added.

@cindex null strings, converting numbers to strings
@cindex type conversion
If, for some reason, you need to force a number to be converted to a
string, concatenate that number with the empty string, @code{""}.
To force a string to be converted to a number, add zero to that string.
A string is converted to a number by interpreting any numeric prefix
of the string as numerals:
@code{"2.5"} converts to 2.5, @code{"1e3"} converts to 1,000, and @code{"25fix"}
has a numeric value of 25.
Strings that can't be interpreted as valid numbers convert to zero.

@cindex @code{CONVFMT} variable
The exact manner in which numbers are converted into strings is controlled
by the @command{awk} predefined variable @code{CONVFMT} (@pxref{Built-in Variables}).
Numbers are converted using the @code{sprintf()} function
with @code{CONVFMT} as the format
specifier
(@pxref{String Functions}).

@code{CONVFMT}'s default value is @code{"%.6g"}, which creates a value with
at most six significant digits.  For some applications, you might want to
change it to specify more precision.
On most modern machines,
17 digits is usually enough to capture a floating-point number's
value exactly.@footnote{Pathological cases can require up to
752 digits (!), but we doubt that you need to worry about this.}

@cindex dark corner, @code{CONVFMT} variable
Strange results can occur if you set @code{CONVFMT} to a string that doesn't
tell @code{sprintf()} how to format floating-point numbers in a useful way.
For example, if you forget the @samp{%} in the format, @command{awk} converts
all numbers to the same constant string.

As a special case, if a number is an integer, then the result of converting
it to a string is @emph{always} an integer, no matter what the value of
@code{CONVFMT} may be.  Given the following code fragment:

@example
CONVFMT = "%2.2f"
a = 12
b = a ""
@end example

@noindent
@code{b} has the value @code{"12"}, not @code{"12.00"}.
@value{DARKCORNER}

@sidebar Pre-POSIX @command{awk} Used @code{OFMT} for String Conversion
@cindex POSIX @command{awk}, @code{OFMT} variable and
@cindex @code{OFMT} variable
@cindex portability, new @command{awk} vs.@: old @command{awk}
@cindex @command{awk}, new vs.@: old, @code{OFMT} variable
Prior to the POSIX standard, @command{awk} used the value
of @code{OFMT} for converting numbers to strings.  @code{OFMT}
specifies the output format to use when printing numbers with @code{print}.
@code{CONVFMT} was introduced in order to separate the semantics of
conversion from the semantics of printing.  Both @code{CONVFMT} and
@code{OFMT} have the same default value: @code{"%.6g"}.  In the vast majority
of cases, old @command{awk} programs do not change their behavior.
@xref{Print} for more information on the @code{print} statement.
@end sidebar

@node Locale influences conversions
@subsubsection Locales Can Influence Conversion

Where you are can matter when it comes to converting between numbers and
strings.  The local character set and language---the @dfn{locale}---can
affect numeric formats.  In particular, for @command{awk} programs,
it affects the decimal point character and the thousands-separator
character.  The @code{"C"} locale, and most English-language locales,
use the period character (@samp{.}) as the decimal point and don't
have a thousands separator.  However, many (if not most) European and
non-English locales use the comma (@samp{,}) as the decimal point
character. European locales often use either a space or a period as
the thousands separator, if they have one.

@cindex dark corner, locale's decimal point character
The POSIX standard says that @command{awk} always uses the period as the decimal
point when reading the @command{awk} program source code, and for
command-line variable assignments (@pxref{Other Arguments}).  However,
when interpreting input data, for @code{print} and @code{printf} output,
and for number-to-string conversion, the local decimal point character
is used.  @value{DARKCORNER} In all cases, numbers in source code and
in input data cannot have a thousands separator.  Here are some examples
indicating the difference in behavior, on a GNU/Linux system:

@example
$ @kbd{export POSIXLY_CORRECT=1}                        @ii{Force POSIX behavior}
$ @kbd{gawk 'BEGIN @{ printf "%g\n", 3.1415927 @}'}
@print{} 3.14159
$ @kbd{LC_ALL=en_DK.utf-8 gawk 'BEGIN @{ printf "%g\n", 3.1415927 @}'}
@print{} 3,14159
$ @kbd{echo 4,321 | gawk '@{ print $1 + 1 @}'}
@print{} 5
$ @kbd{echo 4,321 | LC_ALL=en_DK.utf-8 gawk '@{ print $1 + 1 @}'}
@print{} 5,321
@end example

@noindent
The @code{en_DK.utf-8} locale is for English in Denmark, where the comma acts as
the decimal point separator.  In the normal @code{"C"} locale, @command{gawk}
treats @samp{4,321} as 4, while in the Danish locale, it's treated
as the full number including the fractional part, 4.321.

Some earlier versions of @command{gawk} fully complied with this aspect
of the standard.  However, many users in non-English locales complained
about this behavior, because their data used a period as the decimal
point, so the default behavior was restored to use a period as the
decimal point character.  You can use the @option{--use-lc-numeric}
option (@pxref{Options}) to force @command{gawk} to use the locale's
decimal point character.  (@command{gawk} also uses the locale's decimal
point character when in POSIX mode, either via @option{--posix} or the
@env{POSIXLY_CORRECT} environment variable, as shown previously.)

@ref{table-locale-affects} describes the cases in which the locale's decimal
point character is used and when a period is used. Some of these
features have not been described yet.

@float Table,table-locale-affects
@caption{Locale decimal point versus a period}
@multitable @columnfractions .15 .20 .45
@headitem Feature @tab Default @tab @option{--posix} or @option{--use-lc-numeric}
@item @code{%'g} @tab Use locale @tab Use locale
@item @code{%g} @tab Use period @tab Use locale
@item Input @tab Use period @tab Use locale
@item @code{strtonum()} @tab Use period @tab Use locale
@end multitable
@end float

Finally, modern-day formal standards and the IEEE standard floating-point
representation can have an unusual but important effect on the way
@command{gawk} converts some special string values to numbers.  The details
are presented in @ref{POSIX Floating Point Problems}.

@node All Operators
@section Operators: Doing Something with Values

This @value{SECTION} introduces the @dfn{operators} that make use
of the values provided by constants and variables.

@menu
* Arithmetic Ops::              Arithmetic operations (@samp{+}, @samp{-},
                                etc.)
* Concatenation::               Concatenating strings.
* Assignment Ops::              Changing the value of a variable or a field.
* Increment Ops::               Incrementing the numeric value of a variable.
@end menu

@node Arithmetic Ops
@subsection Arithmetic Operators
@cindex arithmetic operators
@cindex operators, arithmetic
@c @cindex addition
@c @cindex subtraction
@c @cindex multiplication
@c @cindex division
@c @cindex remainder
@c @cindex quotient
@c @cindex exponentiation

The @command{awk} language uses the common arithmetic operators when
evaluating expressions.  All of these arithmetic operators follow normal
precedence rules and work as you would expect them to.

The following example uses a file named @file{grades}, which contains
a list of student names as well as three test scores per student (it's
a small class):

@example
Pat   100 97 58
Sandy  84 72 93
Chris  72 92 89
@end example

@noindent
This program takes the file @file{grades} and prints the average
of the scores:

@example
$ @kbd{awk '@{ sum = $2 + $3 + $4 ; avg = sum / 3}
>        @kbd{print $1, avg @}' grades}
@print{} Pat 85
@print{} Sandy 83
@print{} Chris 84.3333
@end example

The following list provides the arithmetic operators in @command{awk},
in order from the highest precedence to the lowest:

@table @code
@cindex common extensions, @code{**} operator
@cindex extensions, common@comma{} @code{**} operator
@cindex POSIX @command{awk}, arithmetic operators and
@item @var{x} ^ @var{y}
@itemx @var{x} ** @var{y}
Exponentiation; @var{x} raised to the @var{y} power.  @samp{2 ^ 3} has
the value eight; the character sequence @samp{**} is equivalent to
@samp{^}. @value{COMMONEXT}

@item - @var{x}
Negation.

@item + @var{x}
Unary plus; the expression is converted to a number.

@item @var{x} * @var{y}
Multiplication.

@cindex troubleshooting, division
@cindex division
@item @var{x} / @var{y}
Division;  because all numbers in @command{awk} are floating-point
numbers, the result is @emph{not} rounded to an integer---@samp{3 / 4} has
the value 0.75.  (It is a common mistake, especially for C programmers,
to forget that @emph{all} numbers in @command{awk} are floating point,
and that division of integer-looking constants produces a real number,
not an integer.)

@item @var{x} % @var{y}
Remainder; further discussion is provided in the text, just
after this list.

@item @var{x} + @var{y}
Addition.

@item @var{x} - @var{y}
Subtraction.
@end table

Unary plus and minus have the same precedence,
the multiplication operators all have the same precedence, and
addition and subtraction have the same precedence.

@cindex differences in @command{awk} and @command{gawk}, trunc-mod operation
@cindex trunc-mod operation
When computing the remainder of @samp{@var{x} % @var{y}},
the quotient is rounded toward zero to an integer and
multiplied by @var{y}. This result is subtracted from @var{x};
this operation is sometimes known as ``trunc-mod.''  The following
relation always holds:

@example
b * int(a / b) + (a % b) == a
@end example

One possibly undesirable effect of this definition of remainder is that
@samp{@var{x} % @var{y}} is negative if @var{x} is negative.  Thus:

@example
-17 % 8 = -1
@end example

In other @command{awk} implementations, the signedness of the remainder
may be machine-dependent.
@c FIXME !!! what does posix say?

@cindex portability, @code{**} operator and
@cindex @code{*} (asterisk), @code{**} operator
@cindex asterisk (@code{*}), @code{**} operator
@quotation NOTE
The POSIX standard only specifies the use of @samp{^}
for exponentiation.
For maximum portability, do not use the @samp{**} operator.
@end quotation

@node Concatenation
@subsection String Concatenation
@cindex Kernighan, Brian
@quotation
@i{It seemed like a good idea at the time.}
@author Brian Kernighan
@end quotation

@cindex string operators
@cindex operators, string
@cindex concatenating
There is only one string operation: concatenation.  It does not have a
specific operator to represent it.  Instead, concatenation is performed by
writing expressions next to one another, with no operator.  For example:

@example
$ @kbd{awk '@{ print "Field number one: " $1 @}' mail-list}
@print{} Field number one: Amelia
@print{} Field number one: Anthony
@dots{}
@end example

Without the space in the string constant after the @samp{:}, the line
runs together.  For example:

@example
$ @kbd{awk '@{ print "Field number one:" $1 @}' mail-list}
@print{} Field number one:Amelia
@print{} Field number one:Anthony
@dots{}
@end example

@cindex troubleshooting, string concatenation
Because string concatenation does not have an explicit operator, it is
often necessary to ensure that it happens at the right time by using
parentheses to enclose the items to concatenate.  For example,
you might expect that the
following code fragment concatenates @code{file} and @code{name}:

@example
file = "file"
name = "name"
print "something meaningful" > file name
@end example

@cindex Brian Kernighan's @command{awk}
@cindex @command{mawk} utility
@noindent
This produces a syntax error with some versions of Unix
@command{awk}.@footnote{It happens that BWK
@command{awk}, @command{gawk}, and @command{mawk} all ``get it right,''
but you should not rely on this.}
It is necessary to use the following:

@example
print "something meaningful" > (file name)
@end example

@cindex order of evaluation, concatenation
@cindex evaluation order, concatenation
@cindex side effects
Parentheses should be used around concatenation in all but the
most common contexts, such as on the righthand side of @samp{=}.
Be careful about the kinds of expressions used in string concatenation.
In particular, the order of evaluation of expressions used for concatenation
is undefined in the @command{awk} language.  Consider this example:

@example
BEGIN @{
    a = "don't"
    print (a " " (a = "panic"))
@}
@end example

@noindent
It is not defined whether the second assignment to @code{a} happens
before or after the value of @code{a} is retrieved for producing the
concatenated value.  The result could be either @samp{don't panic},
or @samp{panic panic}.
@c see test/nasty.awk for a worse example

The precedence of concatenation, when mixed with other operators, is often
counter-intuitive.  Consider this example:

@ignore
> To: bug-gnu-utils@@gnu.org
> CC: arnold@@gnu.org
> Subject: gawk 3.0.4 bug with {print -12 " " -24}
> From: Russell Schulz <Russell_Schulz@locutus.ofB.ORG>
> Date: Tue, 8 Feb 2000 19:56:08 -0700
>
> gawk 3.0.4 on NT gives me:
>
> prompt> cat bad.awk
> BEGIN { print -12 " " -24; }
>
> prompt> gawk -f bad.awk
> -12-24
>
> when I would expect
>
> -12 -24
>
> I have not investigated the source, or other implementations.  The
> bug is there on my NT and DOS versions 2.15.6 .
@end ignore

@example
$ @kbd{awk 'BEGIN @{ print -12 " " -24 @}'}
@print{} -12-24
@end example

This ``obviously'' is concatenating @minus{}12, a space, and @minus{}24.
But where did the space disappear to?
The answer lies in the combination of operator precedences and
@command{awk}'s automatic conversion rules.  To get the desired result,
write the program this way:

@example
$ @kbd{awk 'BEGIN @{ print -12 " " (-24) @}'}
@print{} -12 -24
@end example

This forces @command{awk} to treat the @samp{-} on the @samp{-24} as unary.
Otherwise, it's parsed as follows:

@display
    @minus{}12 (@code{"@ "} @minus{} 24)
@result{} @minus{}12 (0 @minus{} 24)
@result{} @minus{}12 (@minus{}24)
@result{} @minus{}12@minus{}24
@end display

As mentioned earlier,
when mixing concatenation with other operators, @emph{parenthesize}.  Otherwise,
you're never quite sure what you'll get.

@node Assignment Ops
@subsection Assignment Expressions
@cindex assignment operators
@cindex operators, assignment
@cindex expressions, assignment
@cindex @code{=} (equals sign), @code{=} operator
@cindex equals sign (@code{=}), @code{=} operator
An @dfn{assignment} is an expression that stores a (usually different)
value into a variable.  For example, let's assign the value one to the variable
@code{z}:

@example
z = 1
@end example

After this expression is executed, the variable @code{z} has the value one.
Whatever old value @code{z} had before the assignment is forgotten.

Assignments can also store string values.  For example, the
following stores
the value @code{"this food is good"} in the variable @code{message}:

@example
thing = "food"
predicate = "good"
message = "this " thing " is " predicate
@end example

@noindent
@cindex side effects, assignment expressions
This also illustrates string concatenation.
The @samp{=} sign is called an @dfn{assignment operator}.  It is the
simplest assignment operator because the value of the righthand
operand is stored unchanged.
Most operators (addition, concatenation, and so on) have no effect
except to compute a value.  If the value isn't used, there's no reason to
use the operator.  An assignment operator is different; it does
produce a value, but even if you ignore it, the assignment still
makes itself felt through the alteration of the variable.  We call this
a @dfn{side effect}.

@cindex lvalues/rvalues
@cindex rvalues/lvalues
@cindex assignment operators, lvalues/rvalues
@cindex operators, assignment
The lefthand operand of an assignment need not be a variable
(@pxref{Variables}); it can also be a field
(@pxref{Changing Fields}) or
an array element (@pxref{Arrays}).
These are all called @dfn{lvalues},
which means they can appear on the lefthand side of an assignment operator.
The righthand operand may be any expression; it produces the new value
that the assignment stores in the specified variable, field, or array
element. (Such values are called @dfn{rvalues}.)

@cindex variables, types of
It is important to note that variables do @emph{not} have permanent types.
A variable's type is simply the type of whatever value was last assigned
to it.  In the following program fragment, the variable
@code{foo} has a numeric value at first, and a string value later on:

@example
@group
foo = 1
print foo
@end group
@group
foo = "bar"
print foo
@end group
@end example

@noindent
When the second assignment gives @code{foo} a string value, the fact that
it previously had a numeric value is forgotten.

String values that do not begin with a digit have a numeric value of
zero. After executing the following code, the value of @code{foo} is five:

@example
foo = "a string"
foo = foo + 5
@end example

@quotation NOTE
Using a variable as a number and then later as a string
can be confusing and is poor programming style.  The previous two examples
illustrate how @command{awk} works, @emph{not} how you should write your
programs!
@end quotation

An assignment is an expression, so it has a value---the same value that
is assigned.  Thus, @samp{z = 1} is an expression with the value one.
One consequence of this is that you can write multiple assignments together,
such as:

@example
x = y = z = 5
@end example

@noindent
This example stores the value five in all three variables
(@code{x}, @code{y}, and @code{z}).
It does so because the
value of @samp{z = 5}, which is five, is stored into @code{y} and then
the value of @samp{y = z = 5}, which is five, is stored into @code{x}.

Assignments may be used anywhere an expression is called for.  For
example, it is valid to write @samp{x != (y = 1)} to set @code{y} to one,
and then test whether @code{x} equals one.  But this style tends to make
programs hard to read; such nesting of assignments should be avoided,
except perhaps in a one-shot program.

@cindex @code{+} (plus sign), @code{+=} operator
@cindex plus sign (@code{+}), @code{+=} operator
Aside from @samp{=}, there are several other assignment operators that
do arithmetic with the old value of the variable.  For example, the
operator @samp{+=} computes a new value by adding the righthand value
to the old value of the variable.  Thus, the following assignment adds
five to the value of @code{foo}:

@example
foo += 5
@end example

@noindent
This is equivalent to the following:

@example
foo = foo + 5
@end example

@noindent
Use whichever makes the meaning of your program clearer.

There are situations where using @samp{+=} (or any assignment operator)
is @emph{not} the same as simply repeating the lefthand operand in the
righthand expression.  For example:

@cindex Rankin, Pat
@example
@group
# Thanks to Pat Rankin for this example
BEGIN  @{
    foo[rand()] += 5
    for (x in foo)
       print x, foo[x]
@end group

@group
    bar[rand()] = bar[rand()] + 5
    for (x in bar)
       print x, bar[x]
@}
@end group
@end example

@cindex operators, assignment, evaluation order
@cindex assignment operators, evaluation order
@noindent
The indices of @code{bar} are practically guaranteed to be different, because
@code{rand()} returns different values each time it is called.
(Arrays and the @code{rand()} function haven't been covered yet.
@xref{Arrays},
and
@ifnotdocbook
@pxref{Numeric Functions}
@end ifnotdocbook
@ifdocbook
@ref{Numeric Functions}
@end ifdocbook
for more information.)
This example illustrates an important fact about assignment
operators: the lefthand expression is only evaluated @emph{once}.

It is up to the implementation as to which expression is evaluated
first, the lefthand or the righthand.
Consider this example:

@example
i = 1
a[i += 2] = i + 1
@end example

@noindent
The value of @code{a[3]} could be either two or four.

@ref{table-assign-ops} lists the arithmetic assignment operators.  In each
case, the righthand operand is an expression whose value is converted
to a number.

@cindex @code{-} (hyphen), @code{-=} operator
@cindex hyphen (@code{-}), @code{-=} operator
@cindex @code{*} (asterisk), @code{*=} operator
@cindex asterisk (@code{*}), @code{*=} operator
@cindex @code{/} (forward slash), @code{/=} operator
@cindex forward slash (@code{/}), @code{/=} operator
@cindex @code{%} (percent sign), @code{%=} operator
@cindex percent sign (@code{%}), @code{%=} operator
@cindex @code{^} (caret), @code{^=} operator
@cindex caret (@code{^}), @code{^=} operator
@cindex @code{*} (asterisk), @code{**=} operator
@cindex asterisk (@code{*}), @code{**=} operator
@float Table,table-assign-ops
@caption{Arithmetic assignment operators}
@multitable @columnfractions .30 .70
@headitem Operator @tab Effect
@item @var{lvalue} @code{+=} @var{increment} @tab Add @var{increment} to the value of @var{lvalue}.
@item @var{lvalue} @code{-=} @var{decrement} @tab Subtract @var{decrement} from the value of @var{lvalue}.
@item @var{lvalue} @code{*=} @var{coefficient} @tab Multiply the value of @var{lvalue} by @var{coefficient}.
@item @var{lvalue} @code{/=} @var{divisor} @tab Divide the value of @var{lvalue} by @var{divisor}.
@item @var{lvalue} @code{%=} @var{modulus} @tab Set @var{lvalue} to its remainder by @var{modulus}.
@cindex common extensions, @code{**=} operator
@cindex extensions, common@comma{} @code{**=} operator
@cindex @command{awk} language, POSIX version
@cindex POSIX @command{awk}
@item @var{lvalue} @code{^=} @var{power} @tab Raise @var{lvalue} to the power @var{power}.
@item @var{lvalue} @code{**=} @var{power} @tab Raise @var{lvalue} to the power @var{power}. @value{COMMONEXT}
@end multitable
@end float

@cindex POSIX @command{awk}, @code{**=} operator and
@cindex portability, @code{**=} operator and
@quotation NOTE
Only the @samp{^=} operator is specified by POSIX.
For maximum portability, do not use the @samp{**=} operator.
@end quotation

@sidebar Syntactic Ambiguities Between @samp{/=} and Regular Expressions
@cindex dark corner, regexp constants, @code{/=} operator and
@cindex @code{/} (forward slash), @code{/=} operator, vs. @code{/=@dots{}/} regexp constant
@cindex forward slash (@code{/}), @code{/=} operator, vs. @code{/=@dots{}/} regexp constant
@cindex regexp constants, @code{/=@dots{}/}, @code{/=} operator and

@c derived from email from  "Nelson H. F. Beebe" <beebe@math.utah.edu>
@c Date: Mon, 1 Sep 1997 13:38:35 -0600 (MDT)

@cindex dark corner, @code{/=} operator vs. @code{/=@dots{}/} regexp constant
@cindex ambiguity, syntactic: @code{/=} operator vs. @code{/=@dots{}/} regexp constant
@cindex syntactic ambiguity: @code{/=} operator vs. @code{/=@dots{}/} regexp constant
@cindex @code{/=} operator vs. @code{/=@dots{}/} regexp constant
There is a syntactic ambiguity between the @code{/=} assignment
operator and regexp constants whose first character is an @samp{=}.
@value{DARKCORNER}
This is most notable in some commercial @command{awk} versions.
For example:

@example
$ @kbd{awk /==/ /dev/null}
@error{} awk: syntax error at source line 1
@error{}  context is
@error{}         >>> /= <<<
@error{} awk: bailing out at source line 1
@end example

@noindent
A workaround is:

@example
awk '/[=]=/' /dev/null
@end example

@command{gawk} does not have this problem; BWK @command{awk}
and @command{mawk} also do not.
@end sidebar

@node Increment Ops
@subsection Increment and Decrement Operators

@cindex increment operators
@cindex operators, decrement/increment
@dfn{Increment} and @dfn{decrement operators} increase or decrease the value of
a variable by one.  An assignment operator can do the same thing, so
the increment operators add no power to the @command{awk} language; however, they
are convenient abbreviations for very common operations.

@cindex side effects
@cindex @code{+} (plus sign), @code{++} operator
@cindex plus sign (@code{+}), @code{++} operator
@cindex side effects, decrement/increment operators
The operator used for adding one is written @samp{++}.  It can be used to increment
a variable either before or after taking its value.
To @dfn{pre-increment} a variable @code{v}, write @samp{++v}.  This adds
one to the value of @code{v}---that new value is also the value of the
expression. (The assignment expression @samp{v += 1} is completely equivalent.)
Writing the @samp{++} after the variable specifies @dfn{post-increment}.  This
increments the variable value just the same; the difference is that the
value of the increment expression itself is the variable's @emph{old}
value.  Thus, if @code{foo} has the value four, then the expression @samp{foo++}
has the value four, but it changes the value of @code{foo} to five.
In other words, the operator returns the old value of the variable,
but with the side effect of incrementing it.

The post-increment @samp{foo++} is nearly the same as writing @samp{(foo
+= 1) - 1}.  It is not perfectly equivalent because all numbers in
@command{awk} are floating point---in floating point, @samp{foo + 1 - 1} does
not necessarily equal @code{foo}.  But the difference is minute as
long as you stick to numbers that are fairly small (less than
@iftex
@math{10^{12}}).
@end iftex
@ifinfo
10e12).
@end ifinfo
@ifnottex
@ifnotinfo
10@sup{12}).
@end ifnotinfo
@end ifnottex

@cindex @code{$} (dollar sign), incrementing fields and arrays
@cindex dollar sign (@code{$}), incrementing fields and arrays
Fields and array elements are incremented
just like variables.  (Use @samp{$(i++)} when you want to do a field reference
and a variable increment at the same time.  The parentheses are necessary
because of the precedence of the field reference operator @samp{$}.)

@cindex decrement operators
The decrement operator @samp{--} works just like @samp{++}, except that
it subtracts one instead of adding it.  As with @samp{++}, it can be used before
the lvalue to pre-decrement or after it to post-decrement.
Following is a summary of increment and decrement expressions:

@table @code
@cindex @code{+} (plus sign), @code{++} operator
@cindex plus sign (@code{+}), @code{++} operator
@item ++@var{lvalue}
Increment @var{lvalue}, returning the new value as the
value of the expression.

@item @var{lvalue}++
Increment @var{lvalue}, returning the @emph{old} value of @var{lvalue}
as the value of the expression.

@cindex @code{-} (hyphen), @code{--} operator
@cindex hyphen (@code{-}), @code{--} operator
@item --@var{lvalue}
Decrement @var{lvalue}, returning the new value as the
value of the expression.
(This expression is
like @samp{++@var{lvalue}}, but instead of adding, it subtracts.)

@item @var{lvalue}--
Decrement @var{lvalue}, returning the @emph{old} value of @var{lvalue}
as the value of the expression.
(This expression is
like @samp{@var{lvalue}++}, but instead of adding, it subtracts.)
@end table

@sidebar Operator Evaluation Order
@cindex precedence
@cindex operators, precedence
@cindex portability, operators
@cindex evaluation order
@cindex Marx, Groucho
@quotation
@i{Doctor, it hurts when I do this!@*
Then don't do that!}
@author Groucho Marx
@end quotation

@noindent
What happens for something like the following?

@example
b = 6
print b += b++
@end example

@noindent
Or something even stranger?

@example
b = 6
b += ++b + b++
print b
@end example

@cindex side effects
In other words, when do the various side effects prescribed by the
postfix operators (@samp{b++}) take effect?
When side effects happen is @dfn{implementation-defined}.
In other words, it is up to the particular version of @command{awk}.
The result for the first example may be 12 or 13, and for the second, it
may be 22 or 23.

In short, doing things like this is not recommended and definitely
not anything that you can rely upon for portability.
You should avoid such things in your own programs.
@c You'll sleep better at night and be able to look at yourself
@c in the mirror in the morning.
@end sidebar

@node Truth Values and Conditions
@section Truth Values and Conditions

In certain contexts, expression values also serve as ``truth values''; i.e.,
they determine what should happen next as the program runs. This
@value{SECTION} describes how @command{awk} defines ``true'' and ``false''
and how values are compared.

@menu
* Truth Values::                What is ``true'' and what is ``false''.
* Typing and Comparison::       How variables acquire types and how this
                                affects comparison of numbers and strings with
                                @samp{<}, etc.
* Boolean Ops::                 Combining comparison expressions using boolean
                                operators @samp{||} (``or''), @samp{&&}
                                (``and'') and @samp{!} (``not'').
* Conditional Exp::             Conditional expressions select between two
                                subexpressions under control of a third
                                subexpression.
@end menu

@node Truth Values
@subsection True and False in @command{awk}
@cindex truth values
@cindex logical false/true
@cindex false, logical
@cindex true, logical

@cindex null strings
Many programming languages have a special representation for the concepts
of ``true'' and ``false.''  Such languages usually use the special
constants @code{true} and @code{false}, or perhaps their uppercase
equivalents.
However, @command{awk} is different.
It borrows a very simple concept of true and
false from C.  In @command{awk}, any nonzero numeric value @emph{or} any
nonempty string value is true.  Any other value (zero or the null
string, @code{""}) is false.  The following program prints @samp{A strange
truth value} three times:

@example
BEGIN @{
   if (3.1415927)
       print "A strange truth value"
   if ("Four Score And Seven Years Ago")
       print "A strange truth value"
   if (j = 57)
       print "A strange truth value"
@}
@end example

@cindex dark corner, @code{"0"} is actually true
There is a surprising consequence of the ``nonzero or non-null'' rule:
the string constant @code{"0"} is actually true, because it is non-null.
@value{DARKCORNER}

@node Typing and Comparison
@subsection Variable Typing and Comparison Expressions
@quotation
@i{The Guide is definitive. Reality is frequently inaccurate.}
@author Douglas Adams, @cite{The Hitchhiker's Guide to the Galaxy}
@end quotation
@c 2/2015: Antonio Colombo points out that this is really from
@c The Restaurant at the End of the Universe. But I'm going to
@c leave it alone.

@cindex comparison expressions
@cindex expressions, comparison
@cindex expressions, matching, See comparison expressions
@cindex matching, expressions, See comparison expressions
@cindex relational operators, See comparison operators
@cindex operators, relational, See operators@comma{} comparison
@cindex variable typing
@cindex variables, types of, comparison expressions and
Unlike in other programming languages, in @command{awk} variables do not have a
fixed type. Instead, they can be either a number or a string, depending
upon the value that is assigned to them.
We look now at how variables are typed, and how @command{awk}
compares variables.

@menu
* Variable Typing::             String type versus numeric type.
* Comparison Operators::        The comparison operators.
* POSIX String Comparison::     String comparison with POSIX rules.
@end menu

@node Variable Typing
@subsubsection String Type versus Numeric Type

Scalar objects in @command{awk} (variables, array elements, and fields)
are @emph{dynamically} typed.  This means their type can change as the
program runs, from @dfn{untyped} before any use,@footnote{@command{gawk}
calls this @dfn{unassigned}, as the following example shows.} to string
or number, and then from string to number or number to string, as the
program progresses.  (@command{gawk} also provides regexp-typed scalars,
but let's ignore that for now; @pxref{Strong Regexp Constants}.)

You can't do much with untyped variables, other than tell that they
are untyped. The following program tests @code{a} against @code{""}
and @code{0}; the test succeeds when @code{a} has never been assigned
a value.  It also uses the built-in @code{typeof()} function
(not presented yet; @pxref{Type Functions}) to show @code{a}'s type:

@example
$ @kbd{gawk 'BEGIN @{ print (a == "" && a == 0 ?}
> @kbd{"a is untyped" : "a has a type!") ; print typeof(a) @}'}
@print{} a is untyped
@print{} unassigned
@end example

A scalar has numeric type when assigned a numeric value,
such as from a numeric constant, or from another scalar
with numeric type:

@example
$ @kbd{gawk 'BEGIN @{ a = 42 ; print typeof(a)}
> @kbd{b = a ; print typeof(b) @}'}
number
number
@end example

Similarly, a scalar has string type when assigned a string
value, such as from a string constant, or from another scalar
with string type:

@example
$ @kbd{gawk 'BEGIN @{ a = "forty two" ; print typeof(a)}
> @kbd{b = a ; print typeof(b) @}'}
string
string
@end example

So far, this is all simple and straightforward.  What happens, though,
when @command{awk} has to process data from a user?  Let's start with
field data.  What should the following command produce as output?

@example
echo hello | awk '@{ printf("%s %s < 42\n", $1,
                           ($1 < 42 ? "is" : "is not")) @}'
@end example

@noindent
Since @samp{hello} is alphabetic data, @command{awk} can only do a string
comparison.  Internally, it converts @code{42} into @code{"42"} and compares
the two string values @code{"hello"} and @code{"42"}. Here's the result:

@example
$ @kbd{echo hello | awk '@{ printf("%s %s < 42\n", $1,}
> @kbd{                           ($1 < 42 ? "is" : "is not")) @}'}
@print{} hello is not < 42
@end example

However, what happens when data from a user @emph{looks like} a number?
On the one hand, in reality, the input data consists of characters, not
binary numeric
values.  But, on the other hand, the data looks numeric, and @command{awk}
really ought to treat it as such. And indeed, it does:

@example
$ @kbd{echo 37 | awk '@{ printf("%s %s < 42\n", $1,}
> @kbd{                        ($1 < 42 ? "is" : "is not")) @}'}
@print{} 37 is < 42
@end example

Here are the rules for when @command{awk}
treats data as a number, and for when it treats data as a string.

@cindex numeric, strings
@cindex strings, numeric
@cindex POSIX @command{awk}, numeric strings and
The POSIX standard uses the term @dfn{numeric string} for input data that
looks numeric.  The @samp{37} in the previous example is a numeric string.
So what is the type of a numeric string? Answer: numeric.

The type of a variable is important because the types of two variables
determine how they are compared.
Variable typing follows these definitions and rules:

@itemize @value{BULLET}
@item
A numeric constant or the result of a numeric operation has the @dfn{numeric}
attribute.

@item
A string constant or the result of a string operation has the @dfn{string}
attribute.

@item
Fields, @code{getline} input, @code{FILENAME}, @code{ARGV} elements,
@code{ENVIRON} elements, and the elements of an array created by
@code{match()}, @code{split()}, and @code{patsplit()} that are numeric
strings have the @dfn{strnum} attribute.@footnote{Thus, a POSIX
numeric string and @command{gawk}'s strnum are the same thing.}
Otherwise, they have
the @dfn{string} attribute.  Uninitialized variables also have the
@dfn{strnum} attribute.

@item
Attributes propagate across assignments but are not changed by
any use.
@c (Although a use may cause the entity to acquire an additional
@c value such that it has both a numeric and string value, this leaves the
@c attribute unchanged.)
@c This is important but not relevant
@end itemize

The last rule is particularly important. In the following program,
@code{a} has numeric type, even though it is later used in a string
operation:

@example
BEGIN @{
     a = 12.345
     b = a " is a cute number"
     print b
@}
@end example

When two operands are compared, either string comparison or numeric comparison
may be used. This depends upon the attributes of the operands, according to the
following symmetric matrix:

@c thanks to Karl Berry, kb@cs.umb.edu, for major help with TeX tables
@tex
\centerline{
\vbox{\bigskip % space above the table (about 1 linespace)
% Because we have vertical rules, we can't let TeX insert interline space
% in its usual way.
\offinterlineskip
%
% Define the table template. & separates columns, and \cr ends the
% template (and each row). # is replaced by the text of that entry on
% each row. The template for the first column breaks down like this:
%   \strut -- a way to make each line have the height and depth
%             of a normal line of type, since we turned off interline spacing.
%   \hfil -- infinite glue; has the effect of right-justifying in this case.
%   #     -- replaced by the text (for instance, `STRNUM', in the last row).
%   \quad -- about the width of an `M'. Just separates the columns.
%
% The second column (\vrule#) is what generates the vertical rule that
% spans table rows.
%
% The doubled && before the next entry means `repeat the following
% template as many times as necessary on each line' -- in our case, twice.
%
% The template itself, \quad#\hfil, left-justifies with a little space before.
%
\halign{\strut\hfil#\quad&\vrule#&&\quad#\hfil\cr
	&&STRING	&NUMERIC	&STRNUM\cr
% The \omit tells TeX to skip inserting the template for this column on
% this particular row. In this case, we only want a little extra space
% to separate the heading row from the rule below it.  the depth 2pt --
% `\vrule depth 2pt' is that little space.
\omit	&depth 2pt\cr
% This is the horizontal rule below the heading. Since it has nothing to
% do with the columns of the table, we use \noalign to get it in there.
\noalign{\hrule}
% Like above, this time a little more space.
\omit	&depth 4pt\cr
% The remaining rows have nothing special about them.
STRING	&&string	&string		&string\cr
NUMERIC	&&string	&numeric	&numeric\cr
STRNUM  &&string	&numeric	&numeric\cr
}}}
@end tex
@ifnottex
@ifnotdocbook
@verbatim
        +----------------------------------------------
        |       STRING          NUMERIC         STRNUM
--------+----------------------------------------------
        |
STRING  |       string          string          string
        |
NUMERIC |       string          numeric         numeric
        |
STRNUM  |       string          numeric         numeric
--------+----------------------------------------------
@end verbatim
@end ifnotdocbook
@end ifnottex
@docbook
<informaltable>
<tgroup cols="4">
<colspec colname="1" align="left"/>
<colspec colname="2" align="left"/>
<colspec colname="3" align="left"/>
<colspec colname="4" align="left"/>
<thead>
<row>
<entry/>
<entry>STRING</entry>
<entry>NUMERIC</entry>
<entry>STRNUM</entry>
</row>
</thead>

<tbody>
<row>
<entry><emphasis role="bold">STRING</emphasis></entry>
<entry>string</entry>
<entry>string</entry>
<entry>string</entry>
</row>

<row>
<entry><emphasis role="bold">NUMERIC</emphasis></entry>
<entry>string</entry>
<entry>numeric</entry>
<entry>numeric</entry>
</row>

<row>
<entry><emphasis role="bold">STRNUM</emphasis></entry>
<entry>string</entry>
<entry>numeric</entry>
<entry>numeric</entry>
</row>

</tbody>
</tgroup>
</informaltable>

@end docbook

The basic idea is that user input that looks numeric---and @emph{only}
user input---should be treated as numeric, even though it is actually
made of characters and is therefore also a string.
Thus, for example, the string constant @w{@code{" +3.14"}},
when it appears in program source code,
is a string---even though it looks numeric---and
is @emph{never} treated as a number for comparison
purposes.

In short, when one operand is a ``pure'' string, such as a string
constant, then a string comparison is performed.  Otherwise, a
numeric comparison is performed.
(The primary difference between a number and a strnum is that
for strnums @command{gawk} preserves the original string value that
the scalar had when it came in.)

This point bears additional emphasis:
Input that looks numeric @emph{is} numeric.
All other input is treated as strings.

Thus, the six-character input string @w{@samp{ +3.14}} receives the
strnum attribute. In contrast, the eight characters
@w{@code{" +3.14"}} appearing in program text comprise a string constant.
The following examples print @samp{1} when the comparison between
the two different constants is true, and @samp{0} otherwise:

@c 22.9.2014: Tested with mawk and BWK awk, got same results.
@example
$ @kbd{echo ' +3.14' | awk '@{ print($0 == " +3.14") @}'}    @ii{True}
@print{} 1
$ @kbd{echo ' +3.14' | awk '@{ print($0 == "+3.14") @}'}     @ii{False}
@print{} 0
$ @kbd{echo ' +3.14' | awk '@{ print($0 == "3.14") @}'}      @ii{False}
@print{} 0
$ @kbd{echo ' +3.14' | awk '@{ print($0 == 3.14) @}'}        @ii{True}
@print{} 1
$ @kbd{echo ' +3.14' | awk '@{ print($1 == " +3.14") @}'}    @ii{False}
@print{} 0
$ @kbd{echo ' +3.14' | awk '@{ print($1 == "+3.14") @}'}     @ii{True}
@print{} 1
$ @kbd{echo ' +3.14' | awk '@{ print($1 == "3.14") @}'}      @ii{False}
@print{} 0
$ @kbd{echo ' +3.14' | awk '@{ print($1 == 3.14) @}'}        @ii{True}
@print{} 1
@end example

You can see the type of an input field (or other user input)
using @code{typeof()}:

@example
$ @kbd{echo hello 37 | gawk '@{ print typeof($1), typeof($2) @}'}
@print{} string strnum
@end example

@node Comparison Operators
@subsubsection Comparison Operators

@dfn{Comparison expressions} compare strings or numbers for
relationships such as equality.  They are written using @dfn{relational
operators}, which are a superset of those in C.
@ref{table-relational-ops} describes them.

@cindex @code{<} (left angle bracket), @code{<} operator
@cindex left angle bracket (@code{<}), @code{<} operator
@cindex @code{<} (left angle bracket), @code{<=} operator
@cindex left angle bracket (@code{<}), @code{<=} operator
@cindex @code{>} (right angle bracket), @code{>=} operator
@cindex right angle bracket (@code{>}), @code{>=} operator
@cindex @code{>} (right angle bracket), @code{>} operator
@cindex right angle bracket (@code{>}), @code{>} operator
@cindex @code{=} (equals sign), @code{==} operator
@cindex equals sign (@code{=}), @code{==} operator
@cindex @code{!} (exclamation point), @code{!=} operator
@cindex exclamation point (@code{!}), @code{!=} operator
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@cindex @code{in} operator
@float Table,table-relational-ops
@caption{Relational operators}
@multitable @columnfractions .25 .75
@headitem Expression @tab Result
@item @var{x} @code{<} @var{y} @tab True if @var{x} is less than @var{y}
@item @var{x} @code{<=} @var{y} @tab True if @var{x} is less than or equal to @var{y}
@item @var{x} @code{>} @var{y} @tab True if @var{x} is greater than @var{y}
@item @var{x} @code{>=} @var{y} @tab True if @var{x} is greater than or equal to @var{y}
@item @var{x} @code{==} @var{y} @tab True if @var{x} is equal to @var{y}
@item @var{x} @code{!=} @var{y} @tab True if @var{x} is not equal to @var{y}
@item @var{x} @code{~} @var{y} @tab True if the string @var{x} matches the regexp denoted by @var{y}
@item @var{x} @code{!~} @var{y} @tab True if the string @var{x} does not match the regexp denoted by @var{y}
@item @var{subscript} @code{in} @var{array} @tab True if the array @var{array} has an element with the subscript @var{subscript}
@end multitable
@end float

Comparison expressions have the value one if true and zero if false.
When comparing operands of mixed types, numeric operands are converted
to strings using the value of @code{CONVFMT}
(@pxref{Conversion}).

Strings are compared
by comparing the first character of each, then the second character of each,
and so on.  Thus, @code{"10"} is less than @code{"9"}.  If there are two
strings where one is a prefix of the other, the shorter string is less than
the longer one.  Thus, @code{"abc"} is less than @code{"abcd"}.

@cindex troubleshooting, @code{==} operator
It is very easy to accidentally mistype the @samp{==} operator and
leave off one of the @samp{=} characters.  The result is still valid
@command{awk} code, but the program does not do what is intended:

@example
@group
if (a = b)   # oops! should be a == b
   @dots{}
else
   @dots{}
@end group
@end example

@noindent
Unless @code{b} happens to be zero or the null string, the @code{if}
part of the test always succeeds.  Because the operators are
so similar, this kind of error is very difficult to spot when
scanning the source code.

The following list of expressions illustrates the kinds of comparisons
@command{awk} performs, as well as what the result of each comparison is:

@table @code
@item 1.5 <= 2.0
Numeric comparison (true)

@item "abc" >= "xyz"
String comparison (false)

@item 1.5 != " +2"
String comparison (true)

@item "1e2" < "3"
String comparison (true)

@item a = 2; b = "2"
@itemx a == b
String comparison (true)

@item a = 2; b = " +2"
@itemx a == b
String comparison (false)
@end table

In this example:

@example
$ @kbd{echo 1e2 3 | awk '@{ print ($1 < $2) ? "true" : "false" @}'}
@print{} false
@end example

@cindex comparison expressions, string vs.@: regexp
@c @cindex string comparison vs.@: regexp comparison
@c @cindex regexp comparison vs.@: string comparison
@noindent
the result is @samp{false} because both @code{$1} and @code{$2}
are user input.  They are numeric strings---therefore both have
the strnum attribute, dictating a numeric comparison.
The purpose of the comparison rules and the use of numeric strings is
to attempt to produce the behavior that is ``least surprising,'' while
still ``doing the right thing.''

String comparisons and regular expression comparisons are very different.
For example:

@example
x == "foo"
@end example

@noindent
has the value one, or is true if the variable @code{x}
is precisely @samp{foo}.  By contrast:

@example
x ~ /foo/
@end example

@noindent
has the value one if @code{x} contains @samp{foo}, such as
@code{"Oh, what a fool am I!"}.

@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
The righthand operand of the @samp{~} and @samp{!~} operators may be
either a regexp constant (@code{/}@dots{}@code{/}) or an ordinary
expression. In the latter case, the value of the expression as a string is used as a
dynamic regexp (@pxref{Regexp Usage}; also
@pxref{Computed Regexps}).

@cindex @command{awk}, regexp constants and
@cindex regexp constants
A constant regular
expression in slashes by itself is also an expression.
@code{/@var{regexp}/} is an abbreviation for the following comparison expression:

@example
$0 ~ /@var{regexp}/
@end example

One special place where @code{/foo/} is @emph{not} an abbreviation for
@samp{$0 ~ /foo/} is when it is the righthand operand of @samp{~} or
@samp{!~}.
@xref{Using Constant Regexps},
where this is discussed in more detail.

@node POSIX String Comparison
@subsubsection String Comparison Based on Locale Collating Order

The POSIX standard used to say that all string comparisons are
performed based on the locale's @dfn{collating order}. This
is the order in which characters sort, as defined by the locale
(for more discussion, @pxref{Locales}).  This order is usually very
different from the results obtained when doing straight byte-by-byte
comparison.@footnote{Technically, string comparison is supposed to behave
the same way as if the strings were compared with the C @code{strcoll()}
function.}

Because this behavior differs considerably from existing practice,
@command{gawk} only implemented it when in POSIX mode (@pxref{Options}).
Here is an example to illustrate the difference, in an @code{en_US.UTF-8}
locale:

@example
$ @kbd{gawk 'BEGIN @{ printf("ABC < abc = %s\n",}
>                     @kbd{("ABC" < "abc" ? "TRUE" : "FALSE")) @}'}
@print{} ABC < abc = TRUE
$ @kbd{gawk --posix 'BEGIN @{ printf("ABC < abc = %s\n",}
>                             @kbd{("ABC" < "abc" ? "TRUE" : "FALSE")) @}'}
@print{} ABC < abc = FALSE
@end example

Fortunately, as of August 2016, comparison based on locale
collating order is no longer required for the @code{==} and @code{!=}
operators.@footnote{See @uref{http://austingroupbugs.net/view.php?id=1070,
the Austin Group website}.} However, comparison based on locales is still
required for @code{<}, @code{<=}, @code{>}, and @code{>=}.  POSIX thus
recommends as follows:

@quotation
Since the @code{==} operator checks whether strings are identical,
not whether they collate equally, applications needing to check whether
strings collate equally can use:

@example
a <= b && a >= b
@end example
@end quotation

As of @value{PVERSION} 4.2, @command{gawk} continues to use locale
collating order for @code{<}, @code{<=}, @code{>}, and @code{>=} only
in POSIX mode.

@ignore
References: http://austingroupbugs.net/view.php?id=963
and http://austingroupbugs.net/view.php?id=1070.
@end ignore

@node Boolean Ops
@subsection Boolean Expressions
@cindex and Boolean-logic operator
@cindex or Boolean-logic operator
@cindex not Boolean-logic operator
@cindex expressions, Boolean
@cindex Boolean expressions
@cindex operators, Boolean, See Boolean expressions
@cindex Boolean operators, See Boolean expressions
@cindex logical operators, See Boolean expressions
@cindex operators, logical, See Boolean expressions

A @dfn{Boolean expression} is a combination of comparison expressions or
matching expressions, using the Boolean operators ``or''
(@samp{||}), ``and'' (@samp{&&}), and ``not'' (@samp{!}), along with
parentheses to control nesting.  The truth value of the Boolean expression is
computed by combining the truth values of the component expressions.
Boolean expressions are also referred to as @dfn{logical expressions}.
The terms are equivalent.

Boolean expressions can be used wherever comparison and matching
expressions can be used.  They can be used in @code{if}, @code{while},
@code{do}, and @code{for} statements
(@pxref{Statements}).
They have numeric values (one if true, zero if false) that come into play
if the result of the Boolean expression is stored in a variable or
used in arithmetic.

In addition, every Boolean expression is also a valid pattern, so
you can use one as a pattern to control the execution of rules.
The Boolean operators are:

@table @code
@item @var{boolean1} && @var{boolean2}
True if both @var{boolean1} and @var{boolean2} are true.  For example,
the following statement prints the current input record if it contains
both @samp{edu} and @samp{li}:

@example
if ($0 ~ /edu/ && $0 ~ /li/) print
@end example

@cindex side effects, Boolean operators
The subexpression @var{boolean2} is evaluated only if @var{boolean1}
is true.  This can make a difference when @var{boolean2} contains
expressions that have side effects. In the case of @samp{$0 ~ /foo/ &&
($2 == bar++)}, the variable @code{bar} is not incremented if there is
no substring @samp{foo} in the record.

@item @var{boolean1} || @var{boolean2}
True if at least one of @var{boolean1} or @var{boolean2} is true.
For example, the following statement prints all records in the input
that contain @emph{either} @samp{edu} or
@samp{li}:

@example
if ($0 ~ /edu/ || $0 ~ /li/) print
@end example

The subexpression @var{boolean2} is evaluated only if @var{boolean1}
is false.  This can make a difference when @var{boolean2} contains
expressions that have side effects.
(Thus, this test never really distinguishes records that contain both
@samp{edu} and @samp{li}---as soon as @samp{edu} is matched,
the full test succeeds.)

@item ! @var{boolean}
True if @var{boolean} is false.  For example,
the following program prints @samp{no home!} in
the unusual event that the @env{HOME} environment
variable is not defined:

@example
BEGIN @{ if (! ("HOME" in ENVIRON))
            print "no home!" @}
@end example

(The @code{in} operator is described in
@ref{Reference to Elements}.)
@end table

@cindex short-circuit operators
@cindex operators, short-circuit
@cindex @code{&} (ampersand), @code{&&} operator
@cindex ampersand (@code{&}), @code{&&} operator
@cindex @code{|} (vertical bar), @code{||} operator
@cindex vertical bar (@code{|}), @code{||} operator
The @samp{&&} and @samp{||} operators are called @dfn{short-circuit}
operators because of the way they work.  Evaluation of the full expression
is ``short-circuited'' if the result can be determined partway through
its evaluation.

@cindex line continuations
Statements that end with @samp{&&} or @samp{||} can be continued simply
by putting a newline after them.  But you cannot put a newline in front
of either of these operators without using backslash continuation
(@pxref{Statements/Lines}).

@cindex @code{!} (exclamation point), @code{!}  operator
@cindex exclamation point (@code{!}), @code{!} operator
@cindex newlines
@cindex variables, flag
@cindex flag variables
The actual value of an expression using the @samp{!} operator is
either one or zero, depending upon the truth value of the expression it
is applied to.
The @samp{!} operator is often useful for changing the sense of a flag
variable from false to true and back again. For example, the following
program is one way to print lines in between special bracketing lines:

@example
$1 == "START"   @{ interested = ! interested; next @}
interested      @{ print @}
$1 == "END"     @{ interested = ! interested; next @}
@end example

@noindent
The variable @code{interested}, as with all @command{awk} variables, starts
out initialized to zero, which is also false.  When a line is seen whose
first field is @samp{START}, the value of @code{interested} is toggled
to true, using @samp{!}. The next rule prints lines as long as
@code{interested} is true.  When a line is seen whose first field is
@samp{END}, @code{interested} is toggled back to false.@footnote{This
program has a bug; it prints lines starting with @samp{END}. How
would you fix it?}

@ignore
Scott Deifik points out that this program isn't robust against
bogus input data, but the point is to illustrate the use of `!',
so we'll leave well enough alone.
@end ignore

Most commonly, the @samp{!} operator is used in the conditions of
@code{if} and @code{while} statements, where it often makes more
sense to phrase the logic in the negative:

@example
if (! @var{some condition} || @var{some other condition}) @{
    @var{@dots{} do whatever processing @dots{}}
@}
@end example

@cindex @code{next} statement
@quotation NOTE
The @code{next} statement is discussed in
@ref{Next Statement}.
@code{next} tells @command{awk} to skip the rest of the rules, get the
next record, and start processing the rules over again at the top.
The reason it's there is to avoid printing the bracketing
@samp{START} and @samp{END} lines.
@end quotation

@node Conditional Exp
@subsection Conditional Expressions
@cindex conditional expressions
@cindex expressions, conditional
@cindex expressions, selecting

A @dfn{conditional expression} is a special kind of expression that has
three operands.  It allows you to use one expression's value to select
one of two other expressions.
The conditional expression in @command{awk} is the same as in the C
language, as shown here:

@example
@var{selector} ? @var{if-true-exp} : @var{if-false-exp}
@end example

@noindent
There are three subexpressions.  The first, @var{selector}, is always
computed first.  If it is ``true'' (not zero or not null), then
@var{if-true-exp} is computed next, and its value becomes the value of
the whole expression.  Otherwise, @var{if-false-exp} is computed next,
and its value becomes the value of the whole expression.
For example, the following expression produces the absolute value of @code{x}:

@example
x >= 0 ? x : -x
@end example

@cindex side effects, conditional expressions
Each time the conditional expression is computed, only one of
@var{if-true-exp} and @var{if-false-exp} is used; the other is ignored.
This is important when the expressions have side effects.  For example,
this conditional expression examines element @code{i} of either array
@code{a} or array @code{b}, and increments @code{i}:

@example
x == y ? a[i++] : b[i++]
@end example

@noindent
This is guaranteed to increment @code{i} exactly once, because each time
only one of the two increment expressions is executed
and the other is not.
@xref{Arrays},
for more information about arrays.

@cindex differences in @command{awk} and @command{gawk}, line continuations
@cindex line continuations, @command{gawk}
@cindex @command{gawk}, line continuation in
As a minor @command{gawk} extension,
a statement that uses @samp{?:} can be continued simply
by putting a newline after either character.
However, putting a newline in front
of either character does not work without using backslash continuation
(@pxref{Statements/Lines}).
If @option{--posix} is specified
(@pxref{Options}), this extension is disabled.

@node Function Calls
@section Function Calls
@cindex function calls

A @dfn{function} is a name for a particular calculation.
This enables you to
ask for it by name at any point in the program.  For
example, the function @code{sqrt()} computes the square root of a number.

@cindex functions, built-in
A fixed set of functions are @dfn{built in}, which means they are
available in every @command{awk} program.  The @code{sqrt()} function is one
of these.  @xref{Built-in} for a list of built-in
functions and their descriptions.  In addition, you can define
functions for use in your program.
@xref{User-defined}
for instructions on how to do this.
Finally, @command{gawk} lets you write functions in C or C++
that may be called from your program (@pxref{Dynamic Extensions}).

@cindex arguments, in function calls
The way to use a function is with a @dfn{function call} expression,
which consists of the function name followed immediately by a list of
@dfn{arguments} in parentheses.  The arguments are expressions that
provide the raw materials for the function's calculations.
When there is more than one argument, they are separated by commas.  If
there are no arguments, just write @samp{()} after the function name.
The following examples show function calls with and without arguments:

@example
sqrt(x^2 + y^2)        @ii{one argument}
atan2(y, x)            @ii{two arguments}
rand()                 @ii{no arguments}
@end example

@cindex troubleshooting, function call syntax
@quotation CAUTION
Do not put any space between the function name and the opening parenthesis!
A user-defined function name looks just like the name of a
variable---a space would make the expression look like concatenation of
a variable with an expression inside parentheses.
With built-in functions, space before the parenthesis is harmless, but
it is best not to get into the habit of using space to avoid mistakes
with user-defined functions.
@end quotation

Each function expects a particular number
of arguments.  For example, the @code{sqrt()} function must be called with
a single argument, the number of which to take the square root:

@example
sqrt(@var{argument})
@end example

Some of the built-in functions have one or
more optional arguments.
If those arguments are not supplied, the functions
use a reasonable default value.
@xref{Built-in} for full details.  If arguments
are omitted in calls to user-defined functions, then those arguments are
treated as local variables. Such local variables act like the
empty string if referenced where a string value is required,
and like zero if referenced where a numeric value is required
(@pxref{User-defined}).

As an advanced feature, @command{gawk} provides indirect function calls,
which is a way to choose the function to call at runtime, instead of
when you write the source code to your program. We defer discussion of
this feature until later; see @ref{Indirect Calls}.

@cindex side effects, function calls
Like every other expression, the function call has a value, often
called the @dfn{return value}, which is computed by the function
based on the arguments you give it.  In this example, the return value
of @samp{sqrt(@var{argument})} is the square root of @var{argument}.
The following program reads numbers, one number per line, and prints
the square root of each one:

@example
$ @kbd{awk '@{ print "The square root of", $1, "is", sqrt($1) @}'}
@kbd{1}
@print{} The square root of 1 is 1
@kbd{3}
@print{} The square root of 3 is 1.73205
@kbd{5}
@print{} The square root of 5 is 2.23607
@kbd{Ctrl-d}
@end example

A function can also have side effects, such as assigning
values to certain variables or doing I/O.
This program shows how the @code{match()} function
(@pxref{String Functions})
changes the variables @code{RSTART} and @code{RLENGTH}:

@example
@{
    if (match($1, $2))
        print RSTART, RLENGTH
    else
        print "no match"
@}
@end example

@noindent
Here is a sample run:

@example
$ @kbd{awk -f matchit.awk}
@kbd{aaccdd  c+}
@print{} 3 2
@kbd{foo     bar}
@print{} no match
@kbd{abcdefg e}
@print{} 5 1
@end example

@node Precedence
@section Operator Precedence (How Operators Nest)
@cindex precedence
@cindex operators, precedence

@dfn{Operator precedence} determines how operators are grouped when
different operators appear close by in one expression.  For example,
@samp{*} has higher precedence than @samp{+}; thus, @samp{a + b * c}
means to multiply @code{b} and @code{c}, and then add @code{a} to the
product (i.e., @samp{a + (b * c)}).

The normal precedence of the operators can be overruled by using parentheses.
Think of the precedence rules as saying where the
parentheses are assumed to be.  In
fact, it is wise to always use parentheses whenever there is an unusual
combination of operators, because other people who read the program may
not remember what the precedence is in this case.
Even experienced programmers occasionally forget the exact rules,
which leads to mistakes.
Explicit parentheses help prevent
any such mistakes.

When operators of equal precedence are used together, the leftmost
operator groups first, except for the assignment, conditional, and
exponentiation operators, which group in the opposite order.
Thus, @samp{a - b + c} groups as @samp{(a - b) + c} and
@samp{a = b = c} groups as @samp{a = (b = c)}.

Normally the precedence of prefix unary operators does not matter,
because there is only one way to interpret
them: innermost first.  Thus, @samp{$++i} means @samp{$(++i)} and
@samp{++$x} means @samp{++($x)}.  However, when another operator follows
the operand, then the precedence of the unary operators can matter.
@samp{$x^2} means @samp{($x)^2}, but @samp{-x^2} means
@samp{-(x^2)}, because @samp{-} has lower precedence than @samp{^},
whereas @samp{$} has higher precedence.
Also, operators cannot be combined in a way that violates the
precedence rules; for example, @samp{$$0++--} is not a valid
expression because the first @samp{$} has higher precedence than the
@samp{++}; to avoid the problem the expression can be rewritten as
@samp{$($0++)--}.

This list presents @command{awk}'s operators, in order of highest
to lowest precedence:

@c @asis for docbook to come out right
@table @asis
@item @code{(}@dots{}@code{)}
Grouping.

@cindex @code{$} (dollar sign), @code{$} field operator
@cindex dollar sign (@code{$}), @code{$} field operator
@item @code{$}
Field reference.

@cindex @code{+} (plus sign), @code{++} operator
@cindex plus sign (@code{+}), @code{++} operator
@cindex @code{-} (hyphen), @code{--} operator
@cindex hyphen (@code{-}), @code{--} operator
@item @code{++ --}
Increment, decrement.

@cindex @code{^} (caret), @code{^} operator
@cindex caret (@code{^}), @code{^} operator
@cindex @code{*} (asterisk), @code{**} operator
@cindex asterisk (@code{*}), @code{**} operator
@item @code{^ **}
Exponentiation.  These operators group right to left.

@cindex @code{+} (plus sign), @code{+} operator
@cindex plus sign (@code{+}), @code{+} operator
@cindex @code{-} (hyphen), @code{-} operator
@cindex hyphen (@code{-}), @code{-} operator
@cindex @code{!} (exclamation point), @code{!} operator
@cindex exclamation point (@code{!}), @code{!} operator
@item @code{+ - !}
Unary plus, minus, logical ``not.''

@cindex @code{*} (asterisk), @code{*} operator, as multiplication operator
@cindex asterisk (@code{*}), @code{*} operator, as multiplication operator
@cindex @code{/} (forward slash), @code{/} operator
@cindex forward slash (@code{/}), @code{/} operator
@cindex @code{%} (percent sign), @code{%} operator
@cindex percent sign (@code{%}), @code{%} operator
@item @code{* / %}
Multiplication, division, remainder.

@cindex @code{+} (plus sign), @code{+} operator
@cindex plus sign (@code{+}), @code{+} operator
@cindex @code{-} (hyphen), @code{-} operator
@cindex hyphen (@code{-}), @code{-} operator
@item @code{+ -}
Addition, subtraction.

@item String concatenation
There is no special symbol for concatenation.
The operands are simply written side by side
(@pxref{Concatenation}).

@cindex @code{<} (left angle bracket), @code{<} operator
@cindex left angle bracket (@code{<}), @code{<} operator
@cindex @code{<} (left angle bracket), @code{<=} operator
@cindex left angle bracket (@code{<}), @code{<=} operator
@cindex @code{>} (right angle bracket), @code{>=} operator
@cindex right angle bracket (@code{>}), @code{>=} operator
@cindex @code{>} (right angle bracket), @code{>} operator
@cindex right angle bracket (@code{>}), @code{>} operator
@cindex @code{=} (equals sign), @code{==} operator
@cindex equals sign (@code{=}), @code{==} operator
@cindex @code{!} (exclamation point), @code{!=} operator
@cindex exclamation point (@code{!}), @code{!=} operator
@cindex @code{>} (right angle bracket), @code{>>} operator (I/O)
@cindex right angle bracket (@code{>}), @code{>>} operator (I/O)
@cindex operators, input/output
@cindex @code{|} (vertical bar), @code{|} operator (I/O)
@cindex vertical bar (@code{|}), @code{|} operator (I/O)
@cindex operators, input/output
@cindex @code{|} (vertical bar), @code{|&} operator (I/O)
@cindex vertical bar (@code{|}), @code{|&} operator (I/O)
@cindex operators, input/output
@item @code{< <= == != > >= >> | |&}
Relational and redirection.
The relational operators and the redirections have the same precedence
level.  Characters such as @samp{>} serve both as relationals and as
redirections; the context distinguishes between the two meanings.

@cindex @code{print} statement, I/O operators in
@cindex @code{printf} statement, I/O operators in
Note that the I/O redirection operators in @code{print} and @code{printf}
statements belong to the statement level, not to expressions.  The
redirection does not produce an expression that could be the operand of
another operator.  As a result, it does not make sense to use a
redirection operator near another operator of lower precedence without
parentheses.  Such combinations (e.g., @samp{print foo > a ? b : c})
result in syntax errors.
The correct way to write this statement is @samp{print foo > (a ? b : c)}.

@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@item @code{~ !~}
Matching, nonmatching.

@cindex @code{in} operator
@item @code{in}
Array membership.

@cindex @code{&} (ampersand), @code{&&} operator
@cindex ampersand (@code{&}), @code{&&} operator
@item @code{&&}
Logical ``and.''

@cindex @code{|} (vertical bar), @code{||} operator
@cindex vertical bar (@code{|}), @code{||} operator
@item @code{||}
Logical ``or.''

@cindex @code{?} (question mark), @code{?:} operator
@cindex question mark (@code{?}), @code{?:} operator
@item @code{?:}
Conditional.  This operator groups right to left.

@cindex @code{+} (plus sign), @code{+=} operator
@cindex plus sign (@code{+}), @code{+=} operator
@cindex @code{-} (hyphen), @code{-=} operator
@cindex hyphen (@code{-}), @code{-=} operator
@cindex @code{*} (asterisk), @code{*=} operator
@cindex asterisk (@code{*}), @code{*=} operator
@cindex @code{*} (asterisk), @code{**=} operator
@cindex asterisk (@code{*}), @code{**=} operator
@cindex @code{/} (forward slash), @code{/=} operator
@cindex forward slash (@code{/}), @code{/=} operator
@cindex @code{%} (percent sign), @code{%=} operator
@cindex percent sign (@code{%}), @code{%=} operator
@cindex @code{^} (caret), @code{^=} operator
@cindex caret (@code{^}), @code{^=} operator
@item @code{= += -= *= /= %= ^= **=}
Assignment.  These operators group right to left.
@end table

@cindex POSIX @command{awk}, @code{**} operator and
@cindex portability, operators, not in POSIX @command{awk}
@quotation NOTE
The @samp{|&}, @samp{**}, and @samp{**=} operators are not specified by POSIX.
For maximum portability, do not use them.
@end quotation

@node Locales
@section Where You Are Makes a Difference
@cindex locale, definition of

Modern systems support the notion of @dfn{locales}: a way to tell the
system about the local character set and language.  The ISO C standard
defines a default @code{"C"} locale, which is an environment that is
typical of what many C programmers are used to.

Once upon a time, the locale setting used to affect regexp matching,
but this is no longer true (@pxref{Ranges and Locales}).

Locales can affect record splitting.  For the normal case of @samp{RS =
"\n"}, the locale is largely irrelevant.  For other single-character
record separators, setting @samp{LC_ALL=C} in the environment will
give you much better performance when reading records.  Otherwise,
@command{gawk} has to make several function calls, @emph{per input
character}, to find the record terminator.

Locales can affect how dates and times are formatted (@pxref{Time
Functions}).  For example, a common way to abbreviate the date September
4, 2015, in the United States is ``9/4/15.''  In many countries in
Europe, however, it is abbreviated ``4.9.15.''  Thus, the @samp{%x}
specification in a @code{"US"} locale might produce @samp{9/4/15},
while in a @code{"EUROPE"} locale, it might produce @samp{4.9.15}.

According to POSIX, string comparison is also affected by locales (similar
to regular expressions).  The details are presented in @ref{POSIX String
Comparison}.

Finally, the locale affects the value of the decimal point character
used when @command{gawk} parses input data.  This is discussed in detail
in @ref{Conversion}.

@node Expressions Summary
@section Summary

@itemize @value{BULLET}
@item
Expressions are the basic elements of computation in programs.  They are
built from constants, variables, function calls, and combinations of the
various kinds of values with operators.

@item
@command{awk} supplies three kinds of constants: numeric, string, and
regexp.  @command{gawk} lets you specify numeric constants in octal
and hexadecimal (bases 8 and 16) as well as decimal (base 10).
In certain contexts, a standalone regexp constant such as @code{/foo/}
has the same meaning as @samp{$0 ~ /foo/}.

@item
Variables hold values between uses in computations. A number of built-in
variables provide information to your @command{awk} program, and a number
of others let you control how @command{awk} behaves.

@item
Numbers are automatically converted to strings, and strings to numbers,
as needed by @command{awk}. Numeric values are converted as if they were
formatted with @code{sprintf()} using the format in @code{CONVFMT}.
Locales can influence the conversions.

@item
@command{awk} provides the usual arithmetic operators (addition,
subtraction, multiplication, division, modulus), and unary plus and minus.
It also provides comparison operators, Boolean operators, an array membership
testing operator, and regexp
matching operators.  String concatenation is accomplished by placing
two expressions next to each other; there is no explicit operator.
The three-operand @samp{?:} operator provides an ``if-else'' test within
expressions.

@item
Assignment operators provide convenient shorthands for common arithmetic
operations.

@item
In @command{awk}, a value is considered to be true if it is nonzero
@emph{or} non-null. Otherwise, the value is false.

@item
A variable's type is set upon each assignment and may change over its
lifetime.  The type determines how it behaves in comparisons (string
or numeric).

@item
Function calls return a value that may be used as part of a larger
expression.  Expressions used to pass parameter values are fully
evaluated before the function is called.  @command{awk} provides
built-in and user-defined functions; this is described in
@ref{Functions}.

@item
Operator precedence specifies the order in which operations are performed,
unless explicitly overridden by parentheses.  @command{awk}'s operator
precedence is compatible with that of C.

@item
Locales can affect the format of data as output by an @command{awk}
program, and occasionally the format for data read as input.

@end itemize


@node Patterns and Actions
@chapter Patterns, Actions, and Variables
@cindex patterns

As you have already seen, each @command{awk} statement consists of
a pattern with an associated action.  This @value{CHAPTER} describes how
you build patterns and actions, what kinds of things you can do within
actions, and @command{awk}'s predefined variables.

The pattern--action rules and the statements available for use
within actions form the core of @command{awk} programming.
In a sense, everything covered
up to here has been the foundation
that programs are built on top of.  Now it's time to start
building something useful.

@menu
* Pattern Overview::            What goes into a pattern.
* Using Shell Variables::       How to use shell variables with @command{awk}.
* Action Overview::             What goes into an action.
* Statements::                  Describes the various control statements in
                                detail.
* Built-in Variables::          Summarizes the predefined variables.
* Pattern Action Summary::      Patterns and Actions summary.
@end menu

@node Pattern Overview
@section Pattern Elements

@menu
* Regexp Patterns::             Using regexps as patterns.
* Expression Patterns::         Any expression can be used as a pattern.
* Ranges::                      Pairs of patterns specify record ranges.
* BEGIN/END::                   Specifying initialization and cleanup rules.
* BEGINFILE/ENDFILE::           Two special patterns for advanced control.
* Empty::                       The empty pattern, which matches every record.
@end menu

@cindex patterns, types of
Patterns in @command{awk} control the execution of rules---a rule is
executed when its pattern matches the current input record.
The following is a summary of the types of @command{awk} patterns:

@table @code
@item /@var{regular expression}/
A regular expression. It matches when the text of the
input record fits the regular expression.
(@xref{Regexp}.)

@item @var{expression}
A single expression.  It matches when its value
is nonzero (if a number) or non-null (if a string).
(@xref{Expression Patterns}.)

@item @var{begpat}, @var{endpat}
A pair of patterns separated by a comma, specifying a @dfn{range} of records.
The range includes both the initial record that matches @var{begpat} and
the final record that matches @var{endpat}.
(@xref{Ranges}.)

@item BEGIN
@itemx END
Special patterns for you to supply startup or cleanup actions for your
@command{awk} program.
(@xref{BEGIN/END}.)

@item BEGINFILE
@itemx ENDFILE
Special patterns for you to supply startup or cleanup actions to be
done on a per-file basis.
(@xref{BEGINFILE/ENDFILE}.)

@item @var{empty}
The empty pattern matches every input record.
(@xref{Empty}.)
@end table

@node Regexp Patterns
@subsection Regular Expressions as Patterns
@cindex patterns, expressions as
@cindex regular expressions, as patterns

Regular expressions are one of the first kinds of patterns presented
in this book.
This kind of pattern is simply a regexp constant in the pattern part of
a rule.  Its  meaning is @samp{$0 ~ /@var{pattern}/}.
The pattern matches when the input record matches the regexp.
For example:

@example
/foo|bar|baz/  @{ buzzwords++ @}
END            @{ print buzzwords, "buzzwords seen" @}
@end example

@node Expression Patterns
@subsection Expressions as Patterns
@cindex expressions, as patterns

Any @command{awk} expression is valid as an @command{awk} pattern.
The pattern matches if the expression's value is nonzero (if a
number) or non-null (if a string).
The expression is reevaluated each time the rule is tested against a new
input record.  If the expression uses fields such as @code{$1}, the
value depends directly on the new input record's text; otherwise, it
depends on only what has happened so far in the execution of the
@command{awk} program.

@cindex comparison expressions, as patterns
@cindex patterns, comparison expressions as
Comparison expressions, using the comparison operators described in
@ref{Typing and Comparison},
are a very common kind of pattern.
Regexp matching and nonmatching are also very common expressions.
The left operand of the @samp{~} and @samp{!~} operators is a string.
The right operand is either a constant regular expression enclosed in
slashes (@code{/@var{regexp}/}), or any expression whose string value
is used as a dynamic regular expression
(@pxref{Computed Regexps}).
The following example prints the second field of each input record
whose first field is precisely @samp{li}:

@cindex @code{/} (forward slash), patterns and
@cindex forward slash (@code{/}), patterns and
@cindex @code{~} (tilde), @code{~} operator
@cindex tilde (@code{~}), @code{~} operator
@cindex @code{!} (exclamation point), @code{!~} operator
@cindex exclamation point (@code{!}), @code{!~} operator
@example
$ @kbd{awk '$1 == "li" @{ print $2 @}' mail-list}
@end example

@noindent
(There is no output, because there is no person with the exact name @samp{li}.)
Contrast this with the following regular expression match, which
accepts any record with a first field that contains @samp{li}:

@example
$ @kbd{awk '$1 ~ /li/ @{ print $2 @}' mail-list}
@print{} 555-5553
@print{} 555-6699
@end example

@cindex regexp constants, as patterns
@cindex patterns, regexp constants as
A regexp constant as a pattern is also a special case of an expression
pattern.  The expression @code{/li/} has the value one if @samp{li}
appears in the current input record. Thus, as a pattern, @code{/li/}
matches any record containing @samp{li}.

@cindex Boolean expressions, as patterns
Boolean expressions are also commonly used as patterns.
Whether the pattern
matches an input record depends on whether its subexpressions match.
For example, the following command prints all the records in
@file{mail-list} that contain both @samp{edu} and @samp{li}:

@example
$ @kbd{awk '/edu/ && /li/' mail-list}
@print{} Samuel       555-3430     samuel.lanceolis@@shu.edu        A
@end example

The following command prints all records in
@file{mail-list} that contain @emph{either} @samp{edu} or @samp{li}
(or both, of course):

@example
$ @kbd{awk '/edu/ || /li/' mail-list}
@print{} Amelia       555-5553     amelia.zodiacusque@@gmail.com    F
@print{} Broderick    555-0542     broderick.aliquotiens@@yahoo.com R
@print{} Fabius       555-1234     fabius.undevicesimus@@ucb.edu    F
@print{} Julie        555-6699     julie.perscrutabor@@skeeve.com   F
@print{} Samuel       555-3430     samuel.lanceolis@@shu.edu        A
@print{} Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@end example

The following command prints all records in
@file{mail-list} that do @emph{not} contain the string @samp{li}:

@example
$ @kbd{awk '! /li/' mail-list}
@print{} Anthony      555-3412     anthony.asserturo@@hotmail.com   A
@print{} Becky        555-7685     becky.algebrarum@@gmail.com      A
@print{} Bill         555-1675     bill.drowning@@hotmail.com       A
@print{} Camilla      555-2912     camilla.infusarum@@skynet.be     R
@print{} Fabius       555-1234     fabius.undevicesimus@@ucb.edu    F
@group
@print{} Martin       555-6480     martin.codicibus@@hotmail.com    A
@print{} Jean-Paul    555-2127     jeanpaul.campanorum@@nyu.edu     R
@end group
@end example

@cindex @code{BEGIN} pattern, Boolean patterns and
@cindex @code{END} pattern, Boolean patterns and
@cindex @code{BEGINFILE} pattern, Boolean patterns and
@cindex @code{ENDFILE} pattern, Boolean patterns and
The subexpressions of a Boolean operator in a pattern can be constant regular
expressions, comparisons, or any other @command{awk} expressions.  Range
patterns are not expressions, so they cannot appear inside Boolean
patterns.  Likewise, the special patterns @code{BEGIN}, @code{END},
@code{BEGINFILE}, and @code{ENDFILE},
which never match any input record, are not expressions and cannot
appear inside Boolean patterns.

The precedence of the different operators that can appear in
patterns is described in @ref{Precedence}.

@node Ranges
@subsection Specifying Record Ranges with Patterns

@cindex range patterns
@cindex patterns, ranges in
@cindex lines, matching ranges of
@cindex @code{,} (comma), in range patterns
@cindex comma (@code{,}), in range patterns
A @dfn{range pattern} is made of two patterns separated by a comma, in
the form @samp{@var{begpat}, @var{endpat}}.  It is used to match ranges of
consecutive input records.  The first pattern, @var{begpat}, controls
where the range begins, while @var{endpat} controls where
the pattern ends.  For example, the following:

@example
awk '$1 == "on", $1 == "off"' myfile
@end example

@noindent
prints every record in @file{myfile} between @samp{on}/@samp{off} pairs, inclusive.

A range pattern starts out by matching @var{begpat} against every
input record.  When a record matches @var{begpat}, the range pattern is
@dfn{turned on}, and the range pattern matches this record as well.  As long as
the range pattern stays turned on, it automatically matches every input
record read.  The range pattern also matches @var{endpat} against every
input record; when this succeeds, the range pattern is @dfn{turned off} again
for the following record.  Then the range pattern goes back to checking
@var{begpat} against each record.

@cindex @code{if} statement, actions@comma{} changing
The record that turns on the range pattern and the one that turns it
off both match the range pattern.  If you don't want to operate on
these records, you can write @code{if} statements in the rule's action
to distinguish them from the records you are interested in.

It is possible for a pattern to be turned on and off by the same
record. If the record satisfies both conditions, then the action is
executed for just that record.
For example, suppose there is text between two identical markers (e.g.,
the @samp{%} symbol), each on its own line, that should be ignored.
A first attempt would be to
combine a range pattern that describes the delimited text with the
@code{next} statement
(not discussed yet, @pxref{Next Statement}).
This causes @command{awk} to skip any further processing of the current
record and start over again with the next input record. Such a program
looks like this:

@example
/^%$/,/^%$/    @{ next @}
               @{ print @}
@end example

@noindent
@cindex lines, skipping between markers
@c @cindex flag variables
This program fails because the range pattern is both turned on and turned off
by the first line, which just has a @samp{%} on it.  To accomplish this task,
write the program in the following manner, using a flag:

@cindex @code{!} (exclamation point), @code{!} operator
@example
/^%$/     @{ skip = ! skip; next @}
skip == 1 @{ next @} # skip lines with `skip' set
@end example

In a range pattern, the comma (@samp{,}) has the lowest precedence of
all the operators (i.e., it is evaluated last).  Thus, the following
program attempts to combine a range pattern with another, simpler test:

@example
echo Yes | awk '/1/,/2/ || /Yes/'
@end example

The intent of this program is @samp{(/1/,/2/) || /Yes/}.
However, @command{awk} interprets this as @samp{/1/, (/2/ || /Yes/)}.
This cannot be changed or worked around; range patterns do not combine
with other patterns:

@example
$ @kbd{echo Yes | gawk '(/1/,/2/) || /Yes/'}
@error{} gawk: cmd. line:1: (/1/,/2/) || /Yes/
@error{} gawk: cmd. line:1:           ^ syntax error
@end example

@cindex range patterns, line continuation and
@cindex dark corner, range patterns, line continuation and
As a minor point of interest, although it is poor style,
POSIX allows you to put a newline after the comma in
a range pattern.  @value{DARKCORNER}

@node BEGIN/END
@subsection The @code{BEGIN} and @code{END} Special Patterns

@cindex @code{BEGIN} pattern
@cindex @code{END} pattern
All the patterns described so far are for matching input records.
The @code{BEGIN} and @code{END} special patterns are different.
They supply startup and cleanup actions for @command{awk} programs.
@code{BEGIN} and @code{END} rules must have actions; there is no default
action for these rules because there is no current record when they run.
@code{BEGIN} and @code{END} rules are often referred to as
``@code{BEGIN} and @code{END} blocks'' by longtime @command{awk}
programmers.

@menu
* Using BEGIN/END::             How and why to use BEGIN/END rules.
* I/O And BEGIN/END::           I/O issues in BEGIN/END rules.
@end menu

@node Using BEGIN/END
@subsubsection Startup and Cleanup Actions

@cindex @code{BEGIN} pattern
@cindex @code{END} pattern
A @code{BEGIN} rule is executed once only, before the first input record
is read. Likewise, an @code{END} rule is executed once only, after all the
input is read.  For example:

@example
$ @kbd{awk '}
> @kbd{BEGIN @{ print "Analysis of \"li\"" @}}
> @kbd{/li/  @{ ++n @}}
> @kbd{END   @{ print "\"li\" appears in", n, "records." @}' mail-list}
@print{} Analysis of "li"
@print{} "li" appears in 4 records.
@end example

@cindex @code{BEGIN} pattern, operators and
@cindex @code{END} pattern, operators and
This program finds the number of records in the input file @file{mail-list}
that contain the string @samp{li}.  The @code{BEGIN} rule prints a title
for the report.  There is no need to use the @code{BEGIN} rule to
initialize the counter @code{n} to zero, as @command{awk} does this
automatically (@pxref{Variables}).
The second rule increments the variable @code{n} every time a
record containing the pattern @samp{li} is read.  The @code{END} rule
prints the value of @code{n} at the end of the run.

The special patterns @code{BEGIN} and @code{END} cannot be used in ranges
or with Boolean operators (indeed, they cannot be used with any operators).
An @command{awk} program may have multiple @code{BEGIN} and/or @code{END}
rules.  They are executed in the order in which they appear: all the @code{BEGIN}
rules at startup and all the @code{END} rules at termination.
@code{BEGIN} and @code{END} rules may be intermixed with other rules.
This feature was added in the 1987 version of @command{awk} and is included
in the POSIX standard.
The original (1978) version of @command{awk}
required the @code{BEGIN} rule to be placed at the beginning of the
program, the @code{END} rule to be placed at the end, and only allowed one of
each.
This is no longer required, but it is a good idea to follow this template
in terms of program organization and readability.

Multiple @code{BEGIN} and @code{END} rules are useful for writing
library functions, because each library file can have its own @code{BEGIN} and/or
@code{END} rule to do its own initialization and/or cleanup.
The order in which library functions are named on the command line
controls the order in which their @code{BEGIN} and @code{END} rules are
executed.  Therefore, you have to be careful when writing such rules in
library files so that the order in which they are executed doesn't matter.
@xref{Options} for more information on
using library functions.
@xref{Library Functions},
for a number of useful library functions.

If an @command{awk} program has only @code{BEGIN} rules and no
other rules, then the program exits after the @code{BEGIN} rules are
run.@footnote{The original version of @command{awk} kept
reading and ignoring input until the end of the file was seen.}  However, if an
@code{END} rule exists, then the input is read, even if there are
no other rules in the program.  This is necessary in case the @code{END}
rule checks the @code{FNR} and @code{NR} variables.

@node I/O And BEGIN/END
@subsubsection Input/Output from @code{BEGIN} and @code{END} Rules

@cindex input/output, from @code{BEGIN} and @code{END}
There are several (sometimes subtle) points to be aware of when doing I/O
from a @code{BEGIN} or @code{END} rule.
The first has to do with the value of @code{$0} in a @code{BEGIN}
rule.  Because @code{BEGIN} rules are executed before any input is read,
there simply is no input record, and therefore no fields, when
executing @code{BEGIN} rules.  References to @code{$0} and the fields
yield a null string or zero, depending upon the context.  One way
to give @code{$0} a real value is to execute a @code{getline} command
without a variable (@pxref{Getline}).
Another way is simply to assign a value to @code{$0}.

@cindex Brian Kernighan's @command{awk}
@cindex differences in @command{awk} and @command{gawk}, @code{BEGIN}/@code{END} patterns
@cindex POSIX @command{awk}, @code{BEGIN}/@code{END} patterns
@cindex @code{print} statement, @code{BEGIN}/@code{END} patterns and
@cindex @code{BEGIN} pattern, @code{print} statement and
@cindex @code{END} pattern, @code{print} statement and
The second point is similar to the first, but from the other direction.
Traditionally, due largely to implementation issues, @code{$0} and
@code{NF} were @emph{undefined} inside an @code{END} rule.
The POSIX standard specifies that @code{NF} is available in an @code{END}
rule. It contains the number of fields from the last input record.
Most probably due to an oversight, the standard does not say that @code{$0}
is also preserved, although logically one would think that it should be.
In fact, all of BWK @command{awk}, @command{mawk}, and @command{gawk}
preserve the value of @code{$0} for use in @code{END} rules.  Be aware,
however, that some other implementations and many older versions
of Unix @command{awk} do not.

The third point follows from the first two.  The meaning of @samp{print}
inside a @code{BEGIN} or @code{END} rule is the same as always:
@samp{print $0}.  If @code{$0} is the null string, then this prints an
empty record.  Many longtime @command{awk} programmers use an unadorned
@samp{print} in @code{BEGIN} and @code{END} rules, to mean @samp{@w{print ""}},
relying on @code{$0} being null.  Although one might generally get away with
this in @code{BEGIN} rules, it is a very bad idea in @code{END} rules,
at least in @command{gawk}.  It is also poor style, because if an empty
line is needed in the output, the program should print one explicitly.

@cindex @code{next} statement, @code{BEGIN}/@code{END} patterns and
@cindex @code{nextfile} statement, @code{BEGIN}/@code{END} patterns and
@cindex @code{BEGIN} pattern, @code{next}/@code{nextfile} statements and
@cindex @code{END} pattern, @code{next}/@code{nextfile} statements and
Finally, the @code{next} and @code{nextfile} statements are not allowed
in a @code{BEGIN} rule, because the implicit
read-a-record-and-match-against-the-rules loop has not started yet.  Similarly, those statements
are not valid in an @code{END} rule, because all the input has been read.
(@xref{Next Statement} and
@ifnotdocbook
@pxref{Nextfile Statement}.)
@end ifnotdocbook
@ifdocbook
@ref{Nextfile Statement}.)
@end ifdocbook

@node BEGINFILE/ENDFILE
@subsection The @code{BEGINFILE} and @code{ENDFILE} Special Patterns
@cindex @code{BEGINFILE} pattern
@cindex @code{ENDFILE} pattern
@cindex differences in @command{awk} and @command{gawk}, @code{BEGINFILE}/@code{ENDFILE} patterns

This @value{SECTION} describes a @command{gawk}-specific feature.

Two special kinds of rule, @code{BEGINFILE} and @code{ENDFILE}, give
you ``hooks'' into @command{gawk}'s command-line file processing loop.
As with the @code{BEGIN} and @code{END} rules
@ifnottex
@ifnotdocbook
(@pxref{BEGIN/END}),
@end ifnotdocbook
@end ifnottex
@iftex
(see the previous @value{SECTION}),
@end iftex
@ifdocbook
(see the previous @value{SECTION}),
@end ifdocbook
all @code{BEGINFILE} rules in a program are merged, in the order they are
read by @command{gawk}, and all @code{ENDFILE} rules are merged as well.

The body of the @code{BEGINFILE} rules is executed just before
@command{gawk} reads the first record from a file.  @code{FILENAME}
is set to the name of the current file, and @code{FNR} is set to zero.

The @code{BEGINFILE} rule provides you the opportunity to accomplish two tasks
that would otherwise be difficult or impossible to perform:

@itemize @value{BULLET}
@item
You can test if the file is readable.  Normally, it is a fatal error if a
file named on the command line cannot be opened for reading.  However,
you can bypass the fatal error and move on to the next file on the
command line.

@cindex @command{gawk}, @code{ERRNO} variable in
@cindex @code{ERRNO} variable, with @code{BEGINFILE} pattern
@cindex @code{nextfile} statement, @code{BEGINFILE}/@code{ENDFILE} patterns and
You do this by checking if the @code{ERRNO} variable is not the empty
string; if so, then @command{gawk} was not able to open the file. In
this case, your program can execute the @code{nextfile} statement
(@pxref{Nextfile Statement}).  This causes @command{gawk} to skip
the file entirely.  Otherwise, @command{gawk} exits with the usual
fatal error.

@item
If you have written extensions that modify the record handling (by
inserting an ``input parser''; @pxref{Input Parsers}), you can invoke
them at this point, before @command{gawk} has started processing the file.
(This is a @emph{very} advanced feature, currently used only by the
@uref{https://sourceforge.net/projects/gawkextlib, @code{gawkextlib} project}.)
@end itemize

The @code{ENDFILE} rule is called when @command{gawk} has finished processing
the last record in an input file.  For the last input file,
it will be called before any @code{END} rules.
The @code{ENDFILE} rule is executed even for empty input files.

Normally, when an error occurs when reading input in the normal
input-processing loop, the error is fatal.  However, if an @code{ENDFILE}
rule is present, the error becomes non-fatal, and instead @code{ERRNO}
is set.  This makes it possible to catch and process I/O errors at the
level of the @command{awk} program.

@cindex @code{next} statement, @code{BEGINFILE}/@code{ENDFILE} patterns and
The @code{next} statement (@pxref{Next Statement}) is not allowed inside
either a @code{BEGINFILE} or an @code{ENDFILE} rule.  The @code{nextfile}
statement is allowed only inside a
@code{BEGINFILE} rule, not inside an @code{ENDFILE} rule.

@cindex @code{getline} statement, @code{BEGINFILE}/@code{ENDFILE} patterns and
The @code{getline} statement (@pxref{Getline}) is restricted inside
both @code{BEGINFILE} and @code{ENDFILE}: only redirected
forms of @code{getline} are allowed.

@code{BEGINFILE} and @code{ENDFILE} are @command{gawk} extensions.
In most other @command{awk} implementations, or if @command{gawk} is in
compatibility mode (@pxref{Options}), they are not special.

@c FIXME: For 4.2 maybe deal with this?
@ignore
Date:     Tue, 17 May 2011 02:06:10 PDT
From:     rankin@pactechdata.com (Pat Rankin)
Message-Id: <110517015127.20240f4a@pactechdata.com>
Subject:  BEGINFILE
To:       arnold@skeeve.com

     The documentation for BEGINFILE states that FNR is 0, which seems
pretty obvious.  It doesn't mention what the value of $0 is, and that's
not obvious.  I think setting it to null before starting the BEGINFILE
action would be preferable to leaving whatever was there in the last
record of the previous file.

     ENDFILE can retain the last record in $0.  I guess it has to if
the END rule's actions see that value too.  But the beginning of a new
file doesn't just mean that the old one has been closed; the old file
is being superseded, so leaving the old data around feels wrong to me.
[If the user wants to keep it on hand, he or she can use an ENDFILE
rule to grab it before moving on to the next file.]
@end ignore

@node Empty
@subsection The Empty Pattern

@cindex empty pattern
@cindex patterns, empty
An empty (i.e., nonexistent) pattern is considered to match @emph{every}
input record.  For example, the program:

@example
awk '@{ print $1 @}' mail-list
@end example

@noindent
prints the first field of every record.

@node Using Shell Variables
@section Using Shell Variables in Programs
@cindex shells, variables
@cindex @command{awk} programs, shell variables in
@c @cindex shell and @command{awk} interaction

@command{awk} programs are often used as components in larger
programs written in shell.
For example, it is very common to use a shell variable to
hold a pattern that the @command{awk} program searches for.
There are two ways to get the value of the shell variable
into the body of the @command{awk} program.

@cindex shells, quoting
A common method is to use shell quoting to substitute
the variable's value into the program inside the script.
For example, consider the following program:

@example
@group
printf "Enter search pattern: "
read pattern
awk "/$pattern/ "'@{ nmatches++ @}
     END @{ print nmatches, "found" @}' /path/to/data
@end group
@end example

@noindent
The @command{awk} program consists of two pieces of quoted text
that are concatenated together to form the program.
The first part is double-quoted, which allows substitution of
the @code{pattern} shell variable inside the quotes.
The second part is single-quoted.

Variable substitution via quoting works, but can potentially be
messy.  It requires a good understanding of the shell's quoting rules
(@pxref{Quoting}),
and it's often difficult to correctly
match up the quotes when reading the program.

A better method is to use @command{awk}'s variable assignment feature
(@pxref{Assignment Options})
to assign the shell variable's value to an @command{awk} variable.
Then use dynamic regexps to match the pattern
(@pxref{Computed Regexps}).
The following shows how to redo the
previous example using this technique:

@example
printf "Enter search pattern: "
read pattern
awk -v pat="$pattern" '$0 ~ pat @{ nmatches++ @}
       END @{ print nmatches, "found" @}' /path/to/data
@end example

@noindent
Now, the @command{awk} program is just one single-quoted string.
The assignment @samp{-v pat="$pattern"} still requires double quotes,
in case there is whitespace in the value of @code{$pattern}.
The @command{awk} variable @code{pat} could be named @code{pattern}
too, but that would be more confusing.  Using a variable also
provides more flexibility, as the variable can be used anywhere inside
the program---for printing, as an array subscript, or for any other
use---without requiring the quoting tricks at every point in the program.

@node Action Overview
@section Actions
@c @cindex action, definition of
@c @cindex curly braces
@c @cindex action, curly braces
@c @cindex action, separating statements
@cindex actions

An @command{awk} program or script consists of a series of
rules and function definitions interspersed.  (Functions are
described later.  @xref{User-defined}.)
A rule contains a pattern and an action, either of which (but not
both) may be omitted.  The purpose of the @dfn{action} is to tell
@command{awk} what to do once a match for the pattern is found.  Thus,
in outline, an @command{awk} program generally looks like this:

@display
[@var{pattern}]  @code{@{ @var{action} @}}
 @var{pattern}  [@code{@{ @var{action} @}}]
@dots{}
@code{function @var{name}(@var{args}) @{ @dots{} @}}
@dots{}
@end display

@cindex @code{@{@}} (braces), actions and
@cindex braces (@code{@{@}}), actions and
@cindex separators, for statements in actions
@cindex newlines, separating statements in actions
@cindex @code{;} (semicolon), separating statements in actions
@cindex semicolon (@code{;}), separating statements in actions
An action consists of one or more @command{awk} @dfn{statements}, enclosed
in braces (@samp{@{@r{@dots{}}@}}).  Each statement specifies one
thing to do.  The statements are separated by newlines or semicolons.
The braces around an action must be used even if the action
contains only one statement, or if it contains no statements at
all.  However, if you omit the action entirely, omit the braces as
well.  An omitted action is equivalent to @samp{@{ print $0 @}}:

@example
/foo/  @{ @}     @ii{match @code{foo}, do nothing --- empty action}
/foo/          @ii{match @code{foo}, print the record --- omitted action}
@end example

The following types of statements are supported in @command{awk}:

@table @asis
@cindex side effects, statements
@item Expressions
Call functions or assign values to variables
(@pxref{Expressions}).  Executing
this kind of statement simply computes the value of the expression.
This is useful when the expression has side effects
(@pxref{Assignment Ops}).

@item Control statements
Specify the control flow of @command{awk}
programs.  The @command{awk} language gives you C-like constructs
(@code{if}, @code{for}, @code{while}, and @code{do}) as well as a few
special ones (@pxref{Statements}).

@item Compound statements
Enclose one or more statements in braces.  A compound statement
is used in order to put several statements together in the body of an
@code{if}, @code{while}, @code{do}, or @code{for} statement.

@item Input statements
Use the @code{getline} command
(@pxref{Getline}).
Also supplied in @command{awk} are the @code{next}
statement (@pxref{Next Statement})
and the @code{nextfile} statement
(@pxref{Nextfile Statement}).

@item Output statements
Such as @code{print} and @code{printf}.
@xref{Printing}.

@item Deletion statements
For deleting array elements.
@xref{Delete}.
@end table

@node Statements
@section Control Statements in Actions
@cindex control statements
@cindex statements, control, in actions
@cindex actions, control statements in

@dfn{Control statements}, such as @code{if}, @code{while}, and so on,
control the flow of execution in @command{awk} programs.  Most of @command{awk}'s
control statements are patterned after similar statements in C.

@cindex compound statements@comma{} control statements and
@cindex statements, compound@comma{} control statements and
@cindex body, in actions
@cindex @code{@{@}} (braces), statements, grouping
@cindex braces (@code{@{@}}), statements, grouping
@cindex newlines, separating statements in actions
@cindex @code{;} (semicolon), separating statements in actions
@cindex semicolon (@code{;}), separating statements in actions
All the control statements start with special keywords, such as @code{if}
and @code{while}, to distinguish them from simple expressions.
Many control statements contain other statements.  For example, the
@code{if} statement contains another statement that may or may not be
executed.  The contained statement is called the @dfn{body}.
To include more than one statement in the body, group them into a
single @dfn{compound statement} with braces, separating them with
newlines or semicolons.

@menu
* If Statement::                Conditionally execute some @command{awk}
                                statements.
* While Statement::             Loop until some condition is satisfied.
* Do Statement::                Do specified action while looping until some
                                condition is satisfied.
* For Statement::               Another looping statement, that provides
                                initialization and increment clauses.
* Switch Statement::            Switch/case evaluation for conditional
                                execution of statements based on a value.
* Break Statement::             Immediately exit the innermost enclosing loop.
* Continue Statement::          Skip to the end of the innermost enclosing
                                loop.
* Next Statement::              Stop processing the current input record.
* Nextfile Statement::          Stop processing the current file.
* Exit Statement::              Stop execution of @command{awk}.
@end menu

@node If Statement
@subsection The @code{if}-@code{else} Statement

@cindex @code{if} statement
The @code{if}-@code{else} statement is @command{awk}'s decision-making
statement.  It looks like this:

@display
@code{if (@var{condition}) @var{then-body}} [@code{else @var{else-body}}]
@end display

@noindent
The @var{condition} is an expression that controls what the rest of the
statement does.  If the @var{condition} is true, @var{then-body} is
executed; otherwise, @var{else-body} is executed.
The @code{else} part of the statement is
optional.  The condition is considered false if its value is zero or
the null string; otherwise, the condition is true.
Refer to the following:

@example
@group
if (x % 2 == 0)
    print "x is even"
else
    print "x is odd"
@end group
@end example

In this example, if the expression @samp{x % 2 == 0} is true (i.e.,
if the value of @code{x} is evenly divisible by two), then the first
@code{print} statement is executed; otherwise, the second @code{print}
statement is executed.
If the @code{else} keyword appears on the same line as @var{then-body} and
@var{then-body} is not a compound statement (i.e., not surrounded by
braces), then a semicolon must separate @var{then-body} from
the @code{else}.
To illustrate this, the previous example can be rewritten as:

@example
if (x % 2 == 0) print "x is even"; else
        print "x is odd"
@end example

@noindent
If the @samp{;} is left out, @command{awk} can't interpret the statement and
it produces a syntax error.  Don't actually write programs this way,
because a human reader might fail to see the @code{else} if it is not
the first thing on its line.

@node While Statement
@subsection The @code{while} Statement
@cindex @code{while} statement
@cindex loops
@cindex loops, @code{while}
@cindex loops, See Also @code{while} statement

In programming, a @dfn{loop} is a part of a program that can
be executed two or more times in succession.
The @code{while} statement is the simplest looping statement in
@command{awk}.  It repeatedly executes a statement as long as a condition is
true.  For example:

@example
while (@var{condition})
  @var{body}
@end example

@cindex body, in loops
@noindent
@var{body} is a statement called the @dfn{body} of the loop,
and @var{condition} is an expression that controls how long the loop
keeps running.
The first thing the @code{while} statement does is test the @var{condition}.
If the @var{condition} is true, it executes the statement @var{body}.
@ifinfo
(The @var{condition} is true when the value
is not zero and not a null string.)
@end ifinfo
After @var{body} has been executed,
@var{condition} is tested again, and if it is still true, @var{body}
executes again.  This process repeats until the @var{condition} is no longer
true.  If the @var{condition} is initially false, the body of the loop
never executes and @command{awk} continues with the statement following
the loop.
This example prints the first three fields of each record, one per line:

@example
awk '
@{
    i = 1
    while (i <= 3) @{
        print $i
        i++
    @}
@}' inventory-shipped
@end example

@noindent
The body of this loop is a compound statement enclosed in braces,
containing two statements.
The loop works in the following manner: first, the value of @code{i} is set to one.
Then, the @code{while} statement tests whether @code{i} is less than or equal to
three.  This is true when @code{i} equals one, so the @code{i}th
field is printed.  Then the @samp{i++} increments the value of @code{i}
and the loop repeats.  The loop terminates when @code{i} reaches four.

A newline is not required between the condition and the
body; however, using one makes the program clearer unless the body is a
compound statement or else is very simple.  The newline after the open brace
that begins the compound statement is not required either, but the
program is harder to read without it.

@node Do Statement
@subsection The @code{do}-@code{while} Statement
@cindex @code{do}-@code{while} statement
@cindex loops, @code{do}-@code{while}

The @code{do} loop is a variation of the @code{while} looping statement.
The @code{do} loop executes the @var{body} once and then repeats the
@var{body} as long as the @var{condition} is true.  It looks like this:

@example
do
  @var{body}
while (@var{condition})
@end example

Even if the @var{condition} is false at the start, the @var{body}
executes at least once (and only once, unless executing @var{body}
makes @var{condition} true).  Contrast this with the corresponding
@code{while} statement:

@example
while (@var{condition})
    @var{body}
@end example

@noindent
This statement does not execute the @var{body} even once if the
@var{condition} is false to begin with.  The following is an example of
a @code{do} statement:

@example
@{
    i = 1
    do @{
        print $0
        i++
    @} while (i <= 10)
@}
@end example

@noindent
This program prints each input record 10 times.  However, it isn't a very
realistic example, because in this case an ordinary @code{while} would do
just as well.  This situation reflects actual experience; only
occasionally is there a real use for a @code{do} statement.

@node For Statement
@subsection The @code{for} Statement
@cindex @code{for} statement
@cindex loops, @code{for}, iterative

The @code{for} statement makes it more convenient to count iterations of a
loop.  The general form of the @code{for} statement looks like this:

@example
for (@var{initialization}; @var{condition}; @var{increment})
  @var{body}
@end example

@noindent
The @var{initialization}, @var{condition}, and @var{increment} parts are
arbitrary @command{awk} expressions, and @var{body} stands for any
@command{awk} statement.

The @code{for} statement starts by executing @var{initialization}.
Then, as long
as the @var{condition} is true, it repeatedly executes @var{body} and then
@var{increment}.  Typically, @var{initialization} sets a variable to
either zero or one, @var{increment} adds one to it, and @var{condition}
compares it against the desired number of iterations.
For example:

@example
awk '
@{
    for (i = 1; i <= 3; i++)
        print $i
@}' inventory-shipped
@end example

@noindent
This prints the first three fields of each input record, with one
input field per output line.

It isn't possible to
set more than one variable in the
@var{initialization} part without using a multiple assignment statement
such as @samp{x = y = 0}. This makes sense only if all the initial values
are equal.  (But it is possible to initialize additional variables by writing
their assignments as separate statements preceding the @code{for} loop.)

@c @cindex comma operator, not supported
The same is true of the @var{increment} part. Incrementing additional
variables requires separate statements at the end of the loop.
The C compound expression, using C's comma operator, is useful in
this context, but it is not supported in @command{awk}.

Most often, @var{increment} is an increment expression, as in the previous
example.  But this is not required; it can be any expression
whatsoever.  For example, the following statement prints all the powers of two
between 1 and 100:

@example
for (i = 1; i <= 100; i *= 2)
    print i
@end example

If there is nothing to be done, any of the three expressions in the
parentheses following the @code{for} keyword may be omitted.  Thus,
@w{@samp{for (; x > 0;)}} is equivalent to @w{@samp{while (x > 0)}}.  If the
@var{condition} is omitted, it is treated as true, effectively
yielding an @dfn{infinite loop} (i.e., a loop that never terminates).

In most cases, a @code{for} loop is an abbreviation for a @code{while}
loop, as shown here:

@example
@var{initialization}
while (@var{condition}) @{
  @var{body}
  @var{increment}
@}
@end example

@cindex loops, @code{continue} statements and
@noindent
The only exception is when the @code{continue} statement
(@pxref{Continue Statement}) is used
inside the loop. Changing a @code{for} statement to a @code{while}
statement in this way can change the effect of the @code{continue}
statement inside the loop.

The @command{awk} language has a @code{for} statement in addition to a
@code{while} statement because a @code{for} loop is often both less work to
type and more natural to think of.  Counting the number of iterations is
very common in loops.  It can be easier to think of this counting as part
of looping rather than as something to do inside the loop.

@cindex @code{in} operator
There is an alternative version of the @code{for} loop, for iterating over
all the indices of an array:

@example
for (i in array)
    @var{do something with} array[i]
@end example

@noindent
@xref{Scanning an Array}
for more information on this version of the @code{for} loop.

@node Switch Statement
@subsection The @code{switch} Statement
@cindex @code{switch} statement
@cindex @code{case} keyword
@cindex @code{default} keyword

This @value{SECTION} describes a @command{gawk}-specific feature.
If @command{gawk} is in compatibility mode (@pxref{Options}),
it is not available.

The @code{switch} statement allows the evaluation of an expression and
the execution of statements based on a @code{case} match. Case statements
are checked for a match in the order they are defined.  If no suitable
@code{case} is found, the @code{default} section is executed, if supplied.

Each @code{case} contains a single constant, be it numeric, string, or
regexp.  The @code{switch} expression is evaluated, and then each
@code{case}'s constant is compared against the result in turn. The type of constant
determines the comparison: numeric or string do the usual comparisons.
A regexp constant does a regular expression match against the string
value of the original expression.  The general form of the @code{switch}
statement looks like this:

@example
switch (@var{expression}) @{
case @var{value or regular expression}:
    @var{case-body}
default:
    @var{default-body}
@}
@end example

Control flow in
the @code{switch} statement works as it does in C. Once a match to a given
case is made, the case statement bodies execute until a @code{break},
@code{continue}, @code{next}, @code{nextfile}, or @code{exit} is encountered,
or the end of the @code{switch} statement itself. For example:

@example
while ((c = getopt(ARGC, ARGV, "aksx")) != -1) @{
    switch (c) @{
    case "a":
        # report size of all files
        all_files = TRUE;
        break
    case "k":
        BLOCK_SIZE = 1024       # 1K block size
        break
    case "s":
        # do sums only
        sum_only = TRUE
        break
    case "x":
        # don't cross filesystems
        fts_flags = or(fts_flags, FTS_XDEV)
        break
    case "?":
    default:
        usage()
        break
    @}
@}
@end example

Note that if none of the statements specified here halt execution
of a matched @code{case} statement, execution falls through to the
next @code{case} until execution halts. In this example, the
@code{case} for @code{"?"} falls through to the @code{default}
case, which is to call a function named @code{usage()}.
(The @code{getopt()} function being called here is
described in @ref{Getopt Function}.)

@node Break Statement
@subsection The @code{break} Statement
@cindex @code{break} statement
@cindex loops, exiting
@cindex loops, @code{break} statement and

The @code{break} statement jumps out of the innermost @code{for},
@code{while}, or @code{do} loop that encloses it.  The following example
finds the smallest divisor of any integer, and also identifies prime
numbers:

@example
@group
# find smallest divisor of num
@{
    num = $1
    for (divisor = 2; divisor * divisor <= num; divisor++) @{
        if (num % divisor == 0)
            break
    @}
@end group
@group
    if (num % divisor == 0)
        printf "Smallest divisor of %d is %d\n", num, divisor
    else
        printf "%d is prime\n", num
@}
@end group
@end example

When the remainder is zero in the first @code{if} statement, @command{awk}
immediately @dfn{breaks out} of the containing @code{for} loop.  This means
that @command{awk} proceeds immediately to the statement following the loop
and continues processing.  (This is very different from the @code{exit}
statement, which stops the entire @command{awk} program.
@xref{Exit Statement}.)

The following program illustrates how the @var{condition} of a @code{for}
or @code{while} statement could be replaced with a @code{break} inside
an @code{if}:

@example
# find smallest divisor of num
@{
    num = $1
    for (divisor = 2; ; divisor++) @{
        if (num % divisor == 0) @{
            printf "Smallest divisor of %d is %d\n", num, divisor
            break
        @}
        if (divisor * divisor > num) @{
            printf "%d is prime\n", num
            break
        @}
    @}
@}
@end example

The @code{break} statement is also used to break out of the
@code{switch} statement.
This is discussed in @ref{Switch Statement}.

@c @cindex @code{break}, outside of loops
@c @cindex historical features
@c @cindex @command{awk} language, POSIX version
@cindex POSIX @command{awk}, @code{break} statement and
@cindex dark corner, @code{break} statement
@cindex @command{gawk}, @code{break} statement in
@cindex Brian Kernighan's @command{awk}
The @code{break} statement has no meaning when
used outside the body of a loop or @code{switch}.
However, although it was never documented,
historical implementations of @command{awk} treated the @code{break}
statement outside of a loop as if it were a @code{next} statement
(@pxref{Next Statement}).
@value{DARKCORNER}
Recent versions of BWK @command{awk} no longer allow this usage,
nor does @command{gawk}.

@node Continue Statement
@subsection The @code{continue} Statement

@cindex @code{continue} statement
Similar to @code{break}, the @code{continue} statement is used only inside
@code{for}, @code{while}, and @code{do} loops.  It skips
over the rest of the loop body, causing the next cycle around the loop
to begin immediately.  Contrast this with @code{break}, which jumps out
of the loop altogether.

The @code{continue} statement in a @code{for} loop directs @command{awk} to
skip the rest of the body of the loop and resume execution with the
increment-expression of the @code{for} statement.  The following program
illustrates this fact:

@example
BEGIN @{
     for (x = 0; x <= 20; x++) @{
         if (x == 5)
             continue
         printf "%d ", x
     @}
     print ""
@}
@end example

@noindent
This program prints all the numbers from 0 to 20---except for 5, for
which the @code{printf} is skipped.  Because the increment @samp{x++}
is not skipped, @code{x} does not remain stuck at 5.  Contrast the
@code{for} loop from the previous example with the following @code{while} loop:

@example
BEGIN @{
     x = 0
     while (x <= 20) @{
         if (x == 5)
             continue
         printf "%d ", x
         x++
     @}
     print ""
@}
@end example

@noindent
This program loops forever once @code{x} reaches 5, because
the increment (@samp{x++}) is never reached.

@c @cindex @code{continue}, outside of loops
@c @cindex historical features
@c @cindex @command{awk} language, POSIX version
@cindex POSIX @command{awk}, @code{continue} statement and
@cindex dark corner, @code{continue} statement
@cindex @command{gawk}, @code{continue} statement in
@cindex Brian Kernighan's @command{awk}
The @code{continue} statement has no special meaning with respect to the
@code{switch} statement, nor does it have any meaning when used outside the
body of a loop.  Historical versions of @command{awk} treated a @code{continue}
statement outside a loop the same way they treated a @code{break}
statement outside a loop: as if it were a @code{next}
statement
@ifset FOR_PRINT
(discussed in the following @value{SECTION}).
@end ifset
@ifclear FOR_PRINT
(@pxref{Next Statement}).
@end ifclear
@value{DARKCORNER}
Recent versions of BWK @command{awk} no longer work this way, nor
does @command{gawk}.

@node Next Statement
@subsection The @code{next} Statement
@cindex @code{next} statement

The @code{next} statement forces @command{awk} to immediately stop processing
the current record and go on to the next record.  This means that no
further rules are executed for the current record, and the rest of the
current rule's action isn't executed.

Contrast this with the effect of the @code{getline} function
(@pxref{Getline}).  That also causes
@command{awk} to read the next record immediately, but it does not alter the
flow of control in any way (i.e., the rest of the current action executes
with a new input record).

@cindex @command{awk} programs, execution of
At the highest level, @command{awk} program execution is a loop that reads
an input record and then tests each rule's pattern against it.  If you
think of this loop as a @code{for} statement whose body contains the
rules, then the @code{next} statement is analogous to a @code{continue}
statement. It skips to the end of the body of this implicit loop and
executes the increment (which reads another record).

For example, suppose an @command{awk} program works only on records
with four fields, and it shouldn't fail when given bad input.  To avoid
complicating the rest of the program, write a ``weed out'' rule near
the beginning, in the following manner:

@example
NF != 4 @{
    printf("%s:%d: skipped: NF != 4\n", FILENAME, FNR) > "/dev/stderr"
    next
@}
@end example

@noindent
Because of the @code{next} statement,
the program's subsequent rules won't see the bad record.  The error
message is redirected to the standard error output stream, as error
messages should be.
For more detail, see
@ref{Special Files}.

If the @code{next} statement causes the end of the input to be reached,
then the code in any @code{END} rules is executed.
@xref{BEGIN/END}.

The @code{next} statement is not allowed inside @code{BEGINFILE} and
@code{ENDFILE} rules. @xref{BEGINFILE/ENDFILE}.

@c @cindex @command{awk} language, POSIX version
@c @cindex @code{next}, inside a user-defined function
@cindex @code{BEGIN} pattern, @code{next}/@code{nextfile} statements and
@cindex @code{END} pattern, @code{next}/@code{nextfile} statements and
@cindex POSIX @command{awk}, @code{next}/@code{nextfile} statements and
@cindex @code{next} statement, user-defined functions and
@cindex functions, user-defined, @code{next}/@code{nextfile} statements and
According to the POSIX standard, the behavior is undefined if the
@code{next} statement is used in a @code{BEGIN} or @code{END} rule.
@command{gawk} treats it as a syntax error.  Although POSIX does not disallow it,
most other @command{awk} implementations don't allow the @code{next}
statement inside function bodies (@pxref{User-defined}).  Just as with any
other @code{next} statement, a @code{next} statement inside a function
body reads the next record and starts processing it with the first rule
in the program.

@node Nextfile Statement
@subsection The @code{nextfile} Statement
@cindex @code{nextfile} statement

The @code{nextfile} statement
is similar to the @code{next} statement.
However, instead of abandoning processing of the current record, the
@code{nextfile} statement instructs @command{awk} to stop processing the
current @value{DF}.

Upon execution of the @code{nextfile} statement,
@code{FILENAME} is
updated to the name of the next @value{DF} listed on the command line,
@code{FNR} is reset to one,
and processing
starts over with the first rule in the program.
If the @code{nextfile} statement causes the end of the input to be reached,
then the code in any @code{END} rules is executed. An exception to this is
when @code{nextfile} is invoked during execution of any statement in an
@code{END} rule; in this case, it causes the program to stop immediately.
@xref{BEGIN/END}.

The @code{nextfile} statement is useful when there are many @value{DF}s
to process but it isn't necessary to process every record in every file.
Without @code{nextfile},
in order to move on to the next @value{DF}, a program
would have to continue scanning the unwanted records.  The @code{nextfile}
statement accomplishes this much more efficiently.

In @command{gawk}, execution of @code{nextfile} causes additional things
to happen: any @code{ENDFILE} rules are executed if @command{gawk} is
not currently in an @code{END} or @code{BEGINFILE} rule, @code{ARGIND} is
incremented, and any @code{BEGINFILE} rules are executed.  (@code{ARGIND}
hasn't been introduced yet. @xref{Built-in Variables}.)

With @command{gawk}, @code{nextfile} is useful inside a @code{BEGINFILE}
rule to skip over a file that would otherwise cause @command{gawk}
to exit with a fatal error. In this case, @code{ENDFILE} rules are not
executed. @xref{BEGINFILE/ENDFILE}.

Although it might seem that @samp{close(FILENAME)} would accomplish
the same as @code{nextfile}, this isn't true.  @code{close()} is
reserved for closing files, pipes, and coprocesses that are
opened with redirections.  It is not related to the main processing that
@command{awk} does with the files listed in @code{ARGV}.

@quotation NOTE
For many years, @code{nextfile} was a
common extension. In September 2012, it was accepted for
inclusion into the POSIX standard.
See @uref{http://austingroupbugs.net/view.php?id=607, the Austin Group website}.
@end quotation

@cindex functions, user-defined, @code{next}/@code{nextfile} statements and
@cindex @code{nextfile} statement, user-defined functions and
@cindex Brian Kernighan's @command{awk}
@cindex @command{mawk} utility
The current version of BWK @command{awk} and @command{mawk}
also support @code{nextfile}.  However, they don't allow the
@code{nextfile} statement inside function bodies (@pxref{User-defined}).
@command{gawk} does; a @code{nextfile} inside a function body reads the
first record from the next file and starts processing it with the first
rule in the program, just as any other @code{nextfile} statement.

@node Exit Statement
@subsection The @code{exit} Statement

@cindex @code{exit} statement
The @code{exit} statement causes @command{awk} to immediately stop
executing the current rule and to stop processing input; any remaining input
is ignored.  The @code{exit} statement is written as follows:

@display
@code{exit} [@var{return code}]
@end display

@cindex @code{BEGIN} pattern, @code{exit} statement and
@cindex @code{END} pattern, @code{exit} statement and
When an @code{exit} statement is executed from a @code{BEGIN} rule, the
program stops processing everything immediately.  No input records are
read.  However, if an @code{END} rule is present,
as part of executing the @code{exit} statement,
the @code{END} rule is executed
(@pxref{BEGIN/END}).
If @code{exit} is used in the body of an @code{END} rule, it causes
the program to stop immediately.

An @code{exit} statement that is not part of a @code{BEGIN} or @code{END}
rule stops the execution of any further automatic rules for the current
record, skips reading any remaining input records, and executes the
@code{END} rule if there is one.  @command{gawk} also skips
any @code{ENDFILE} rules; they do not execute.

In such a case,
if you don't want the @code{END} rule to do its job, set a variable
to a nonzero value before the @code{exit} statement and check that variable in
the @code{END} rule.
@xref{Assert Function}
for an example that does this.

@cindex dark corner, @code{exit} statement
If an argument is supplied to @code{exit}, its value is used as the exit
status code for the @command{awk} process.  If no argument is supplied,
@code{exit} causes @command{awk} to return a ``success'' status.
In the case where an argument
is supplied to a first @code{exit} statement, and then @code{exit} is
called a second time from an @code{END} rule with no argument,
@command{awk} uses the previously supplied exit value.  @value{DARKCORNER}
@xref{Exit Status} for more information.

@cindex programming conventions, @code{exit} statement
For example, suppose an error condition occurs that is difficult or
impossible to handle.  Conventionally, programs report this by
exiting with a nonzero status.  An @command{awk} program can do this
using an @code{exit} statement with a nonzero argument, as shown
in the following example:

@example
@group
BEGIN @{
    if (("date" | getline date_now) <= 0) @{
        print "Can't get system date" > "/dev/stderr"
        exit 1
    @}
@end group
@group
    print "current date is", date_now
    close("date")
@}
@end group
@end example

@quotation NOTE
For full portability, exit values should be between zero and 126, inclusive.
Negative values, and values of 127 or greater, may not produce consistent
results across different operating systems.
@end quotation


@node Built-in Variables
@section Predefined Variables
@cindex predefined variables
@cindex variables, predefined

Most @command{awk} variables are available to use for your own
purposes; they never change unless your program assigns values to
them, and they never affect anything unless your program examines them.
However, a few variables in @command{awk} have special built-in meanings.
@command{awk} examines some of these automatically, so that they enable you
to tell @command{awk} how to do certain things.  Others are set
automatically by @command{awk}, so that they carry information from the
internal workings of @command{awk} to your program.

@cindex @command{gawk}, predefined variables and
This @value{SECTION} documents all of @command{gawk}'s predefined variables,
most of which are also documented in the @value{CHAPTER}s describing
their areas of activity.

@menu
* User-modified::               Built-in variables that you change to control
                                @command{awk}.
* Auto-set::                    Built-in variables where @command{awk} gives
                                you information.
* ARGC and ARGV::               Ways to use @code{ARGC} and @code{ARGV}.
@end menu

@node User-modified
@subsection Built-in Variables That Control @command{awk}
@cindex predefined variables, user-modifiable
@cindex user-modifiable variables

The following is an alphabetical list of variables that you can change to
control how @command{awk} does certain things.

The variables that are specific to @command{gawk} are marked with a pound
sign (@samp{#}).  These variables are @command{gawk} extensions.  In other
@command{awk} implementations or if @command{gawk} is in compatibility
mode (@pxref{Options}), they are not special.  (Any exceptions are noted
in the description of each variable.)

@table @code
@cindex @code{BINMODE} variable
@cindex binary input/output
@cindex input/output, binary
@cindex differences in @command{awk} and @command{gawk}, @code{BINMODE} variable
@item BINMODE #
On non-POSIX systems, this variable specifies use of binary mode
for all I/O.  Numeric values of one, two, or three specify that input
files, output files, or all files, respectively, should use binary I/O.
A numeric value less than zero is treated as zero, and a numeric value
greater than three is treated as three.  Alternatively, string values
of @code{"r"} or @code{"w"} specify that input files and output files,
respectively, should use binary I/O.  A string value of @code{"rw"} or
@code{"wr"} indicates that all files should use binary I/O.  Any other
string value is treated the same as @code{"rw"}, but causes @command{gawk}
to generate a warning message.  @code{BINMODE} is described in more
detail in @ref{PC Using}.  @command{mawk} (@pxref{Other Versions})
also supports this variable, but only using numeric values.

@cindex @code{CONVFMT} variable
@cindex POSIX @command{awk}, @code{CONVFMT} variable and
@cindex numbers, converting, to strings
@cindex strings, converting, numbers to
@item @code{CONVFMT}
A string that controls the conversion of numbers to
strings (@pxref{Conversion}).
It works by being passed, in effect, as the first argument to the
@code{sprintf()} function
(@pxref{String Functions}).
Its default value is @code{"%.6g"}.
@code{CONVFMT} was introduced by the POSIX standard.

@cindex @command{gawk}, @code{FIELDWIDTHS} variable in
@cindex @code{FIELDWIDTHS} variable
@cindex differences in @command{awk} and @command{gawk}, @code{FIELDWIDTHS} variable
@cindex field separators, @code{FIELDWIDTHS} variable and
@cindex separators, field, @code{FIELDWIDTHS} variable and
@item FIELDWIDTHS #
A space-separated list of columns that tells @command{gawk}
how to split input with fixed columnar boundaries.
Starting in @value{PVERSION} 4.2, each field width may optionally be
preceded by a colon-separated value specifying the number of characters to skip
before the field starts.
Assigning a value to @code{FIELDWIDTHS}
overrides the use of @code{FS} and @code{FPAT} for field splitting.
@xref{Constant Size} for more information.

@cindex @command{gawk}, @code{FPAT} variable in
@cindex @code{FPAT} variable
@cindex differences in @command{awk} and @command{gawk}, @code{FPAT} variable
@cindex field separators, @code{FPAT} variable and
@cindex separators, field, @code{FPAT} variable and
@item FPAT #
A regular expression (as a string) that tells @command{gawk}
to create the fields based on text that matches the regular expression.
Assigning a value to @code{FPAT}
overrides the use of @code{FS} and @code{FIELDWIDTHS} for field splitting.
@xref{Splitting By Content} for more information.

@cindex @code{FS} variable
@cindex separators, field
@cindex field separators
@item FS
The input field separator (@pxref{Field Separators}).
The value is a single-character string or a multicharacter regular
expression that matches the separations between fields in an input
record.  If the value is the null string (@code{""}), then each
character in the record becomes a separate field.
(This behavior is a @command{gawk} extension. POSIX @command{awk} does not
specify the behavior when @code{FS} is the null string.
Nonetheless, some other versions of @command{awk} also treat
@code{""} specially.)

The default value is @w{@code{" "}}, a string consisting of a single
space.  As a special exception, this value means that any sequence of
spaces, TABs, and/or newlines is a single separator.  It also causes
spaces, TABs, and newlines at the beginning and end of a record to
be ignored.

You can set the value of @code{FS} on the command line using the
@option{-F} option:

@example
awk -F, '@var{program}' @var{input-files}
@end example

@cindex @command{gawk}, field separators and
If @command{gawk} is using @code{FIELDWIDTHS} or @code{FPAT}
for field splitting,
assigning a value to @code{FS} causes @command{gawk} to return to
the normal, @code{FS}-based field splitting. An easy way to do this
is to simply say @samp{FS = FS}, perhaps with an explanatory comment.

@cindex @command{gawk}, @code{IGNORECASE} variable in
@cindex @code{IGNORECASE} variable
@cindex differences in @command{awk} and @command{gawk}, @code{IGNORECASE} variable
@cindex case sensitivity, and string comparisons
@cindex case sensitivity, and regexps
@cindex regular expressions, case sensitivity
@item IGNORECASE #
If @code{IGNORECASE} is nonzero or non-null, then all string comparisons
and all regular expression matching are case-independent.
This applies to
regexp matching with @samp{~} and @samp{!~},
the @code{gensub()}, @code{gsub()}, @code{index()}, @code{match()},
@code{patsplit()}, @code{split()}, and @code{sub()} functions,
record termination with @code{RS}, and field splitting with
@code{FS} and @code{FPAT}.
However, the value of @code{IGNORECASE} does @emph{not} affect array subscripting
and it does not affect field splitting when using a single-character
field separator.
@xref{Case-sensitivity}.

@cindex @command{gawk}, @code{LINT} variable in
@cindex @code{LINT} variable
@cindex differences in @command{awk} and @command{gawk}, @code{LINT} variable
@cindex lint checking
@item LINT #
When this variable is true (nonzero or non-null), @command{gawk}
behaves as if the @option{--lint} command-line option is in effect
(@pxref{Options}).
With a value of @code{"fatal"}, lint warnings become fatal errors.
With a value of @code{"invalid"}, only warnings about things that are
actually invalid are issued. (This is not fully implemented yet.)
Any other true value prints nonfatal warnings.
Assigning a false value to @code{LINT} turns off the lint warnings.

This variable is a @command{gawk} extension.  It is not special
in other @command{awk} implementations.  Unlike with the other special variables,
changing @code{LINT} does affect the production of lint warnings,
even if @command{gawk} is in compatibility mode.  Much as
the @option{--lint} and @option{--traditional} options independently
control different aspects of @command{gawk}'s behavior, the control
of lint warnings during program execution is independent of the flavor
of @command{awk} being executed.

@cindex @code{OFMT} variable
@cindex numbers, converting, to strings
@cindex strings, converting, numbers to
@item OFMT
A string that controls conversion of numbers to
strings (@pxref{Conversion}) for
printing with the @code{print} statement.  It works by being passed
as the first argument to the @code{sprintf()} function
(@pxref{String Functions}).
Its default value is @code{"%.6g"}.  Earlier versions of @command{awk}
used @code{OFMT} to specify the format for converting numbers to
strings in general expressions; this is now done by @code{CONVFMT}.

@cindex @code{sprintf()} function, @code{OFMT} variable and
@cindex @code{print} statement, @code{OFMT} variable and
@cindex @code{OFS} variable
@cindex separators, field
@cindex field separators
@item OFS
The output field separator (@pxref{Output Separators}).  It is
output between the fields printed by a @code{print} statement.  Its
default value is @w{@code{" "}}, a string consisting of a single space.

@cindex @code{ORS} variable
@item ORS
The output record separator.  It is output at the end of every
@code{print} statement.  Its default value is @code{"\n"}, the newline
character.  (@xref{Output Separators}.)

@cindex @code{PREC} variable
@item PREC #
The working precision of arbitrary-precision floating-point numbers,
53 bits by default (@pxref{Setting precision}).

@cindex @code{ROUNDMODE} variable
@item ROUNDMODE #
The rounding mode to use for arbitrary-precision arithmetic on
numbers, by default @code{"N"} (@code{roundTiesToEven} in
the IEEE 754 standard; @pxref{Setting the rounding mode}).

@cindex @code{RS} variable
@cindex separators, for records
@cindex record separators
@item @code{RS}
The input record separator.  Its default value is a string
containing a single newline character, which means that an input record
consists of a single line of text.
It can also be the null string, in which case records are separated by
runs of blank lines.
If it is a regexp, records are separated by
matches of the regexp in the input text.
(@xref{Records}.)

The ability for @code{RS} to be a regular expression
is a @command{gawk} extension.
In most other @command{awk} implementations,
or if @command{gawk} is in compatibility mode
(@pxref{Options}),
just the first character of @code{RS}'s value is used.

@cindex @code{SUBSEP} variable
@cindex separators, subscript
@cindex subscript separators
@item @code{SUBSEP}
The subscript separator.  It has the default value of
@code{"\034"} and is used to separate the parts of the indices of a
multidimensional array.  Thus, the expression @samp{@w{foo["A", "B"]}}
really accesses @code{foo["A\034B"]}
(@pxref{Multidimensional}).

@cindex @command{gawk}, @code{TEXTDOMAIN} variable in
@cindex @code{TEXTDOMAIN} variable
@cindex differences in @command{awk} and @command{gawk}, @code{TEXTDOMAIN} variable
@cindex internationalization, localization
@item TEXTDOMAIN #
Used for internationalization of programs at the
@command{awk} level.  It sets the default text domain for specially
marked string constants in the source text, as well as for the
@code{dcgettext()}, @code{dcngettext()}, and @code{bindtextdomain()} functions
(@pxref{Internationalization}).
The default value of @code{TEXTDOMAIN} is @code{"messages"}.
@end table

@node Auto-set
@subsection Built-in Variables That Convey Information

@cindex predefined variables, conveying information
@cindex variables, predefined conveying information
The following is an alphabetical list of variables that @command{awk}
sets automatically on certain occasions in order to provide
information to your program.

The variables that are specific to @command{gawk} are marked with a pound
sign (@samp{#}).  These variables are @command{gawk} extensions.  In other
@command{awk} implementations or if @command{gawk} is in compatibility
mode (@pxref{Options}), they are not special:

@c @asis for docbook
@table @asis
@cindex @code{ARGC}/@code{ARGV} variables
@cindex arguments, command-line
@cindex command line, arguments
@item @code{ARGC}, @code{ARGV}
The command-line arguments available to @command{awk} programs are stored in
an array called @code{ARGV}.  @code{ARGC} is the number of command-line
arguments present.  @xref{Other Arguments}.
Unlike most @command{awk} arrays,
@code{ARGV} is indexed from 0 to @code{ARGC} @minus{} 1.
In the following example:

@example
@group
$ @kbd{awk 'BEGIN @{}
>         @kbd{for (i = 0; i < ARGC; i++)}
>             @kbd{print ARGV[i]}
>      @kbd{@}' inventory-shipped mail-list}
@print{} awk
@print{} inventory-shipped
@print{} mail-list
@end group
@end example

@noindent
@code{ARGV[0]} contains @samp{awk}, @code{ARGV[1]}
contains @samp{inventory-shipped}, and @code{ARGV[2]} contains
@samp{mail-list}.  The value of @code{ARGC} is three, one more than the
index of the last element in @code{ARGV}, because the elements are numbered
from zero.

@cindex programming conventions, @code{ARGC}/@code{ARGV} variables
The names @code{ARGC} and @code{ARGV}, as well as the convention of indexing
the array from 0 to @code{ARGC} @minus{} 1, are derived from the C language's
method of accessing command-line arguments.

@cindex dark corner, value of @code{ARGV[0]}
The value of @code{ARGV[0]} can vary from system to system.
Also, you should note that the program text is @emph{not} included in
@code{ARGV}, nor are any of @command{awk}'s command-line options.
@xref{ARGC and ARGV} for information
about how @command{awk} uses these variables.
@value{DARKCORNER}

@cindex @code{ARGIND} variable
@cindex differences in @command{awk} and @command{gawk}, @code{ARGIND} variable
@item @code{ARGIND #}
The index in @code{ARGV} of the current file being processed.
Every time @command{gawk} opens a new @value{DF} for processing, it sets
@code{ARGIND} to the index in @code{ARGV} of the @value{FN}.
When @command{gawk} is processing the input files,
@samp{FILENAME == ARGV[ARGIND]} is always true.

@cindex files, processing@comma{} @code{ARGIND} variable and
This variable is useful in file processing; it allows you to tell how far
along you are in the list of @value{DF}s as well as to distinguish between
successive instances of the same @value{FN} on the command line.

@cindex file names, distinguishing
While you can change the value of @code{ARGIND} within your @command{awk}
program, @command{gawk} automatically sets it to a new value when it
opens the next file.

@cindex @code{ENVIRON} array
@cindex environment variables, in @code{ENVIRON} array
@item @code{ENVIRON}
An associative array containing the values of the environment.  The array
indices are the environment variable names; the elements are the values of
the particular environment variables.  For example,
@code{ENVIRON["HOME"]} might be @code{/home/arnold}.

For POSIX @command{awk}, changing this array does not affect the
environment passed on to any programs that @command{awk} may spawn via
redirection or the @code{system()} function.

However, beginning with @value{PVERSION} 4.2, if not in POSIX
compatibility mode, @command{gawk} does update its own environment when
@code{ENVIRON} is changed, thus changing the environment seen by programs
that it creates.  You should therefore be especially careful if you
modify @code{ENVIRON["PATH"]}, which is the search path for finding
executable programs.

This can also affect the running @command{gawk} program, since some of the
built-in functions may pay attention to certain environment variables.
The most notable instance of this is @code{mktime()} (@pxref{Time
Functions}), which pays attention the value of the @env{TZ} environment
variable on many systems.

Some operating systems may not have environment variables.
On such systems, the @code{ENVIRON} array is empty (except for
@w{@code{ENVIRON["AWKPATH"]}} and
@w{@code{ENVIRON["AWKLIBPATH"]}};
@pxref{AWKPATH Variable} and
@ifdocbook
@ref{AWKLIBPATH Variable}).
@end ifdocbook
@ifnotdocbook
@pxref{AWKLIBPATH Variable}).
@end ifnotdocbook

@cindex @command{gawk}, @code{ERRNO} variable in
@cindex @code{ERRNO} variable
@cindex differences in @command{awk} and @command{gawk}, @code{ERRNO} variable
@cindex error handling, @code{ERRNO} variable and
@item @code{ERRNO #}
If a system error occurs during a redirection for @code{getline}, during
a read for @code{getline}, or during a @code{close()} operation, then
@code{ERRNO} contains a string describing the error.

In addition, @command{gawk} clears @code{ERRNO} before opening each
command-line input file. This enables checking if the file is readable
inside a @code{BEGINFILE} pattern (@pxref{BEGINFILE/ENDFILE}).

Otherwise, @code{ERRNO} works similarly to the C variable @code{errno}.
Except for the case just mentioned, @command{gawk} @emph{never} clears
it (sets it to zero or @code{""}).  Thus, you should only expect its
value to be meaningful when an I/O operation returns a failure value,
such as @code{getline} returning @minus{}1.  You are, of course, free
to clear it yourself before doing an I/O operation.

If the value of @code{ERRNO} corresponds to a system error in the C
@code{errno} variable, then @code{PROCINFO["errno"]} will be set to the value
of @code{errno}.  For non-system errors, @code{PROCINFO["errno"]} will
be zero.

@cindex @code{FILENAME} variable
@cindex dark corner, @code{FILENAME} variable
@item @code{FILENAME}
The name of the current input file.  When no @value{DF}s are listed
on the command line, @command{awk} reads from the standard input and
@code{FILENAME} is set to @code{"-"}.  @code{FILENAME} changes each
time a new file is read (@pxref{Reading Files}).  Inside a @code{BEGIN}
rule, the value of @code{FILENAME} is @code{""}, because there are no input
files being processed yet.@footnote{Some early implementations of Unix
@command{awk} initialized @code{FILENAME} to @code{"-"}, even if there
were @value{DF}s to be processed. This behavior was incorrect and should
not be relied upon in your programs.} @value{DARKCORNER} Note, though,
that using @code{getline} (@pxref{Getline}) inside a @code{BEGIN} rule
can give @code{FILENAME} a value.

@cindex @code{FNR} variable
@item @code{FNR}
The current record number in the current file.  @command{awk} increments
@code{FNR} each time it reads a new record (@pxref{Records}).
@command{awk} resets @code{FNR} to zero each time it starts a new
input file.

@cindex @code{NF} variable
@item @code{NF}
The number of fields in the current input record.
@code{NF} is set each time a new record is read, when a new field is
created, or when @code{$0} changes (@pxref{Fields}).

Unlike most of the variables described in this @value{SUBSECTION},
assigning a value to @code{NF} has the potential to affect
@command{awk}'s internal workings.  In particular, assignments
to @code{NF} can be used to create fields in or remove fields from the
current record. @xref{Changing Fields}.

@cindex @code{FUNCTAB} array
@cindex @command{gawk}, @code{FUNCTAB} array in
@cindex differences in @command{awk} and @command{gawk}, @code{FUNCTAB} variable
@item @code{FUNCTAB #}
An array whose indices and corresponding values are the names of all
the built-in, user-defined, and extension functions in the program.

@quotation NOTE
Attempting to use the @code{delete} statement with the @code{FUNCTAB}
array causes a fatal error.  Any attempt to assign to an element of
@code{FUNCTAB} also causes a fatal error.
@end quotation

@cindex @code{NR} variable
@item @code{NR}
The number of input records @command{awk} has processed since
the beginning of the program's execution
(@pxref{Records}).
@command{awk} increments @code{NR} each time it reads a new record.

@cindex @command{gawk}, @code{PROCINFO} array in
@cindex @code{PROCINFO} array
@cindex differences in @command{awk} and @command{gawk}, @code{PROCINFO} array
@item @code{PROCINFO #}
The elements of this array provide access to information about the
running @command{awk} program.
The following elements (listed alphabetically)
are guaranteed to be available:

@table @code
@item PROCINFO["argv"]
@cindex command line arguments, @code{PROCINFO["argv"}
The @code{PROCINFO["argv"]} array contains all of the command-line arguments
(after glob expansion and redirection processing on platforms where that must
be done manually by the program) with subscripts ranging from 0 through
@code{argc} @minus{} 1.  For example, @code{PROCINFO["argv"][0]} will contain
the name by which @command{gawk} was invoked.  Here is an example of how this
feature may be used:

@example
gawk '
BEGIN @{
        for (i = 0; i < length(PROCINFO["argv"]); i++)
                print i, PROCINFO["argv"][i]
@}'
@end example

Please note that this differs from the standard @code{ARGV} array which does
not include command-line arguments that have already been processed by
@command{gawk} (@pxref{ARGC and ARGV}).

@cindex effective group ID of @command{gawk} user
@item PROCINFO["egid"]
The value of the @code{getegid()} system call.

@item PROCINFO["errno"]
The value of the C @code{errno} variable when @code{ERRNO} is set to
the associated error message.

@item PROCINFO["euid"]
@cindex effective user ID of @command{gawk} user
The value of the @code{geteuid()} system call.

@item PROCINFO["FS"]
This is
@code{"FS"} if field splitting with @code{FS} is in effect,
@code{"FIELDWIDTHS"} if field splitting with @code{FIELDWIDTHS} is in effect,
@code{"FPAT"} if field matching with @code{FPAT} is in effect,
or @code{"API"} if field splitting is controlled by an API input parser.

@item PROCINFO["gid"]
@cindex group ID of @command{gawk} user
The value of the @code{getgid()} system call.

@item PROCINFO["identifiers"]
@cindex program identifiers
A subarray, indexed by the names of all identifiers used in the text of
the @command{awk} program.  An @dfn{identifier} is simply the name of a variable
(be it scalar or array), built-in function, user-defined function, or
extension function.  For each identifier, the value of the element is
one of the following:

@table @code
@item "array"
The identifier is an array.

@item "builtin"
The identifier is a built-in function.

@item "extension"
The identifier is an extension function loaded via
@code{@@load} or @option{-l}.

@item "scalar"
The identifier is a scalar.

@item "untyped"
The identifier is untyped (could be used as a scalar or an array;
@command{gawk} doesn't know yet).

@item "user"
The identifier is a user-defined function.
@end table

@noindent
The values indicate what @command{gawk} knows about the identifiers
after it has finished parsing the program; they are @emph{not} updated
while the program runs.

@item PROCINFO["pgrpid"]
@cindex process group ID of @command{gawk} process
The process group ID of the current process.

@item PROCINFO["pid"]
@cindex process ID of @command{gawk} process
The process ID of the current process.

@item PROCINFO["ppid"]
@cindex parent process ID of @command{gawk} process
The parent process ID of the current process.

@item PROCINFO["strftime"]
The default time format string for @code{strftime()}.
Assigning a new value to this element changes the default.
@xref{Time Functions}.

@item PROCINFO["uid"]
The value of the @code{getuid()} system call.

@item PROCINFO["version"]
@cindex version of @command{gawk}
@cindex @command{gawk} version
The version of @command{gawk}.
@end table

The following additional elements in the array
are available to provide information about the MPFR and GMP libraries
if your version of @command{gawk} supports arbitrary-precision arithmetic
(@pxref{Arbitrary Precision Arithmetic}):

@table @code
@item PROCINFO["gmp_version"]
@cindex version of GNU MP library
The version of the GNU MP library.

@cindex version of GNU MPFR library
@item PROCINFO["mpfr_version"]
The version of the GNU MPFR library.

@item PROCINFO["prec_max"]
@cindex maximum precision supported by MPFR library
The maximum precision supported by MPFR.

@item PROCINFO["prec_min"]
@cindex minimum precision required by MPFR library
The minimum precision required by MPFR.
@end table

The following additional elements in the array are available to provide
information about the version of the extension API, if your version
of @command{gawk} supports dynamic loading of extension functions
(@pxref{Dynamic Extensions}):

@table @code
@item PROCINFO["api_major"]
@cindex version of @command{gawk} extension API
@cindex extension API, version number
The major version of the extension API.

@item PROCINFO["api_minor"]
The minor version of the extension API.
@end table

@cindex supplementary groups of @command{gawk} process
On some systems, there may be elements in the array, @code{"group1"}
through @code{"group@var{N}"} for some @var{N}. @var{N} is the number of
supplementary groups that the process has.  Use the @code{in} operator
to test for these elements
(@pxref{Reference to Elements}).

The following elements allow you to change @command{gawk}'s behavior:

@table @code
@item PROCINFO["NONFATAL"]
If this element exists, then I/O errors for all redirections become nonfatal.
@xref{Nonfatal}.

@item PROCINFO["@var{name}", "NONFATAL"]
Make I/O errors for @var{name} be nonfatal.
@xref{Nonfatal}.

@item PROCINFO["@var{command}", "pty"]
For two-way communication to @var{command}, use a pseudo-tty instead
of setting up a two-way pipe.
@xref{Two-way I/O} for more information.

@item PROCINFO["@var{input_name}", "READ_TIMEOUT"]
Set a timeout for reading from input redirection @var{input_name}.
@xref{Read Timeout} for more information.

@item PROCINFO["@var{input_name}", "RETRY"]
If an I/O error that may be retried occurs when reading data from
@var{input_name}, and this array entry exists, then @code{getline} returns
@minus{}2 instead of following the default behavior of returning @minus{}1
and configuring @var{input_name} to return no further data.  An I/O error
that may be retried is one where @code{errno} has the value @code{EAGAIN},
@code{EWOULDBLOCK}, @code{EINTR}, or @code{ETIMEDOUT}.  This may be useful
in conjunction with @code{PROCINFO["@var{input_name}", "READ_TIMEOUT"]}
or situations where a file descriptor has been configured to behave in
a non-blocking fashion.
@xref{Retrying Input} for more information.

@item PROCINFO["sorted_in"]
If this element exists in @code{PROCINFO}, its value controls the
order in which array indices will be processed by
@samp{for (@var{indx} in @var{array})} loops.
This is an advanced feature, so we defer the
full description until later; see
@ref{Scanning an Array}.
@end table

@cindex @code{RLENGTH} variable
@item @code{RLENGTH}
The length of the substring matched by the
@code{match()} function
(@pxref{String Functions}).
@code{RLENGTH} is set by invoking the @code{match()} function.  Its value
is the length of the matched string, or @minus{}1 if no match is found.

@cindex @code{RSTART} variable
@item @code{RSTART}
The start index in characters of the substring that is matched by the
@code{match()} function
(@pxref{String Functions}).
@code{RSTART} is set by invoking the @code{match()} function.  Its value
is the position of the string where the matched substring starts, or zero
if no match was found.

@cindex @command{gawk}, @code{RT} variable in
@cindex @code{RT} variable
@cindex differences in @command{awk} and @command{gawk}, @code{RT} variable
@item @code{RT #}
The input text that matched the text denoted by @code{RS},
the record separator.  It is set every time a record is read.

@cindex @command{gawk}, @code{SYMTAB} array in
@cindex @code{SYMTAB} array
@cindex differences in @command{awk} and @command{gawk}, @code{SYMTAB} variable
@item @code{SYMTAB #}
An array whose indices are the names of all defined global variables and
arrays in the program.  @code{SYMTAB} makes @command{gawk}'s symbol table
visible to the @command{awk} programmer.  It is built as @command{gawk}
parses the program and is complete before the program starts to run.

The array may be used for indirect access to read or write the value of
a variable:

@example
foo = 5
SYMTAB["foo"] = 4
print foo    # prints 4
@end example

@noindent
The @code{isarray()} function (@pxref{Type Functions}) may be used to test
if an element in @code{SYMTAB} is an array.
Also, you may not use the @code{delete} statement with the
@code{SYMTAB} array.

You may use an index for @code{SYMTAB} that is not a predefined identifier:

@example
SYMTAB["xxx"] = 5
print SYMTAB["xxx"]
@end example

@noindent
This works as expected: in this case @code{SYMTAB} acts just like
a regular array.  The only difference is that you can't then delete
@code{SYMTAB["xxx"]}.

@cindex Schorr, Andrew
The @code{SYMTAB} array is more interesting than it looks. Andrew Schorr
points out that it effectively gives @command{awk} data pointers. Consider his
example:

@example
@group
# Indirect multiply of any variable by amount, return result

function multiply(variable, amount)
@{
    return SYMTAB[variable] *= amount
@}
@end group
@end example

@noindent
You would use it like this:

@example
BEGIN @{
    answer = 10.5
    multiply("answer", 4)
    print "The answer is", answer
@}
@end example

@noindent
When run, this produces:

@example
$ @kbd{gawk -f answer.awk}
@print{} The answer is 42
@end example

@quotation NOTE
In order to avoid severe time-travel paradoxes,@footnote{Not to mention
difficult implementation issues.} neither @code{FUNCTAB} nor @code{SYMTAB}
is available as an element within the @code{SYMTAB} array.
@end quotation
@end table

@sidebar Changing @code{NR} and @code{FNR}
@cindex @code{NR} variable, changing
@cindex @code{FNR} variable, changing
@cindex dark corner, @code{FNR}/@code{NR} variables
@command{awk} increments @code{NR} and @code{FNR}
each time it reads a record, instead of setting them to the absolute
value of the number of records read.  This means that a program can
change these variables and their new values are incremented for
each record.
@value{DARKCORNER}
The following example shows this:

@example
$ @kbd{echo '1}
> @kbd{2}
> @kbd{3}
> @kbd{4' | awk 'NR == 2 @{ NR = 17 @}}
> @kbd{@{ print NR @}'}
@print{} 1
@print{} 17
@print{} 18
@print{} 19
@end example

@noindent
Before @code{FNR} was added to the @command{awk} language
(@pxref{V7/SVR3.1}),
many @command{awk} programs used this feature to track the number of
records in a file by resetting @code{NR} to zero when @code{FILENAME}
changed.
@end sidebar

@node ARGC and ARGV
@subsection Using @code{ARGC} and @code{ARGV}
@cindex @code{ARGC}/@code{ARGV} variables, how to use
@cindex arguments, command-line
@cindex command line, arguments

@ref{Auto-set}
presented the following program describing the information contained in @code{ARGC}
and @code{ARGV}:

@example
@group
$ @kbd{awk 'BEGIN @{}
>        @kbd{for (i = 0; i < ARGC; i++)}
>            @kbd{print ARGV[i]}
>      @kbd{@}' inventory-shipped mail-list}
@print{} awk
@print{} inventory-shipped
@print{} mail-list
@end group
@end example

@noindent
In this example, @code{ARGV[0]} contains @samp{awk}, @code{ARGV[1]}
contains @samp{inventory-shipped}, and @code{ARGV[2]} contains
@samp{mail-list}.
Notice that the @command{awk} program is not entered in @code{ARGV}.  The
other command-line options, with their arguments, are also not
entered.  This includes variable assignments done with the @option{-v}
option (@pxref{Options}).
Normal variable assignments on the command line @emph{are}
treated as arguments and do show up in the @code{ARGV} array.
Given the following program in a file named @file{showargs.awk}:

@example
BEGIN @{
    printf "A=%d, B=%d\n", A, B
    for (i = 0; i < ARGC; i++)
        printf "\tARGV[%d] = %s\n", i, ARGV[i]
@}
END   @{ printf "A=%d, B=%d\n", A, B @}
@end example

@noindent
Running it produces the following:

@example
$ @kbd{awk -v A=1 -f showargs.awk B=2 /dev/null}
@print{} A=1, B=0
@print{}        ARGV[0] = awk
@print{}        ARGV[1] = B=2
@print{}        ARGV[2] = /dev/null
@print{} A=1, B=2
@end example

A program can alter @code{ARGC} and the elements of @code{ARGV}.
Each time @command{awk} reaches the end of an input file, it uses the next
element of @code{ARGV} as the name of the next input file.  By storing a
different string there, a program can change which files are read.
Use @code{"-"} to represent the standard input.  Storing
additional elements and incrementing @code{ARGC} causes
additional files to be read.

If the value of @code{ARGC} is decreased, that eliminates input files
from the end of the list.  By recording the old value of @code{ARGC}
elsewhere, a program can treat the eliminated arguments as
something other than @value{FN}s.

To eliminate a file from the middle of the list, store the null string
(@code{""}) into @code{ARGV} in place of the file's name.  As a
special feature, @command{awk} ignores @value{FN}s that have been
replaced with the null string.
Another option is to
use the @code{delete} statement to remove elements from
@code{ARGV} (@pxref{Delete}).

All of these actions are typically done in the @code{BEGIN} rule,
before actual processing of the input begins.
@xref{Split Program} and
@ifnotdocbook
@pxref{Tee Program}
@end ifnotdocbook
@ifdocbook
@ref{Tee Program}
@end ifdocbook
for examples
of each way of removing elements from @code{ARGV}.

To actually get options into an @command{awk} program,
end the @command{awk} options with @option{--} and then supply
the @command{awk} program's options, in the following manner:

@example
awk -f myprog.awk -- -v -q file1 file2 @dots{}
@end example

The following fragment processes @code{ARGV} in order to examine, and
then remove, the previously mentioned command-line options:

@example
BEGIN @{
    for (i = 1; i < ARGC; i++) @{
        if (ARGV[i] == "-v")
            verbose = 1
        else if (ARGV[i] == "-q")
            debug = 1
        else if (ARGV[i] ~ /^-./) @{
            e = sprintf("%s: unrecognized option -- %c",
                    ARGV[0], substr(ARGV[i], 2, 1))
            print e > "/dev/stderr"
        @} else
            break
        delete ARGV[i]
    @}
@}
@end example

@cindex differences in @command{awk} and @command{gawk}, @code{ARGC}/@code{ARGV} variables
Ending the @command{awk} options with @option{--} isn't
necessary in @command{gawk}. Unless @option{--posix} has
been specified, @command{gawk} silently puts any unrecognized options
into @code{ARGV} for the @command{awk} program to deal with.  As soon
as it sees an unknown option, @command{gawk} stops looking for other
options that it might otherwise recognize.  The previous command line with
@command{gawk} would be:

@example
gawk -f myprog.awk -q -v file1 file2 @dots{}
@end example

@noindent
Because @option{-q} is not a valid @command{gawk} option, it and the
following @option{-v} are passed on to the @command{awk} program.
(@xref{Getopt Function} for an @command{awk} library function that
parses command-line options.)

When designing your program, you should choose options that don't
conflict with @command{gawk}'s, because it will process any options
that it accepts before passing the rest of the command line on to
your program.  Using @samp{#!} with the @option{-E} option may help
(@pxref{Executable Scripts}
and
@ifnotdocbook
@pxref{Options}).
@end ifnotdocbook
@ifdocbook
@ref{Options}).
@end ifdocbook

@node Pattern Action Summary
@section Summary

@itemize @value{BULLET}
@item
Pattern--action pairs make up the basic elements of an @command{awk}
program.  Patterns are either normal expressions, range expressions,
or regexp constants; one of the special keywords @code{BEGIN}, @code{END},
@code{BEGINFILE}, or @code{ENDFILE}; or empty.  The action executes if
the current record matches the pattern.  Empty (missing) patterns match
all records.

@item
I/O from @code{BEGIN} and @code{END} rules has certain constraints.
This is also true, only more so, for @code{BEGINFILE} and @code{ENDFILE}
rules.  The latter two give you ``hooks'' into @command{gawk}'s file
processing, allowing you to recover from a file that otherwise would
cause a fatal error (such as a file that cannot be opened).

@item
Shell variables can be used in @command{awk} programs by careful
use of shell quoting.  It is easier to pass a shell variable into
@command{awk} by using the @option{-v} option and an @command{awk}
variable.

@item
Actions consist of statements enclosed in curly braces. Statements
are built up from expressions, control statements, compound statements,
input and output statements, and deletion statements.

@item
The control statements in @command{awk} are @code{if}-@code{else},
@code{while}, @code{for}, and @code{do}-@code{while}.  @command{gawk}
adds the @code{switch} statement.  There are two flavors of @code{for}
statement: one for performing general looping, and the other for iterating
through an array.

@item
@code{break} and @code{continue} let you exit early or start the next
iteration of a loop (or get out of a @code{switch}).

@item
@code{next} and @code{nextfile} let you read the next record and start
over at the top of your program or skip to the next input file and
start over, respectively.

@item
The @code{exit} statement terminates your program. When executed
from an action (or function body), it transfers control to the
@code{END} statements. From an @code{END} statement body, it exits
immediately.  You may pass an optional numeric value to be used
as @command{awk}'s exit status.

@item
Some predefined variables provide control over @command{awk}, mainly for I/O.
Other variables convey information from @command{awk} to your program.

@item
@code{ARGC} and @code{ARGV} make the command-line arguments available
to your program. Manipulating them from a @code{BEGIN} rule lets you
control how @command{awk} will process the provided @value{DF}s.

@end itemize

@node Arrays
@chapter Arrays in @command{awk}
@cindex arrays

An @dfn{array} is a table of values called @dfn{elements}.  The
elements of an array are distinguished by their @dfn{indices}.  Indices
may be either numbers or strings.

This @value{CHAPTER} describes how arrays work in @command{awk},
how to use array elements, how to scan through every element in an array,
and how to remove array elements.
It also describes how @command{awk} simulates multidimensional
arrays, as well as some of the less obvious points about array usage.
The @value{CHAPTER} moves on to discuss @command{gawk}'s facility
for sorting arrays, and ends with a brief description of @command{gawk}'s
ability to support true arrays of arrays.

@menu
* Array Basics::                The basics of arrays.
* Numeric Array Subscripts::    How to use numbers as subscripts in
                                @command{awk}.
* Uninitialized Subscripts::    Using Uninitialized variables as subscripts.
* Delete::                      The @code{delete} statement removes an element
                                from an array.
* Multidimensional::            Emulating multidimensional arrays in
                                @command{awk}.
* Arrays of Arrays::            True multidimensional arrays.
* Arrays Summary::              Summary of arrays.
@end menu

@node Array Basics
@section The Basics of Arrays

This @value{SECTION} presents the basics: working with elements
in arrays one at a time, and traversing all of the elements in
an array.

@menu
* Array Intro::                 Introduction to Arrays
* Reference to Elements::       How to examine one element of an array.
* Assigning Elements::          How to change an element of an array.
* Array Example::               Basic Example of an Array
* Scanning an Array::           A variation of the @code{for} statement. It
                                loops through the indices of an array's
                                existing elements.
* Controlling Scanning::        Controlling the order in which arrays are
                                scanned.
@end menu

@node Array Intro
@subsection Introduction to Arrays

@cindex Wall, Larry
@quotation
@i{Doing linear scans over an associative array is like trying to club someone
to death with a loaded Uzi.}
@author Larry Wall
@end quotation

The @command{awk} language provides one-dimensional arrays
for storing groups of related strings or numbers.
Every @command{awk} array must have a name.  Array names have the same
syntax as variable names; any valid variable name would also be a valid
array name.  But one name cannot be used in both ways (as an array and
as a variable) in the same @command{awk} program.

Arrays in @command{awk} superficially resemble arrays in other programming
languages, but there are fundamental differences.  In @command{awk}, it
isn't necessary to specify the size of an array before starting to use it.
Additionally, any number or string, not just consecutive integers,
may be used as an array index.

In most other languages, arrays must be @dfn{declared} before use,
including a specification of
how many elements or components they contain.  In such languages, the
declaration causes a contiguous block of memory to be allocated for that
many elements.  Usually, an index in the array must be a nonnegative integer.
For example, the index zero specifies the first element in the array, which is
actually stored at the beginning of the block of memory.  Index one
specifies the second element, which is stored in memory right after the
first element, and so on.  It is impossible to add more elements to the
array, because it has room only for as many elements as given in
the declaration.
(Some languages allow arbitrary starting and ending
indices---e.g., @samp{15 .. 27}---but the size of the array is still fixed when
the array is declared.)

@c 1/2015: Do not put the numeric values into @code. Array element
@c values are no different than scalar variable values.
A contiguous array of four elements might look like
@ifnotdocbook
@ref{figure-array-elements},
@end ifnotdocbook
@ifdocbook
@inlineraw{docbook, <xref linkend="figure-array-elements"/>},
@end ifdocbook
conceptually, if the element values are eight, @code{"foo"},
@code{""}, and 30.

@ifnotdocbook
@float Figure,figure-array-elements
@caption{A contiguous array}
@center @image{array-elements, , , A Contiguous Array}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-array-elements" float="0">
<title>A contiguous array</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="array-elements.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

@noindent
Only the values are stored; the indices are implicit from the order of
the values. Here, eight is the value at index zero, because eight appears in the
position with zero elements before it.

@cindex arrays, indexing
@cindex indexing arrays
@cindex associative arrays
@cindex arrays, associative
Arrays in @command{awk} are different---they are @dfn{associative}.  This means
that each array is a collection of pairs---an index and its corresponding
array element value:

@ifnotdocbook
@c extra empty column to indent it right
@multitable @columnfractions .1 .1 .1
@headitem @tab Index @tab Value
@item @tab @code{3} @tab @code{30}
@item @tab @code{1} @tab @code{"foo"}
@item @tab @code{0} @tab @code{8}
@item @tab @code{2} @tab @code{""}
@end multitable
@end ifnotdocbook

@docbook
<informaltable>
<tgroup cols="2">
<colspec colname="1" align="left"/>
<colspec colname="2" align="left"/>
<thead>
<row>
<entry>Index</entry>
<entry>Value</entry>
</row>
</thead>

<tbody>
<row>
<entry><literal>3</literal></entry>
<entry><literal>30</literal></entry>
</row>

<row>
<entry><literal>1</literal></entry>
<entry><literal>"foo"</literal></entry>
</row>

<row>
<entry><literal>0</literal></entry>
<entry><literal>8</literal></entry>
</row>

<row>
<entry><literal>2</literal></entry>
<entry><literal>""</literal></entry>
</row>

</tbody>
</tgroup>
</informaltable>

@end docbook

@noindent
The pairs are shown in jumbled order because their order is
irrelevant.@footnote{The ordering will vary among @command{awk}
implementations, which typically use hash tables to store array elements
and values.}

One advantage of associative arrays is that new pairs can be added
at any time.  For example, suppose a tenth element is added to the array
whose value is @w{@code{"number ten"}}.  The result is:

@ifnotdocbook
@c extra empty column to indent it right
@multitable @columnfractions .1 .1 .2
@headitem @tab Index @tab Value
@item @tab @code{10} @tab @code{"number ten"}
@item @tab @code{3} @tab @code{30}
@item @tab @code{1} @tab @code{"foo"}
@item @tab @code{0} @tab @code{8}
@item @tab @code{2} @tab @code{""}
@end multitable
@end ifnotdocbook

@docbook
<informaltable>
<tgroup cols="2">
<colspec colname="1" align="left"/>
<colspec colname="2" align="left"/>
<thead>
<row>
<entry>Index</entry>
<entry>Value</entry>
</row>
</thead>
<tbody>

<row>
<entry><literal>10</literal></entry>
<entry><literal>"number ten"</literal></entry>
</row>

<row>
<entry><literal>3</literal></entry>
<entry><literal>30</literal></entry>
</row>

<row>
<entry><literal>1</literal></entry>
<entry><literal>"foo"</literal></entry>
</row>

<row>
<entry><literal>0</literal></entry>
<entry><literal>8</literal></entry>
</row>

<row>
<entry><literal>2</literal></entry>
<entry><literal>""</literal></entry>
</row>

</tbody>
</tgroup>
</informaltable>

@end docbook

@noindent
@cindex sparse arrays
@cindex arrays, sparse
Now the array is @dfn{sparse}, which just means some indices are missing.
It has elements 0--3 and 10, but doesn't have elements 4, 5, 6, 7, 8, or 9.

Another consequence of associative arrays is that the indices don't
have to be nonnegative integers.  Any number, or even a string, can be
an index.  For example, the following is an array that translates words from
English to French:

@ifnotdocbook
@multitable @columnfractions .1 .1 .1
@headitem @tab Index @tab Value
@item @tab @code{"dog"} @tab @code{"chien"}
@item @tab @code{"cat"} @tab @code{"chat"}
@item @tab @code{"one"} @tab @code{"un"}
@item @tab @code{1} @tab @code{"un"}
@end multitable
@end ifnotdocbook

@docbook
<informaltable>
<tgroup cols="2">
<colspec colname="1" align="left"/>
<colspec colname="2" align="left"/>
<thead>
<row>
<entry>Index</entry>
<entry>Value</entry>
</row>
</thead>
<tbody>
<row>
<entry><literal>"dog"</literal></entry>
<entry><literal>"chien"</literal></entry>
</row>

<row>
<entry><literal>"cat"</literal></entry>
<entry><literal>"chat"</literal></entry>
</row>

<row>
<entry><literal>"one"</literal></entry>
<entry><literal>"un"</literal></entry>
</row>

<row>
<entry><literal>1</literal></entry>
<entry><literal>"un"</literal></entry>
</row>

</tbody>
</tgroup>
</informaltable>

@end docbook

@noindent
Here we decided to translate the number one in both spelled-out and
numeric form---thus illustrating that a single array can have both
numbers and strings as indices.
(In fact, array subscripts are always strings.
There are some subtleties to how numbers work when used as
array subscripts; this is discussed in more detail in
@ref{Numeric Array Subscripts}.)
Here, the number @code{1} isn't double-quoted, because @command{awk}
automatically converts it to a string.

@cindex @command{gawk}, @code{IGNORECASE} variable in
@cindex case sensitivity, array indices and
@cindex arrays, and @code{IGNORECASE} variable
@cindex @code{IGNORECASE} variable, and array indices
The value of @code{IGNORECASE} has no effect upon array subscripting.
The identical string value used to store an array element must be used
to retrieve it.
When @command{awk} creates an array (e.g., with the @code{split()}
built-in function),
that array's indices are consecutive integers starting at one.
(@xref{String Functions}.)

@command{awk}'s arrays are efficient---the time to access an element
is independent of the number of elements in the array.

@node Reference to Elements
@subsection Referring to an Array Element
@cindex arrays, referencing elements
@cindex array members
@cindex elements of arrays

The principal way to use an array is to refer to one of its elements.
An @dfn{array reference} is an expression as follows:

@example
@var{array}[@var{index-expression}]
@end example

@noindent
Here, @var{array} is the name of an array.  The expression @var{index-expression} is
the index of the desired element of the array.

@c 1/2015: Having the 4.3 in @samp is a little iffy. It's essentially
@c an expression though, so leave be. It's to early in the discussion
@c to mention that it's really a string.
The value of the array reference is the current value of that array
element.  For example, @code{foo[4.3]} is an expression referencing the element
of array @code{foo} at index @samp{4.3}.

@cindex arrays, unassigned elements
@cindex unassigned array elements
@cindex empty array elements
A reference to an array element that has no recorded value yields a value of
@code{""}, the null string.  This includes elements
that have not been assigned any value as well as elements that have been
deleted (@pxref{Delete}).

@cindex non-existent array elements
@cindex arrays, elements that don't exist
@quotation NOTE
A reference to an element that does not exist @emph{automatically} creates
that array element, with the null string as its value.  (In some cases,
this is unfortunate, because it might waste memory inside @command{awk}.)

Novice @command{awk} programmers often make the mistake of checking if
an element exists by checking if the value is empty:

@example
# Check if "foo" exists in a:         @ii{Incorrect!}
if (a["foo"] != "") @dots{}
@end example

@noindent
This is incorrect for two reasons. First, it @emph{creates} @code{a["foo"]}
if it didn't exist before! Second, it is valid (if a bit unusual) to set
an array element equal to the empty string.
@end quotation

@c @cindex arrays, @code{in} operator and
@cindex @code{in} operator, testing if array element exists
To determine whether an element exists in an array at a certain index, use
the following expression:

@example
@var{indx} in @var{array}
@end example

@cindex side effects, array indexing
@noindent
This expression tests whether the particular index @var{indx} exists,
without the side effect of creating that element if it is not present.
The expression has the value one (true) if @code{@var{array}[@var{indx}]}
exists and zero (false) if it does not exist.
(We use @var{indx} here, because @samp{index} is the name of a built-in
function.)
For example, this statement tests whether the array @code{frequencies}
contains the index @samp{2}:

@example
@group
if (2 in frequencies)
    print "Subscript 2 is present."
@end group
@end example

Note that this is @emph{not} a test of whether the array
@code{frequencies} contains an element whose @emph{value} is two.
There is no way to do that except to scan all the elements.  Also, this
@emph{does not} create @code{frequencies[2]}, while the following
(incorrect) alternative does:

@example
@group
if (frequencies[2] != "")
    print "Subscript 2 is present."
@end group
@end example

@node Assigning Elements
@subsection Assigning Array Elements
@cindex arrays, elements, assigning values
@cindex elements in arrays, assigning values

Array elements can be assigned values just like
@command{awk} variables:

@example
@var{array}[@var{index-expression}] = @var{value}
@end example

@noindent
@var{array} is the name of an array.  The expression
@var{index-expression} is the index of the element of the array that is
assigned a value.  The expression @var{value} is the value to
assign to that element of the array.

@node Array Example
@subsection Basic Array Example
@cindex arrays, an example of using

The following program takes a list of lines, each beginning with a line
number, and prints them out in order of line number.  The line numbers
are not in order when they are first read---instead, they
are scrambled.  This program sorts the lines by making an array using
the line numbers as subscripts.  The program then prints out the lines
in sorted order of their numbers.  It is a very simple program and gets
confused upon encountering repeated numbers, gaps, or lines that don't
begin with a number:

@example
@c file eg/misc/arraymax.awk
@{
    if ($1 > max)
        max = $1
    arr[$1] = $0
@}

END @{
    for (x = 1; x <= max; x++)
        print arr[x]
@}
@c endfile
@end example

The first rule keeps track of the largest line number seen so far;
it also stores each line into the array @code{arr}, at an index that
is the line's number.
The second rule runs after all the input has been read, to print out
all the lines.
When this program is run with the following input:

@example
@group
@c file eg/misc/arraymax.data
5  I am the Five man
2  Who are you?  The new number two!
4  . . . And four on the floor
1  Who is number one?
3  I three you.
@c endfile
@end group
@end example

@noindent
Its output is:

@example
@group
1  Who is number one?
2  Who are you?  The new number two!
3  I three you.
4  . . . And four on the floor
5  I am the Five man
@end group
@end example

If a line number is repeated, the last line with a given number overrides
the others.
Gaps in the line numbers can be handled with an easy improvement to the
program's @code{END} rule, as follows:

@example
@group
END @{
    for (x = 1; x <= max; x++)
        if (x in arr)
            print arr[x]
@}
@end group
@end example

@node Scanning an Array
@subsection Scanning All Elements of an Array
@cindex elements in arrays, scanning
@cindex scanning arrays
@cindex arrays, scanning
@cindex loops, @code{for}, array scanning

In programs that use arrays, it is often necessary to use a loop that
executes once for each element of an array.  In other languages, where
arrays are contiguous and indices are limited to nonnegative integers,
this is easy: all the valid indices can be found by counting from
the lowest index up to the highest.  This technique won't do the job
in @command{awk}, because any number or string can be an array index.
So @command{awk} has a special kind of @code{for} statement for scanning
an array:

@example
@group
for (@var{var} in @var{array})
    @var{body}
@end group
@end example

@noindent
@cindex @code{in} operator, use in loops
This loop executes @var{body} once for each index in @var{array} that the
program has previously used, with the variable @var{var} set to that index.

@cindex arrays, @code{for} statement and
@cindex @code{for} statement, looping over arrays
The following program uses this form of the @code{for} statement.  The
first rule scans the input records and notes which words appear (at
least once) in the input, by storing a one into the array @code{used} with
the word as the index.  The second rule scans the elements of @code{used} to
find all the distinct words that appear in the input.  It prints each
word that is more than 10 characters long and also prints the number of
such words.
@xref{String Functions}
for more information on the built-in function @code{length()}.

@example
@group
# Record a 1 for each word that is used at least once
@{
    for (i = 1; i <= NF; i++)
        used[$i] = 1
@}
@end group

@group
# Find number of distinct words more than 10 characters long
END @{
    for (x in used) @{
        if (length(x) > 10) @{
            ++num_long_words
            print x
        @}
    @}
    print num_long_words, "words longer than 10 characters"
@}
@end group
@end example

@noindent
@xref{Word Sorting}
for a more detailed example of this type.

@cindex arrays, elements, order of access by @code{in} operator
@cindex elements in arrays, order of access by @code{in} operator
@cindex @code{in} operator, order of array access
The order in which elements of the array are accessed by this statement
is determined by the internal arrangement of the array elements within
@command{awk} and in standard @command{awk} cannot be controlled
or changed.  This can lead to problems if new elements are added to
@var{array} by statements in the loop body; it is not predictable whether
the @code{for} loop will reach them.  Similarly, changing @var{var} inside
the loop may produce strange results.  It is best to avoid such things.

As a point of information, @command{gawk} sets up the list of elements
to be iterated over before the loop starts, and does not change it.
But not all @command{awk} versions do so. Consider this program, named
@file{loopcheck.awk}:

@example
BEGIN @{
    a["here"] = "here"
    a["is"] = "is"
    a["a"] = "a"
    a["loop"] = "loop"
    for (i in a) @{
        j++
        a[j] = j
        print i
    @}
@}
@end example

Here is what happens when run with @command{gawk} (and @command{mawk}):

@example
$ @kbd{gawk -f loopcheck.awk}
@print{} here
@print{} loop
@print{} a
@print{} is
@end example

Contrast this to BWK @command{awk}:

@example
$ @kbd{nawk -f loopcheck.awk}
@print{} loop
@print{} here
@print{} is
@print{} a
@print{} 1
@end example

@node Controlling Scanning
@subsection Using Predefined Array Scanning Orders with @command{gawk}

This @value{SUBSECTION} describes a feature that is specific to @command{gawk}.

By default, when a @code{for} loop traverses an array, the order
is undefined, meaning that the @command{awk} implementation
determines the order in which the array is traversed.
This order is usually based on the internal implementation of arrays
and will vary from one version of @command{awk} to the next.

@cindex array scanning order, controlling
@cindex controlling array scanning order
Often, though, you may wish to do something simple, such as
``traverse the array by comparing the indices in ascending order,''
or ``traverse the array by comparing the values in descending order.''
@command{gawk} provides two mechanisms that give you this control:

@itemize @value{BULLET}
@item
Set @code{PROCINFO["sorted_in"]} to one of a set of predefined values.
We describe this now.

@item
Set @code{PROCINFO["sorted_in"]} to the name of a user-defined function
to use for comparison of array elements. This advanced feature
is described later in @ref{Array Sorting}.
@end itemize

@cindex @code{PROCINFO}, values of @code{sorted_in}
The following special values for @code{PROCINFO["sorted_in"]} are available:

@table @code
@item "@@unsorted"
Array elements are processed in arbitrary order, which is the default
@command{awk} behavior.

@item "@@ind_str_asc"
Order by indices in ascending order compared as strings; this is the most basic sort.
(Internally, array indices are always strings, so with @samp{a[2*5] = 1}
the index is @code{"10"} rather than numeric 10.)

@item "@@ind_num_asc"
Order by indices in ascending order but force them to be treated as numbers in the process.
Any index with a non-numeric value will end up positioned as if it were zero.

@item "@@val_type_asc"
Order by element values in ascending order (rather than by indices).
Ordering is by the type assigned to the element
(@pxref{Typing and Comparison}).
All numeric values come before all string values,
which in turn come before all subarrays.
(Subarrays have not been described yet;
@pxref{Arrays of Arrays}.)

@item "@@val_str_asc"
Order by element values in ascending order (rather than by indices).  Scalar values are
compared as strings.  Subarrays, if present, come out last.

@item "@@val_num_asc"
Order by element values in ascending order (rather than by indices).  Scalar values are
compared as numbers.  Subarrays, if present, come out last.
When numeric values are equal, the string values are used to provide
an ordering: this guarantees consistent results across different
versions of the C @code{qsort()} function,@footnote{When two elements
compare as equal, the C @code{qsort()} function does not guarantee
that they will maintain their original relative order after sorting.
Using the string value to provide a unique ordering when the numeric
values are equal ensures that @command{gawk} behaves consistently
across different environments.} which @command{gawk} uses internally
to perform the sorting.

@item "@@ind_str_desc"
Like @code{"@@ind_str_asc"}, but the
string indices are ordered from high to low.

@item "@@ind_num_desc"
Like @code{"@@ind_num_asc"}, but the
numeric indices are ordered from high to low.

@item "@@val_type_desc"
Like @code{"@@val_type_asc"}, but the
element values, based on type, are ordered from high to low.
Subarrays, if present, come out first.

@item "@@val_str_desc"
Like @code{"@@val_str_asc"}, but the
element values, treated as strings, are ordered from high to low.
Subarrays, if present, come out first.

@item "@@val_num_desc"
Like @code{"@@val_num_asc"}, but the
element values, treated as numbers, are ordered from high to low.
Subarrays, if present, come out first.
@end table

The array traversal order is determined before the @code{for} loop
starts to run. Changing @code{PROCINFO["sorted_in"]} in the loop body
does not affect the loop.
For example:

@example
$ @kbd{gawk '}
> @kbd{BEGIN @{}
> @kbd{   a[4] = 4}
> @kbd{   a[3] = 3}
> @kbd{   for (i in a)}
> @kbd{       print i, a[i]}
> @kbd{@}'}
@print{} 4 4
@print{} 3 3
$ @kbd{gawk '}
> @kbd{BEGIN @{}
> @kbd{   PROCINFO["sorted_in"] = "@@ind_str_asc"}
> @kbd{   a[4] = 4}
> @kbd{   a[3] = 3}
> @kbd{   for (i in a)}
> @kbd{       print i, a[i]}
> @kbd{@}'}
@print{} 3 3
@print{} 4 4
@end example

When sorting an array by element values, if a value happens to be
a subarray then it is considered to be greater than any string or
numeric value, regardless of what the subarray itself contains,
and all subarrays are treated as being equal to each other.  Their
order relative to each other is determined by their index strings.

Here are some additional things to bear in mind about sorted
array traversal:

@itemize @value{BULLET}
@item
The value of @code{PROCINFO["sorted_in"]} is global. That is, it affects
all array traversal @code{for} loops.  If you need to change it within your
own code, you should see if it's defined and save and restore the value:

@example
@dots{}
if ("sorted_in" in PROCINFO) @{
    save_sorted = PROCINFO["sorted_in"]
    PROCINFO["sorted_in"] = "@@val_str_desc" # or whatever
@}
@dots{}
if (save_sorted)
    PROCINFO["sorted_in"] = save_sorted
@end example

@item
As already mentioned, the default array traversal order is represented by
@code{"@@unsorted"}.  You can also get the default behavior by assigning
the null string to @code{PROCINFO["sorted_in"]} or by just deleting the
@code{"sorted_in"} element from the @code{PROCINFO} array with
the @code{delete} statement.
(The @code{delete} statement hasn't been described yet; @pxref{Delete}.)
@end itemize

In addition, @command{gawk} provides built-in functions for
sorting arrays; see @ref{Array Sorting Functions}.

@node Numeric Array Subscripts
@section Using Numbers to Subscript Arrays

@cindex numbers, as array subscripts
@cindex arrays, numeric subscripts
@cindex subscripts in arrays, numbers as
@cindex @code{CONVFMT} variable, and array subscripts
An important aspect to remember about arrays is that @emph{array subscripts
are always strings}.  When a numeric value is used as a subscript,
it is converted to a string value before being used for subscripting
(@pxref{Conversion}).
This means that the value of the predefined variable @code{CONVFMT} can
affect how your program accesses elements of an array.  For example:

@example
xyz = 12.153
data[xyz] = 1
CONVFMT = "%2.2f"
if (xyz in data)
    printf "%s is in data\n", xyz
else
    printf "%s is not in data\n", xyz
@end example

@noindent
This prints @samp{12.15 is not in data}.  The first statement gives
@code{xyz} a numeric value.  Assigning to
@code{data[xyz]} subscripts @code{data} with the string value @code{"12.153"}
(using the default conversion value of @code{CONVFMT}, @code{"%.6g"}).
Thus, the array element @code{data["12.153"]} is assigned the value one.
The program then changes
the value of @code{CONVFMT}.  The test @samp{(xyz in data)} generates a new
string value from @code{xyz}---this time @code{"12.15"}---because the value of
@code{CONVFMT} only allows two significant digits.  This test fails,
because @code{"12.15"} is different from @code{"12.153"}.

@cindex converting integer array subscripts
@cindex integer array indices
According to the rules for conversions
(@pxref{Conversion}), integer
values always convert to strings as integers, no matter what the
value of @code{CONVFMT} may happen to be.  So the usual case of
the following works:

@example
for (i = 1; i <= maxsub; i++)
    @ii{do something with} array[i]
@end example

The ``integer values always convert to strings as integers'' rule
has an additional consequence for array indexing.
Octal and hexadecimal constants
@ifnotdocbook
(@pxref{Nondecimal-numbers})
@end ifnotdocbook
@ifdocbook
(covered in @ref{Nondecimal-numbers})
@end ifdocbook
are converted internally into numbers, and their original form
is forgotten.  This means, for example, that @code{array[17]},
@code{array[021]}, and @code{array[0x11]} all refer to the same element!

As with many things in @command{awk}, the majority of the time
things work as you would expect them to.  But it is useful to have a precise
knowledge of the actual rules, as they can sometimes have a subtle
effect on your programs.

@node Uninitialized Subscripts
@section Using Uninitialized Variables as Subscripts

@cindex variables, uninitialized@comma{} as array subscripts
@cindex uninitialized variables, as array subscripts
@cindex subscripts in arrays, uninitialized variables as
@cindex arrays, subscripts, uninitialized variables as
Suppose it's necessary to write a program
to print the input data in reverse order.
A reasonable attempt to do so (with some test
data) might look like this:

@example
$ @kbd{echo 'line 1}
> @kbd{line 2}
> @kbd{line 3' | awk '@{ l[lines] = $0; ++lines @}}
> @kbd{END @{}
>     @kbd{for (i = lines - 1; i >= 0; i--)}
>        @kbd{print l[i]}
> @kbd{@}'}
@print{} line 3
@print{} line 2
@end example

Unfortunately, the very first line of input data did not appear in the
output!

Upon first glance, we would think that this program should have worked.
The variable @code{lines}
is uninitialized, and uninitialized variables have the numeric value zero.
So, @command{awk} should have printed the value of @code{l[0]}.

The issue here is that subscripts for @command{awk} arrays are @emph{always}
strings. Uninitialized variables, when used as strings, have the
value @code{""}, not zero.  Thus, @samp{line 1} ends up stored in
@code{l[""]}.
The following version of the program works correctly:

@example
@{ l[lines++] = $0 @}
END @{
    for (i = lines - 1; i >= 0; i--)
       print l[i]
@}
@end example

Here, the @samp{++} forces @code{lines} to be numeric, thus making
the ``old value'' numeric zero. This is then converted to @code{"0"}
as the array subscript.

@cindex null strings, as array subscripts
@cindex dark corner, array subscripts
@cindex lint checking, array subscripts
Even though it is somewhat unusual, the null string
(@code{""}) is a valid array subscript.
@value{DARKCORNER}
@command{gawk} warns about the use of the null string as a subscript
if @option{--lint} is provided
on the command line (@pxref{Options}).

@node Delete
@section The @code{delete} Statement
@cindex @code{delete} statement
@cindex deleting elements in arrays
@cindex arrays, elements, deleting
@cindex elements in arrays, deleting

To remove an individual element of an array, use the @code{delete}
statement:

@example
delete @var{array}[@var{index-expression}]
@end example

Once an array element has been deleted, any value the element once
had is no longer available. It is as if the element had never
been referred to or been given a value.
The following is an example of deleting elements in an array:

@example
for (i in frequencies)
    delete frequencies[i]
@end example

@noindent
This example removes all the elements from the array @code{frequencies}.
Once an element is deleted, a subsequent @code{for} statement to scan the array
does not report that element and using the @code{in} operator to check for
the presence of that element returns zero (i.e., false):

@example
delete foo[4]
if (4 in foo)
    print "This will never be printed"
@end example

@cindex null strings, and deleting array elements
It is important to note that deleting an element is @emph{not} the
same as assigning it a null value (the empty string, @code{""}).
For example:

@example
@group
foo[4] = ""
if (4 in foo)
  print "This is printed, even though foo[4] is empty"
@end group
@end example

@cindex lint checking, array elements
It is not an error to delete an element that does not exist.
However, if @option{--lint} is provided on the command line
(@pxref{Options}),
@command{gawk} issues a warning message when an element that
is not in the array is deleted.

@cindex common extensions, @code{delete} to delete entire arrays
@cindex extensions, common@comma{} @code{delete} to delete entire arrays
@cindex arrays, deleting entire contents
@cindex deleting entire arrays
@cindex @code{delete} @var{array}
@cindex differences in @command{awk} and @command{gawk}, array elements, deleting
All the elements of an array may be deleted with a single statement
by leaving off the subscript in the @code{delete} statement,
as follows:


@example
delete @var{array}
@end example

Using this version of the @code{delete} statement is about three times
more efficient than the equivalent loop that deletes each element one
at a time.

This form of the @code{delete} statement is also supported
by BWK @command{awk} and @command{mawk}, as well as
by a number of other implementations.

@cindex Brian Kernighan's @command{awk}
@quotation NOTE
For many years, using @code{delete} without a subscript was a common
extension.  In September 2012, it was accepted for inclusion into the
POSIX standard.  See @uref{http://austingroupbugs.net/view.php?id=544,
the Austin Group website}.
@end quotation

@cindex portability, deleting array elements
@cindex Brennan, Michael
The following statement provides a portable but nonobvious way to clear
out an array:@footnote{Thanks to Michael Brennan for pointing this out.}

@example
split("", array)
@end example

@cindex @code{split()} function, array elements@comma{} deleting
The @code{split()} function
(@pxref{String Functions})
clears out the target array first. This call asks it to split
apart the null string. Because there is no data to split out, the
function simply clears the array and then returns.

@quotation CAUTION
Deleting all the elements from an array does not change its type; you cannot
clear an array and then use the array's name as a scalar
(i.e., a regular variable). For example, the following does not work:

@example
a[1] = 3
delete a
a = 3
@end example
@end quotation

@node Multidimensional
@section Multidimensional Arrays

@menu
* Multiscanning::               Scanning multidimensional arrays.
@end menu

@cindex subscripts in arrays, multidimensional
@cindex arrays, multidimensional
A @dfn{multidimensional array} is an array in which an element is identified
by a sequence of indices instead of a single index.  For example, a
two-dimensional array requires two indices.  The usual way (in many
languages, including @command{awk}) to refer to an element of a
two-dimensional array named @code{grid} is with
@code{grid[@var{x},@var{y}]}.

@cindex @code{SUBSEP} variable, and multidimensional arrays
Multidimensional arrays are supported in @command{awk} through
concatenation of indices into one string.
@command{awk} converts the indices into strings
(@pxref{Conversion}) and
concatenates them together, with a separator between them.  This creates
a single string that describes the values of the separate indices.  The
combined string is used as a single index into an ordinary,
one-dimensional array.  The separator used is the value of the built-in
variable @code{SUBSEP}.

For example, suppose we evaluate the expression @samp{foo[5,12] = "value"}
when the value of @code{SUBSEP} is @code{"@@"}.  The numbers 5 and 12 are
converted to strings and
concatenated with an @samp{@@} between them, yielding @code{"5@@12"}; thus,
the array element @code{foo["5@@12"]} is set to @code{"value"}.

Once the element's value is stored, @command{awk} has no record of whether
it was stored with a single index or a sequence of indices.  The two
expressions @samp{foo[5,12]} and @w{@samp{foo[5 SUBSEP 12]}} are always
equivalent.

The default value of @code{SUBSEP} is the string @code{"\034"},
which contains a nonprinting character that is unlikely to appear in an
@command{awk} program or in most input data.
The usefulness of choosing an unlikely character comes from the fact
that index values that contain a string matching @code{SUBSEP} can lead to
combined strings that are ambiguous.  Suppose that @code{SUBSEP} is
@code{"@@"}; then @w{@samp{foo["a@@b", "c"]}} and @w{@samp{foo["a",
"b@@c"]}} are indistinguishable because both are actually
stored as @samp{foo["a@@b@@c"]}.

@cindex @code{in} operator, index existence in multidimensional arrays
To test whether a particular index sequence exists in a
multidimensional array, use the same operator (@code{in}) that is
used for single-dimensional arrays.  Write the whole sequence of indices
in parentheses, separated by commas, as the left operand:

@example
if ((@var{subscript1}, @var{subscript2}, @dots{}) in @var{array})
    @dots{}
@end example

Here is an example that treats its input as a two-dimensional array of
fields; it rotates this array 90 degrees clockwise and prints the
result.  It assumes that all lines have the same number of
elements:

@example
@{
     if (max_nf < NF)
          max_nf = NF
     max_nr = NR
     for (x = 1; x <= NF; x++)
          vector[x, NR] = $x
@}

END @{
     for (x = 1; x <= max_nf; x++) @{
          for (y = max_nr; y >= 1; --y)
               printf("%s ", vector[x, y])
          printf("\n")
     @}
@}
@end example

@noindent
When given the input:

@example
@group
1 2 3 4 5 6
2 3 4 5 6 1
3 4 5 6 1 2
4 5 6 1 2 3
@end group
@end example

@noindent
the program produces the following output:

@example
@group
4 3 2 1
5 4 3 2
6 5 4 3
1 6 5 4
2 1 6 5
3 2 1 6
@end group
@end example

@node Multiscanning
@subsection Scanning Multidimensional Arrays

There is no special @code{for} statement for scanning a
``multidimensional'' array. There cannot be one, because, in truth,
@command{awk} does not have
multidimensional arrays or elements---there is only a
multidimensional @emph{way of accessing} an array.

@cindex subscripts in arrays, multidimensional, scanning
@cindex arrays, multidimensional, scanning
@cindex scanning multidimensional arrays
However, if your program has an array that is always accessed as
multidimensional, you can get the effect of scanning it by combining
the scanning @code{for} statement
(@pxref{Scanning an Array}) with the
built-in @code{split()} function
(@pxref{String Functions}).
It works in the following manner:

@example
for (combined in array) @{
    split(combined, separate, SUBSEP)
    @dots{}
@}
@end example

@noindent
This sets the variable @code{combined} to
each concatenated combined index in the array, and splits it
into the individual indices by breaking it apart where the value of
@code{SUBSEP} appears.  The individual indices then become the elements of
the array @code{separate}.

Thus, if a value is previously stored in @code{array[1, "foo"]}, then
an element with index @code{"1\034foo"} exists in @code{array}.  (Recall
that the default value of @code{SUBSEP} is the character with code 034.)
Sooner or later, the @code{for} statement finds that index and does an
iteration with the variable @code{combined} set to @code{"1\034foo"}.
Then the @code{split()} function is called as follows:

@example
split("1\034foo", separate, "\034")
@end example

@noindent
The result is to set @code{separate[1]} to @code{"1"} and
@code{separate[2]} to @code{"foo"}.  Presto! The original sequence of
separate indices is recovered.


@node Arrays of Arrays
@section Arrays of Arrays
@cindex arrays of arrays

@command{gawk} goes beyond standard @command{awk}'s multidimensional
array access and provides true arrays of
arrays. Elements of a subarray are referred to by their own indices
enclosed in square brackets, just like the elements of the main array.
For example, the following creates a two-element subarray at index @code{1}
of the main array @code{a}:

@example
a[1][1] = 1
a[1][2] = 2
@end example

This simulates a true two-dimensional array. Each subarray element can
contain another subarray as a value, which in turn can hold other arrays
as well. In this way, you can create arrays of three or more dimensions.
The indices can be any @command{awk} expressions, including scalars
separated by commas (i.e., a regular @command{awk} simulated
multidimensional subscript). So the following is valid in
@command{gawk}:

@example
a[1][3][1, "name"] = "barney"
@end example

Each subarray and the main array can be of different length. In fact, the
elements of an array or its subarray do not all have to have the same
type. This means that the main array and any of its subarrays can be
nonrectangular, or jagged in structure. You can assign a scalar value to
the index @code{4} of the main array @code{a}, even though @code{a[1]}
is itself an array and not a scalar:

@example
a[4] = "An element in a jagged array"
@end example

The terms @dfn{dimension}, @dfn{row}, and @dfn{column} are
meaningless when applied
to such an array, but we will use ``dimension'' henceforth to imply the
maximum number of indices needed to refer to an existing element. The
type of any element that has already been assigned cannot be changed
by assigning a value of a different type. You have to first delete the
current element, which effectively makes @command{gawk} forget about
the element at that index:

@example
delete a[4]
a[4][5][6][7] = "An element in a four-dimensional array"
@end example

@noindent
This removes the scalar value from index @code{4} and then inserts a
three-level nested subarray
containing a scalar. You can also
delete an entire subarray or subarray of subarrays:

@example
delete a[4][5]
a[4][5] = "An element in subarray a[4]"
@end example

But recall that you can not delete the main array @code{a} and then use it
as a scalar.

The built-in functions that take array arguments can also be used
with subarrays. For example, the following code fragment uses @code{length()}
(@pxref{String Functions})
to determine the number of elements in the main array @code{a} and
its subarrays:

@example
print length(a), length(a[1]), length(a[1][3])
@end example

@noindent
This results in the following output for our main array @code{a}:

@example
2, 3, 1
@end example

@noindent
The @samp{@var{subscript} in @var{array}} expression
(@pxref{Reference to Elements}) works similarly for both
regular @command{awk}-style
arrays and arrays of arrays. For example, the tests @samp{1 in a},
@samp{3 in a[1]}, and @samp{(1, "name") in a[1][3]} all evaluate to
one (true) for our array @code{a}.

The @samp{for (item in array)} statement (@pxref{Scanning an Array})
can be nested to scan all the
elements of an array of arrays if it is rectangular in structure. In order
to print the contents (scalar values) of a two-dimensional array of arrays
(i.e., in which each first-level element is itself an
array, not necessarily of the same length),
you could use the following code:

@example
for (i in array)
    for (j in array[i])
        print array[i][j]
@end example

The @code{isarray()} function (@pxref{Type Functions})
lets you test if an array element is itself an array:

@example
for (i in array) @{
    if (isarray(array[i]) @{
        for (j in array[i]) @{
            print array[i][j]
        @}
    @}
    else
        print array[i]
@}
@end example

If the structure of a jagged array of arrays is known in advance,
you can often devise workarounds using control statements. For example,
the following code prints the elements of our main array @code{a}:

@example
@group
for (i in a) @{
    for (j in a[i]) @{
        if (j == 3) @{
            for (k in a[i][j])
                print a[i][j][k]
@end group
@group
        @} else
            print a[i][j]
    @}
@}
@end group
@end example

@noindent
@xref{Walking Arrays} for a user-defined function that ``walks'' an
arbitrarily dimensioned array of arrays.

Recall that a reference to an uninitialized array element yields a value
of @code{""}, the null string. This has one important implication when you
intend to use a subarray as an argument to a function, as illustrated by
the following example:

@example
$ @kbd{gawk 'BEGIN @{ split("a b c d", b[1]); print b[1][1] @}'}
@error{} gawk: cmd. line:1: fatal: split: second argument is not an array
@end example

The way to work around this is to first force @code{b[1]} to be an array by
creating an arbitrary index:

@example
$ @kbd{gawk 'BEGIN @{ b[1][1] = ""; split("a b c d", b[1]); print b[1][1] @}'}
@print{} a
@end example

@node Arrays Summary
@section Summary

@itemize @value{BULLET}
@item
Standard @command{awk} provides one-dimensional associative arrays
(arrays indexed by string values).  All arrays are associative; numeric
indices are converted automatically to strings.

@item
Array elements are referenced as @code{@var{array}[@var{indx}]}.
Referencing an element creates it if it did not exist previously.

@item
The proper way to see if an array has an element with a given index
is to use the @code{in} operator: @samp{@var{indx} in @var{array}}.

@item
Use @samp{for (@var{indx} in @var{array}) @dots{}} to scan through all the
individual elements of an array. In the body of the loop, @var{indx} takes
on the value of each element's index in turn.

@item
The order in which a @samp{for (@var{indx} in @var{array})} loop
traverses an array is undefined in POSIX @command{awk} and varies among
implementations.  @command{gawk} lets you control the order by assigning
special predefined values to @code{PROCINFO["sorted_in"]}.

@item
Use @samp{delete @var{array}[@var{indx}]} to delete an individual element.
To delete all of the elements in an array,
use @samp{delete @var{array}}.
This latter feature has been a common extension for many
years and is now standard, but may not be supported by all commercial
versions of @command{awk}.

@item
Standard @command{awk} simulates multidimensional arrays by separating
subscript values with commas.  The values are concatenated into a
single string, separated by the value of @code{SUBSEP}.  The fact
that such a subscript was created in this way is not retained; thus,
changing @code{SUBSEP} may have unexpected consequences.  You can use
@samp{(@var{sub1}, @var{sub2}, @dots{}) in @var{array}} to see if such
a multidimensional subscript exists in @var{array}.

@item
@command{gawk} provides true arrays of arrays. You use a separate
set of square brackets for each dimension in such an array:
@code{data[row][col]}, for example. Array elements may thus be either
scalar values (number or string) or other arrays.

@item
Use the @code{isarray()} built-in function to determine if an array
element is itself a subarray.

@end itemize


@node Functions
@chapter Functions

@cindex functions, built-in
@cindex built-in functions
This @value{CHAPTER} describes @command{awk}'s built-in functions,
which fall into three categories: numeric, string, and I/O.
@command{gawk} provides additional groups of functions
to work with values that represent time, do
bit manipulation, sort arrays,
provide type information, and internationalize and localize programs.

Besides the built-in functions, @command{awk} has provisions for
writing new functions that the rest of a program can use.
The second half of this @value{CHAPTER} describes these
@dfn{user-defined} functions.
Finally, we explore indirect function calls, a @command{gawk}-specific
extension that lets you determine at runtime what function is to
be called.

@menu
* Built-in::                    Summarizes the built-in functions.
* User-defined::                Describes User-defined functions in detail.
* Indirect Calls::              Choosing the function to call at runtime.
* Functions Summary::           Summary of functions.
@end menu

@node Built-in
@section Built-in Functions

@dfn{Built-in} functions are always available for your @command{awk}
program to call.  This @value{SECTION} defines all the built-in functions
in @command{awk}; some of these are mentioned in other @value{SECTION}s
but are summarized here for your convenience.

@menu
* Calling Built-in::            How to call built-in functions.
* Numeric Functions::           Functions that work with numbers, including
                                @code{int()}, @code{sin()} and @code{rand()}.
* String Functions::            Functions for string manipulation, such as
                                @code{split()}, @code{match()} and
                                @code{sprintf()}.
* I/O Functions::               Functions for files and shell commands.
* Time Functions::              Functions for dealing with timestamps.
* Bitwise Functions::           Functions for bitwise operations.
* Type Functions::              Functions for type information.
* I18N Functions::              Functions for string translation.
@end menu

@node Calling Built-in
@subsection Calling Built-in Functions

To call one of @command{awk}'s built-in functions, write the name of
the function followed
by arguments in parentheses.  For example, @samp{atan2(y + z, 1)}
is a call to the function @code{atan2()} and has two arguments.

@cindex programming conventions, functions, calling
@cindex whitespace, functions@comma{} calling
Whitespace is ignored between the built-in function name and the
opening parenthesis, but nonetheless it is good practice to avoid using whitespace
there.  User-defined functions do not permit whitespace in this way, and
it is easier to avoid mistakes by following a simple
convention that always works---no whitespace after a function name.

@cindex troubleshooting, @command{gawk}, fatal errors@comma{} function arguments
@cindex @command{gawk}, function arguments and
@cindex differences in @command{awk} and @command{gawk}, function arguments (@command{gawk})
Each built-in function accepts a certain number of arguments.
In some cases, arguments can be omitted. The defaults for omitted
arguments vary from function to function and are described under the
individual functions.  In some @command{awk} implementations, extra
arguments given to built-in functions are ignored.  However, in @command{gawk},
it is a fatal error to give extra arguments to a built-in function.

When a function is called, expressions that create the function's actual
parameters are evaluated completely before the call is performed.
For example, in the following code fragment:

@example
i = 4
j = sqrt(i++)
@end example

@cindex evaluation order, functions
@cindex functions, built-in, evaluation order
@cindex built-in functions, evaluation order
@noindent
the variable @code{i} is incremented to the value five before @code{sqrt()}
is called with a value of four for its actual parameter.
The order of evaluation of the expressions used for the function's
parameters is undefined.  Thus, avoid writing programs that
assume that parameters are evaluated from left to right or from
right to left.  For example:

@example
i = 5
j = atan2(++i, i *= 2)
@end example

If the order of evaluation is left to right, then @code{i} first becomes
six, and then 12, and @code{atan2()} is called with the two arguments six
and 12.  But if the order of evaluation is right to left, @code{i}
first becomes 10, then 11, and @code{atan2()} is called with the
two arguments 11 and 10.

@node Numeric Functions
@subsection Numeric Functions
@cindex numeric functions

The following list describes all of
the built-in functions that work with numbers.
Optional parameters are enclosed in square brackets@w{ ([ ]):}

@c @asis for docbook
@table @asis
@item @code{atan2(@var{y}, @var{x})}
@cindexawkfunc{atan2}
@cindex arctangent
Return the arctangent of @code{@var{y} / @var{x}} in radians.
You can use @samp{pi = atan2(0, -1)} to retrieve the value of
@value{PI}.

@item @code{cos(@var{x})}
@cindexawkfunc{cos}
@cindex cosine
Return the cosine of @var{x}, with @var{x} in radians.

@item @code{exp(@var{x})}
@cindexawkfunc{exp}
@cindex exponent
Return the exponential of @var{x} (@code{e ^ @var{x}}) or report
an error if @var{x} is out of range.  The range of values @var{x} can have
depends on your machine's floating-point representation.

@item @code{int(@var{x})}
@cindexawkfunc{int}
@cindex round to nearest integer
Return the nearest integer to @var{x}, located between @var{x} and zero and
truncated toward zero.
For example, @code{int(3)} is 3, @code{int(3.9)} is 3, @code{int(-3.9)}
is @minus{}3, and @code{int(-3)} is @minus{}3 as well.

@ifset INTDIV
@item @code{intdiv0(@var{numerator}, @var{denominator}, @var{result})}
@cindexawkfunc{intdiv0}
@cindex intdiv0
Perform integer division, similar to the standard C @code{div()} function.
First, truncate @code{numerator} and @code{denominator}
towards zero, creating integer values.  Clear the @code{result}
array, and then set @code{result["quotient"]} to the result of
@samp{numerator / denominator}, truncated towards zero to an integer,
and set @code{result["remainder"]} to the result of @samp{numerator %
denominator}, truncated towards zero to an integer.
Attempting division by zero causes a fatal error.
The function returns zero upon success, and @minus{}1 upon error.

This function is
primarily intended for use with arbitrary length integers; it avoids
creating MPFR arbitrary precision floating-point values (@pxref{Arbitrary
Precision Integers}).

This function is a @code{gawk} extension.  It is not available in
compatibility mode (@pxref{Options}).
@end ifset

@item @code{log(@var{x})}
@cindexawkfunc{log}
@cindex logarithm
Return the natural logarithm of @var{x}, if @var{x} is positive;
otherwise, return @code{NaN} (``not a number'') on IEEE 754 systems.
Additionally, @command{gawk} prints a warning message when @code{x}
is negative.

@item @code{rand()}
@cindexawkfunc{rand}
@cindex random numbers, @code{rand()}/@code{srand()} functions
Return a random number.  The values of @code{rand()} are
uniformly distributed between zero and one.
The value could be zero but is never one.@footnote{The C version of @code{rand()}
on many Unix systems
is known to produce fairly poor sequences of random numbers.
However, nothing requires that an @command{awk} implementation use the C
@code{rand()} to implement the @command{awk} version of @code{rand()}.
In fact, @command{gawk} uses the BSD @code{random()} function, which is
considerably better than @code{rand()}, to produce random numbers.}

Often random integers are needed instead.  Following is a user-defined function
that can be used to obtain a random nonnegative integer less than @var{n}:

@example
function randint(n)
@{
    return int(n * rand())
@}
@end example

@noindent
The multiplication produces a random number greater than or equal to
zero and less than @code{n}.  Using @code{int()}, this result is made into
an integer between zero and @code{n} @minus{} 1, inclusive.

The following example uses a similar function to produce random integers
between one and @var{n}.  This program prints a new random number for
each input record:

@example
# Function to roll a simulated die.
function roll(n) @{ return 1 + int(rand() * n) @}

# Roll 3 six-sided dice and
# print total number of points.
@{
    printf("%d points\n", roll(6) + roll(6) + roll(6))
@}
@end example

@cindex seeding random number generator
@cindex random numbers, seed of
@quotation CAUTION
In most @command{awk} implementations, including @command{gawk},
@code{rand()} starts generating numbers from the same
starting number, or @dfn{seed}, each time you run @command{awk}.@footnote{@command{mawk}
uses a different seed each time.}  Thus,
a program generates the same results each time you run it.
The numbers are random within one @command{awk} run but predictable
from run to run.  This is convenient for debugging, but if you want
a program to do different things each time it is used, you must change
the seed to a value that is different in each run.  To do this,
use @code{srand()}.
@end quotation

@item @code{sin(@var{x})}
@cindexawkfunc{sin}
@cindex sine
Return the sine of @var{x}, with @var{x} in radians.

@item @code{sqrt(@var{x})}
@cindexawkfunc{sqrt}
@cindex square root
Return the positive square root of @var{x}.
@command{gawk} prints a warning message
if @var{x} is negative.  Thus, @code{sqrt(4)} is 2.

@item @code{srand(}[@var{x}]@code{)}
@cindexawkfunc{srand}
Set the starting point, or seed,
for generating random numbers to the value @var{x}.

Each seed value leads to a particular sequence of random
numbers.@footnote{Computer-generated random numbers really are not truly
random.  They are technically known as @dfn{pseudorandom}.  This means
that although the numbers in a sequence appear to be random, you can in
fact generate the same sequence of random numbers over and over again.}
Thus, if the seed is set to the same value a second time,
the same sequence of random numbers is produced again.

@quotation CAUTION
Different @command{awk} implementations use different random-number
generators internally.  Don't expect the same @command{awk} program
to produce the same series of random numbers when executed by
different versions of @command{awk}.
@end quotation

If the argument @var{x} is omitted, as in @samp{srand()}, then the current
date and time of day are used for a seed.  This is the way to get random
numbers that are truly unpredictable.

The return value of @code{srand()} is the previous seed.  This makes it
easy to keep track of the seeds in case you need to consistently reproduce
sequences of random numbers.

POSIX does not specify the initial seed; it differs among @command{awk}
implementations.
@end table

@node String Functions
@subsection String-Manipulation Functions
@cindex string-manipulation functions

The functions in this @value{SECTION} look at or change the text of one
or more strings.

@command{gawk} understands locales (@pxref{Locales}) and does all
string processing in terms of @emph{characters}, not @emph{bytes}.
This distinction is particularly important to understand for locales
where one character may be represented by multiple bytes.  Thus, for
example, @code{length()} returns the number of characters in a string,
and not the number of bytes used to represent those characters. Similarly,
@code{index()} works with character indices, and not byte indices.

@quotation CAUTION
A number of functions deal with indices into strings.  For these
functions, the first character of a string is at position (index) one.
This is different from C and the languages descended from it, where the
first character is at position zero.  You need to remember this when
doing index calculations, particularly if you are used to C.
@end quotation

In the following list, optional parameters are enclosed in square brackets@w{ ([ ]).}
Several functions perform string substitution; the full discussion is
provided in the description of the @code{sub()} function, which comes
toward the end, because the list is presented alphabetically.

Those functions that are specific to @command{gawk} are marked with a
pound sign (@samp{#}).  They are not available in compatibility mode
(@pxref{Options}):


@menu
* Gory Details::                More than you want to know about @samp{\} and
                                @samp{&} with @code{sub()}, @code{gsub()}, and
                                @code{gensub()}.
@end menu

@c @asis for docbook
@table @asis
@item @code{asort(}@var{source} [@code{,} @var{dest} [@code{,} @var{how} ] ]@code{) #}
@itemx @code{asorti(}@var{source} [@code{,} @var{dest} [@code{,} @var{how} ] ]@code{) #}
@cindexgawkfunc{asorti}
@cindex sort array
@cindex arrays, elements, retrieving number of
@cindexgawkfunc{asort}
@cindex sort array indices
These two functions are similar in behavior, so they are described
together.

@quotation NOTE
The following description ignores the third argument, @var{how}, as it
requires understanding features that we have not discussed yet.  Thus,
the discussion here is a deliberate simplification.  (We do provide all
the details later on; see @ref{Array Sorting Functions} for the full story.)
@end quotation

Both functions return the number of elements in the array @var{source}.
For @command{asort()}, @command{gawk} sorts the values of @var{source}
and replaces the indices of the sorted values of @var{source} with
sequential integers starting with one.  If the optional array @var{dest}
is specified, then @var{source} is duplicated into @var{dest}.  @var{dest}
is then sorted, leaving the indices of @var{source} unchanged.

@cindex @command{gawk}, @code{IGNORECASE} variable in
When comparing strings, @code{IGNORECASE} affects the sorting
(@pxref{Array Sorting Functions}).  If the
@var{source} array contains subarrays as values (@pxref{Arrays of
Arrays}), they will come last, after all scalar values.
Subarrays are @emph{not} recursively sorted.

For example, if the contents of @code{a} are as follows:

@example
a["last"] = "de"
a["first"] = "sac"
a["middle"] = "cul"
@end example

@noindent
A call to @code{asort()}:

@example
asort(a)
@end example

@noindent
results in the following contents of @code{a}:

@example
@group
a[1] = "cul"
a[2] = "de"
a[3] = "sac"
@end group
@end example

The @code{asorti()} function works similarly to @code{asort()}; however,
the @emph{indices} are sorted, instead of the values. Thus, in the
previous example, starting with the same initial set of indices and
values in @code{a}, calling @samp{asorti(a)} would yield:

@example
a[1] = "first"
a[2] = "last"
a[3] = "middle"
@end example

@item @code{gensub(@var{regexp}, @var{replacement}, @var{how}} [@code{, @var{target}}]@code{) #}
@cindexgawkfunc{gensub}
@cindex search and replace in strings
@cindex substitute in string
Search the target string @var{target} for matches of the regular
expression @var{regexp}.  If @var{how} is a string beginning with
@samp{g} or @samp{G} (short for ``global''), then replace all matches of @var{regexp} with
@var{replacement}.  Otherwise, @var{how} is treated as a number indicating
which match of @var{regexp} to replace. If no @var{target} is supplied,
use @code{$0}.  It returns the modified string as the result
of the function and the original target string is @emph{not} changed.

@code{gensub()} is a general substitution function.  Its purpose is
to provide more features than the standard @code{sub()} and @code{gsub()}
functions.

@code{gensub()} provides an additional feature that is not available
in @code{sub()} or @code{gsub()}: the ability to specify components of a
regexp in the replacement text.  This is done by using parentheses in
the regexp to mark the components and then specifying @samp{\@var{N}}
in the replacement text, where @var{N} is a digit from 1 to 9.
For example:

@example
$ @kbd{gawk '}
> @kbd{BEGIN @{}
>      @kbd{a = "abc def"}
>      @kbd{b = gensub(/(.+) (.+)/, "\\2 \\1", "g", a)}
>      @kbd{print b}
> @kbd{@}'}
@print{} def abc
@end example

@noindent
As with @code{sub()}, you must type two backslashes in order
to get one into the string.
In the replacement text, the sequence @samp{\0} represents the entire
matched text, as does the character @samp{&}.

The following example shows how you can use the third argument to control
which match of the regexp should be changed:

@example
$ @kbd{echo a b c a b c |}
> @kbd{gawk '@{ print gensub(/a/, "AA", 2) @}'}
@print{} a b c AA b c
@end example

In this case, @code{$0} is the default target string.
@code{gensub()} returns the new string as its result, which is
passed directly to @code{print} for printing.

@c @cindex automatic warnings
@c @cindex warnings, automatic
If the @var{how} argument is a string that does not begin with @samp{g} or
@samp{G}, or if it is a number that is less than or equal to zero, only one
substitution is performed.  If @var{how} is zero, @command{gawk} issues
a warning message.

If @var{regexp} does not match @var{target}, @code{gensub()}'s return value
is the original unchanged value of @var{target}.

@item @code{gsub(@var{regexp}, @var{replacement}} [@code{, @var{target}}]@code{)}
@cindexawkfunc{gsub}
Search @var{target} for
@emph{all} of the longest, leftmost, @emph{nonoverlapping} matching
substrings it can find and replace them with @var{replacement}.
The @samp{g} in @code{gsub()} stands for
``global,'' which means replace everywhere.  For example:

@example
@{ gsub(/Britain/, "United Kingdom"); print @}
@end example

@noindent
replaces all occurrences of the string @samp{Britain} with @samp{United
Kingdom} for all input records.

The @code{gsub()} function returns the number of substitutions made.  If
the variable to search and alter (@var{target}) is
omitted, then the entire input record (@code{$0}) is used.
As in @code{sub()}, the characters @samp{&} and @samp{\} are special,
and the third argument must be assignable.

@item @code{index(@var{in}, @var{find})}
@cindexawkfunc{index}
@cindex search in string
@cindex find substring in string
Search the string @var{in} for the first occurrence of the string
@var{find}, and return the position in characters where that occurrence
begins in the string @var{in}.  Consider the following example:

@example
$ @kbd{awk 'BEGIN @{ print index("peanut", "an") @}'}
@print{} 3
@end example

@noindent
If @var{find} is not found, @code{index()} returns zero.

@cindex dark corner, regexp as second argument to @code{index()}
With BWK @command{awk} and @command{gawk},
it is a fatal error to use a regexp constant for @var{find}.
Other implementations allow it, simply treating the regexp
constant as an expression meaning @samp{$0 ~ /regexp/}. @value{DARKCORNER}

@item @code{length(}[@var{string}]@code{)}
@cindexawkfunc{length}
@cindex string length
@cindex length of string
Return the number of characters in @var{string}.  If
@var{string} is a number, the length of the digit string representing
that number is returned.  For example, @code{length("abcde")} is five.  By
contrast, @code{length(15 * 35)} works out to three. In this example,
@iftex
@math{15 @cdot 35 = 525},
@end iftex
@ifnottex
@ifnotdocbook
15 * 35 = 525,
@end ifnotdocbook
@end ifnottex
@docbook
15 &sdot; 35 = 525,
@end docbook
and 525 is then converted to the string @code{"525"}, which has
three characters.

@cindex length of input record
@cindex input record, length of
If no argument is supplied, @code{length()} returns the length of @code{$0}.

@c @cindex historical features
@cindex portability, @code{length()} function
@cindex POSIX @command{awk}, functions and, @code{length()}
@quotation NOTE
In older versions of @command{awk}, the @code{length()} function could
be called
without any parentheses.  Doing so is considered poor practice,
although the 2008 POSIX standard explicitly allows it, to
support historical practice.  For programs to be maximally portable,
always supply the parentheses.
@end quotation

@cindex dark corner, @code{length()} function
If @code{length()} is called with a variable that has not been used,
@command{gawk} forces the variable to be a scalar.  Other
implementations of @command{awk} leave the variable without a type.
@value{DARKCORNER}
Consider:

@example
$ @kbd{gawk 'BEGIN @{ print length(x) ; x[1] = 1 @}'}
@print{} 0
@error{} gawk: fatal: attempt to use scalar `x' as array

$ @kbd{nawk 'BEGIN @{ print length(x) ; x[1] = 1 @}'}
@print{} 0
@end example

@noindent
If @option{--lint} has
been specified on the command line, @command{gawk} issues a
warning about this.

@cindex common extensions, @code{length()} applied to an array
@cindex extensions, common@comma{} @code{length()} applied to an array
@cindex differences between @command{gawk} and @command{awk}
@cindex number of array elements
@cindex array, number of elements
With @command{gawk} and several other @command{awk} implementations, when given an
array argument, the @code{length()} function returns the number of elements
in the array. @value{COMMONEXT}
This is less useful than it might seem at first, as the
array is not guaranteed to be indexed from one to the number of elements
in it.
If @option{--lint} is provided on the command line
(@pxref{Options}),
@command{gawk} warns that passing an array argument is not portable.
If @option{--posix} is supplied, using an array argument is a fatal error
(@pxref{Arrays}).

@item @code{match(@var{string}, @var{regexp}} [@code{, @var{array}}]@code{)}
@cindexawkfunc{match}
@cindex string, regular expression match
@cindex match regexp in string
Search @var{string} for the
longest, leftmost substring matched by the regular expression
@var{regexp} and return the character position (index)
at which that substring begins (one, if it starts at the beginning of
@var{string}).  If no match is found, return zero.

The @var{regexp} argument may be either a regexp constant
(@code{/}@dots{}@code{/}) or a string constant (@code{"}@dots{}@code{"}).
In the latter case, the string is treated as a regexp to be matched.
@xref{Computed Regexps} for a
discussion of the difference between the two forms, and the
implications for writing your program correctly.

The order of the first two arguments is the opposite of most other string
functions that work with regular expressions, such as
@code{sub()} and @code{gsub()}.  It might help to remember that
for @code{match()}, the order is the same as for the @samp{~} operator:
@samp{@var{string} ~ @var{regexp}}.

@cindex @code{RSTART} variable, @code{match()} function and
@cindex @code{RLENGTH} variable, @code{match()} function and
@cindex @code{match()} function, @code{RSTART}/@code{RLENGTH} variables
The @code{match()} function sets the predefined variable @code{RSTART} to
the index.  It also sets the predefined variable @code{RLENGTH} to the
length in characters of the matched substring.  If no match is found,
@code{RSTART} is set to zero, and @code{RLENGTH} to @minus{}1.

For example:

@example
@c file eg/misc/findpat.awk
@{
    if ($1 == "FIND")
        regex = $2
    else @{
        where = match($0, regex)
        if (where != 0)
            print "Match of", regex, "found at", where, "in", $0
       @}
@}
@c endfile
@end example

@noindent
This program looks for lines that match the regular expression stored in
the variable @code{regex}.  This regular expression can be changed.  If the
first word on a line is @samp{FIND}, @code{regex} is changed to be the
second word on that line.  Therefore, if given:

@example
@c file eg/misc/findpat.data
FIND ru+n
My program runs
but not very quickly
FIND Melvin
JF+KM
This line is property of Reality Engineering Co.
Melvin was here.
@c endfile
@end example

@noindent
@command{awk} prints:

@example
Match of ru+n found at 12 in My program runs
Match of Melvin found at 1 in Melvin was here.
@end example

@cindex differences in @command{awk} and @command{gawk}, @code{match()} function
If @var{array} is present, it is cleared, and then the zeroth element
of @var{array} is set to the entire portion of @var{string}
matched by @var{regexp}.  If @var{regexp} contains parentheses,
the integer-indexed elements of @var{array} are set to contain the
portion of @var{string} matching the corresponding parenthesized
subexpression.
For example:

@example
$ @kbd{echo foooobazbarrrrr |}
> @kbd{gawk '@{ match($0, /(fo+).+(bar*)/, arr)}
>         @kbd{print arr[1], arr[2] @}'}
@print{} foooo barrrrr
@end example

In addition,
multidimensional subscripts are available providing
the start index and length of each matched subexpression:

@example
$ @kbd{echo foooobazbarrrrr |}
> @kbd{gawk '@{ match($0, /(fo+).+(bar*)/, arr)}
>           @kbd{print arr[1], arr[2]}
>           @kbd{print arr[1, "start"], arr[1, "length"]}
>           @kbd{print arr[2, "start"], arr[2, "length"]}
> @kbd{@}'}
@print{} foooo barrrrr
@print{} 1 5
@print{} 9 7
@end example

There may not be subscripts for the start and index for every parenthesized
subexpression, because they may not all have matched text; thus, they
should be tested for with the @code{in} operator
(@pxref{Reference to Elements}).

@cindex troubleshooting, @code{match()} function
The @var{array} argument to @code{match()} is a
@command{gawk} extension.  In compatibility mode
(@pxref{Options}),
using a third argument is a fatal error.

@item @code{patsplit(@var{string}, @var{array}} [@code{, @var{fieldpat}} [@code{, @var{seps}} ] ]@code{) #}
@cindexgawkfunc{patsplit}
@cindex split string into array
Divide
@var{string} into pieces (or ``fields'') defined by @var{fieldpat}
and store the pieces in @var{array} and the separator strings in the
@var{seps} array.  The first piece is stored in
@code{@var{array}[1]}, the second piece in @code{@var{array}[2]}, and so
forth.  The third argument, @var{fieldpat}, is
a regexp describing the fields in @var{string} (just as @code{FPAT} is
a regexp describing the fields in input records).
It may be either a regexp constant or a string.
If @var{fieldpat} is omitted, the value of @code{FPAT} is used.
@code{patsplit()} returns the number of elements created.
@code{@var{seps}[@var{i}]} is
the possibly null separator string
after @code{@var{array}[@var{i}]}.
The possibly null leading separator will be in @code{@var{seps}[0]}.
So a non-null @var{string} with @var{n} fields will have @var{n+1} separators.
A null @var{string} will not have neither fields nor separators.

The @code{patsplit()} function splits strings into pieces in a
manner similar to the way input lines are split into fields using @code{FPAT}
(@pxref{Splitting By Content}).

Before splitting the string, @code{patsplit()} deletes any previously existing
elements in the arrays @var{array} and @var{seps}.

@item @code{split(@var{string}, @var{array}} [@code{, @var{fieldsep}} [@code{, @var{seps}} ] ]@code{)}
@cindexawkfunc{split}
Divide @var{string} into pieces separated by @var{fieldsep}
and store the pieces in @var{array} and the separator strings in the
@var{seps} array.  The first piece is stored in
@code{@var{array}[1]}, the second piece in @code{@var{array}[2]}, and so
forth.  The string value of the third argument, @var{fieldsep}, is
a regexp describing where to split @var{string} (much as @code{FS} can
be a regexp describing where to split input records).
If @var{fieldsep} is omitted, the value of @code{FS} is used.
@code{split()} returns the number of elements created.
@var{seps} is a @command{gawk} extension, with @code{@var{seps}[@var{i}]}
being the separator string
between @code{@var{array}[@var{i}]} and @code{@var{array}[@var{i}+1]}.
If @var{fieldsep} is a single
space, then any leading whitespace goes into @code{@var{seps}[0]} and
any trailing
whitespace goes into @code{@var{seps}[@var{n}]}, where @var{n} is the
return value of
@code{split()} (i.e., the number of elements in @var{array}).

The @code{split()} function splits strings into pieces in a
manner similar to the way input lines are split into fields.  For example:

@example
split("cul-de-sac", a, "-", seps)
@end example

@noindent
@cindex strings splitting, example
splits the string @code{"cul-de-sac"} into three fields using @samp{-} as the
separator.  It sets the contents of the array @code{a} as follows:

@example
a[1] = "cul"
a[2] = "de"
a[3] = "sac"
@end example

and sets the contents of the array @code{seps} as follows:

@example
seps[1] = "-"
seps[2] = "-"
@end example

@noindent
The value returned by this call to @code{split()} is three.

@cindex differences in @command{awk} and @command{gawk}, @code{split()} function
As with input field-splitting, when the value of @var{fieldsep} is
@w{@code{" "}}, leading and trailing whitespace is ignored in values assigned to
the elements of
@var{array} but not in @var{seps}, and the elements
are separated by runs of whitespace.
Also, as with input field splitting, if @var{fieldsep} is the null string, each
individual character in the string is split into its own array element.
@value{COMMONEXT}

Note, however, that @code{RS} has no effect on the way @code{split()}
works. Even though @samp{RS = ""} causes the newline character to also be an input
field separator, this does not affect how @code{split()} splits strings.

@cindex dark corner, @code{split()} function
Modern implementations of @command{awk}, including @command{gawk}, allow
the third argument to be a regexp constant (@w{@code{/}@dots{}@code{/}})
as well as a string.  @value{DARKCORNER}
The POSIX standard allows this as well.
@xref{Computed Regexps} for a
discussion of the difference between using a string constant or a regexp constant,
and the implications for writing your program correctly.

Before splitting the string, @code{split()} deletes any previously existing
elements in the arrays @var{array} and @var{seps}.

If @var{string} is null, the array has no elements. (So this is a portable
way to delete an entire array with one statement.
@xref{Delete}.)

If @var{string} does not match @var{fieldsep} at all (but is not null),
@var{array} has one element only. The value of that element is the original
@var{string}.

In POSIX mode (@pxref{Options}), the fourth argument is not allowed.

@item @code{sprintf(@var{format}, @var{expression1}, @dots{})}
@cindexawkfunc{sprintf}
@cindex formatting strings
Return (without printing) the string that @code{printf} would
have printed out with the same arguments
(@pxref{Printf}).
For example:

@example
pival = sprintf("pi = %.2f (approx.)", 22/7)
@end example

@noindent
assigns the string @w{@samp{pi = 3.14 (approx.)}} to the variable @code{pival}.

@cindexgawkfunc{strtonum}
@cindex convert string to number
@item @code{strtonum(@var{str}) #}
Examine @var{str} and return its numeric value.  If @var{str}
begins with a leading @samp{0}, @code{strtonum()} assumes that @var{str}
is an octal number.  If @var{str} begins with a leading @samp{0x} or
@samp{0X}, @code{strtonum()} assumes that @var{str} is a hexadecimal number.
For example:

@example
$ @kbd{echo 0x11 |}
> @kbd{gawk '@{ printf "%d\n", strtonum($1) @}'}
@print{} 17
@end example

Using the @code{strtonum()} function is @emph{not} the same as adding zero
to a string value; the automatic coercion of strings to numbers
works only for decimal data, not for octal or hexadecimal.@footnote{Unless
you use the @option{--non-decimal-data} option, which isn't recommended.
@xref{Nondecimal Data} for more information.}

Note also that @code{strtonum()} uses the current locale's decimal point
for recognizing numbers (@pxref{Locales}).

@item @code{sub(@var{regexp}, @var{replacement}} [@code{, @var{target}}]@code{)}
@cindexawkfunc{sub}
@cindex replace in string
Search @var{target}, which is treated as a string, for the
leftmost, longest substring matched by the regular expression @var{regexp}.
Modify the entire string
by replacing the matched text with @var{replacement}.
The modified string becomes the new value of @var{target}.
Return the number of substitutions made (zero or one).

The @var{regexp} argument may be either a regexp constant
(@code{/}@dots{}@code{/}) or a string constant (@code{"}@dots{}@code{"}).
In the latter case, the string is treated as a regexp to be matched.
@xref{Computed Regexps} for a
discussion of the difference between the two forms, and the
implications for writing your program correctly.

This function is peculiar because @var{target} is not simply
used to compute a value, and not just any expression will do---it
must be a variable, field, or array element so that @code{sub()} can
store a modified value there.  If this argument is omitted, then the
default is to use and alter @code{$0}.@footnote{Note that this means
that the record will first be regenerated using the value of @code{OFS} if
any fields have been changed, and that the fields will be updated
after the substitution, even if the operation is a ``no-op'' such
as @samp{sub(/^/, "")}.}
For example:

@example
str = "water, water, everywhere"
sub(/at/, "ith", str)
@end example

@noindent
sets @code{str} to @w{@samp{wither, water, everywhere}}, by replacing the
leftmost longest occurrence of @samp{at} with @samp{ith}.

If the special character @samp{&} appears in @var{replacement}, it
stands for the precise substring that was matched by @var{regexp}.  (If
the regexp can match more than one string, then this precise substring
may vary.)  For example:

@example
@{ sub(/candidate/, "& and his wife"); print @}
@end example

@noindent
changes the first occurrence of @samp{candidate} to @samp{candidate
and his wife} on each input line.
Here is another example:

@example
$ @kbd{awk 'BEGIN @{}
>         @kbd{str = "daabaaa"}
>         @kbd{sub(/a+/, "C&C", str)}
>         @kbd{print str}
> @kbd{@}'}
@print{} dCaaCbaaa
@end example

@noindent
This shows how @samp{&} can represent a nonconstant string and also
illustrates the ``leftmost, longest'' rule in regexp matching
(@pxref{Leftmost Longest}).

The effect of this special character (@samp{&}) can be turned off by putting a
backslash before it in the string.  As usual, to insert one backslash in
the string, you must write two backslashes.  Therefore, write @samp{\\&}
in a string constant to include a literal @samp{&} in the replacement.
For example, the following shows how to replace the first @samp{|} on each line with
an @samp{&}:

@example
@{ sub(/\|/, "\\&"); print @}
@end example

@cindex @code{sub()} function, arguments of
@cindex @code{gsub()} function, arguments of
As mentioned, the third argument to @code{sub()} must
be a variable, field, or array element.
Some versions of @command{awk} allow the third argument to
be an expression that is not an lvalue.  In such a case, @code{sub()}
still searches for the pattern and returns zero or one, but the result of
the substitution (if any) is thrown away because there is no place
to put it.  Such versions of @command{awk} accept expressions
like the following:

@example
sub(/USA/, "United States", "the USA and Canada")
@end example

@noindent
@cindex troubleshooting, @code{gsub()}/@code{sub()} functions
For historical compatibility, @command{gawk} accepts such erroneous code.
However, using any other nonchangeable
object as the third parameter causes a fatal error and your program
will not run.

Finally, if the @var{regexp} is not a regexp constant, it is converted into a
string, and then the value of that string is treated as the regexp to match.

@item @code{substr(@var{string}, @var{start}} [@code{, @var{length}} ]@code{)}
@cindexawkfunc{substr}
@cindex substring
Return a @var{length}-character-long substring of @var{string},
starting at character number @var{start}.  The first character of a
string is character number one.@footnote{This is different from
C and C++, in which the first character is number zero.}
For example, @code{substr("washington", 5, 3)} returns @code{"ing"}.

If @var{length} is not present, @code{substr()} returns the whole suffix of
@var{string} that begins at character number @var{start}.  For example,
@code{substr("washington", 5)} returns @code{"ington"}.  The whole
suffix is also returned
if @var{length} is greater than the number of characters remaining
in the string, counting from character @var{start}.

@cindex Brian Kernighan's @command{awk}
If @var{start} is less than one, @code{substr()} treats it as
if it was one. (POSIX doesn't specify what to do in this case:
BWK @command{awk} acts this way, and therefore @command{gawk}
does too.)
If @var{start} is greater than the number of characters
in the string, @code{substr()} returns the null string.
Similarly, if @var{length} is present but less than or equal to zero,
the null string is returned.

@cindex troubleshooting, @code{substr()} function
The string returned by @code{substr()} @emph{cannot} be
assigned.  Thus, it is a mistake to attempt to change a portion of
a string, as shown in the following example:

@example
string = "abcdef"
# try to get "abCDEf", won't work
substr(string, 3, 3) = "CDE"
@end example

@noindent
It is also a mistake to use @code{substr()} as the third argument
of @code{sub()} or @code{gsub()}:

@example
gsub(/xyz/, "pdq", substr($0, 5, 20))  # WRONG
@end example

@cindex portability, @code{substr()} function
(Some commercial versions of @command{awk} treat
@code{substr()} as assignable, but doing so is not portable.)

If you need to replace bits and pieces of a string, combine @code{substr()}
with string concatenation, in the following manner:

@example
string = "abcdef"
@dots{}
string = substr(string, 1, 2) "CDE" substr(string, 6)
@end example

@cindex case sensitivity, converting case
@cindex strings, converting letter case
@item @code{tolower(@var{string})}
@cindexawkfunc{tolower}
@cindex convert string to lower case
Return a copy of @var{string}, with each uppercase character
in the string replaced with its corresponding lowercase character.
Nonalphabetic characters are left unchanged.  For example,
@code{tolower("MiXeD cAsE 123")} returns @code{"mixed case 123"}.

@item @code{toupper(@var{string})}
@cindexawkfunc{toupper}
@cindex convert string to upper case
Return a copy of @var{string}, with each lowercase character
in the string replaced with its corresponding uppercase character.
Nonalphabetic characters are left unchanged.  For example,
@code{toupper("MiXeD cAsE 123")} returns @code{"MIXED CASE 123"}.
@end table

@sidebar Matching the Null String
@cindex matching, null strings
@cindex null strings, matching
@cindex @code{*} (asterisk), @code{*} operator, null strings@comma{} matching
@cindex asterisk (@code{*}), @code{*} operator, null strings@comma{} matching

In @command{awk}, the @samp{*} operator can match the null string.
This is particularly important for the @code{sub()}, @code{gsub()},
and @code{gensub()} functions.  For example:

@example
$ @kbd{echo abc | awk '@{ gsub(/m*/, "X"); print @}'}
@print{} XaXbXcX
@end example

@noindent
Although this makes a certain amount of sense, it can be surprising.
@end sidebar


@node Gory Details
@subsubsection More about @samp{\} and @samp{&} with @code{sub()}, @code{gsub()}, and @code{gensub()}

@cindex escape processing, @code{gsub()}/@code{gensub()}/@code{sub()} functions
@cindex @code{sub()} function, escape processing
@cindex @code{gsub()} function, escape processing
@cindex @code{gensub()} function (@command{gawk}), escape processing
@cindex @code{\} (backslash), @code{gsub()}/@code{gensub()}/@code{sub()} functions and
@cindex backslash (@code{\}), @code{gsub()}/@code{gensub()}/@code{sub()} functions and
@cindex @code{&} (ampersand), @code{gsub()}/@code{gensub()}/@code{sub()} functions and
@cindex ampersand (@code{&}), @code{gsub()}/@code{gensub()}/@code{sub()} functions and

@quotation CAUTION
This subsubsection has been reported to cause headaches.
You might want to skip it upon first reading.
@end quotation

When using @code{sub()}, @code{gsub()}, or @code{gensub()}, and trying to get literal
backslashes and ampersands into the replacement text, you need to remember
that there are several levels of @dfn{escape processing} going on.

First, there is the @dfn{lexical} level, which is when @command{awk} reads
your program
and builds an internal copy of it to execute.
Then there is the runtime level, which is when @command{awk} actually scans the
replacement string to determine what to generate.

@cindex Brian Kernighan's @command{awk}
At both levels, @command{awk} looks for a defined set of characters that
can come after a backslash.  At the lexical level, it looks for the
escape sequences listed in @ref{Escape Sequences}.
Thus, for every @samp{\} that @command{awk} processes at the runtime
level, you must type two backslashes at the lexical level.
When a character that is not valid for an escape sequence follows the
@samp{\}, BWK @command{awk} and @command{gawk} both simply remove the initial
@samp{\} and put the next character into the string. Thus, for
example, @code{"a\qb"} is treated as @code{"aqb"}.

At the runtime level, the various functions handle sequences of
@samp{\} and @samp{&} differently.  The situation is (sadly) somewhat complex.
Historically, the @code{sub()} and @code{gsub()} functions treated the
two-character sequence @samp{\&} specially; this sequence was replaced in
the generated text with a single @samp{&}.  Any other @samp{\} within
the @var{replacement} string that did not precede an @samp{&} was passed
through unchanged.  This is illustrated in @ref{table-sub-escapes}.

@c Thank to Karl Berry for help with the TeX stuff.
@float Table,table-sub-escapes
@caption{Historical escape sequence processing for @code{sub()} and @code{gsub()}}
@tex
\vbox{\bigskip
% We need more characters for escape and tab ...
\catcode`_ = 0
\catcode`! = 4
% ... since this table has lots of &'s and \'s, so we unspecialize them.
\catcode`\& = \other \catcode`\\ = \other
_halign{_hfil#!_qquad_hfil#!_qquad#_hfil_cr
      You type!@code{sub()} sees!@code{sub()} generates_cr
_hrulefill!_hrulefill!_hrulefill_cr
     @code{\&}!       @code{&}!The matched text_cr
    @code{\\&}!      @code{\&}!A literal @samp{&}_cr
   @code{\\\&}!      @code{\&}!A literal @samp{&}_cr
  @code{\\\\&}!     @code{\\&}!A literal @samp{\&}_cr
 @code{\\\\\&}!     @code{\\&}!A literal @samp{\&}_cr
@code{\\\\\\&}!    @code{\\\&}!A literal @samp{\\&}_cr
    @code{\\q}!      @code{\q}!A literal @samp{\q}_cr
}
_bigskip}
@end tex
@ifdocbook
@multitable @columnfractions .20 .20 .60
@headitem You type @tab @code{sub()} sees @tab @code{sub()} generates
@item @code{\&}      @tab @code{&}    @tab The matched text
@item @code{\\&}     @tab @code{\&}   @tab A literal @samp{&}
@item @code{\\\&}    @tab @code{\&}   @tab A literal @samp{&}
@item @code{\\\\&}   @tab @code{\\&}  @tab A literal @samp{\&}
@item @code{\\\\\&}  @tab @code{\\&}  @tab A literal @samp{\&}
@item @code{\\\\\\&} @tab @code{\\\&} @tab A literal @samp{\\&}
@item @code{\\q}     @tab @code{\q}   @tab A literal @samp{\q}
@end multitable
@end ifdocbook
@ifnottex
@ifnotdocbook
@display
 You type         @code{sub()} sees          @code{sub()} generates
 --------         ----------          ---------------
     @code{\&}              @code{&}            The matched text
    @code{\\&}             @code{\&}            A literal @samp{&}
   @code{\\\&}             @code{\&}            A literal @samp{&}
  @code{\\\\&}            @code{\\&}            A literal @samp{\&}
 @code{\\\\\&}            @code{\\&}            A literal @samp{\&}
@code{\\\\\\&}           @code{\\\&}            A literal @samp{\\&}
    @code{\\q}             @code{\q}            A literal @samp{\q}
@end display
@end ifnotdocbook
@end ifnottex
@end float

@noindent
This table shows the lexical-level processing, where
an odd number of backslashes becomes an even number at the runtime level,
as well as the runtime processing done by @code{sub()}.
(For the sake of simplicity, the rest of the following tables only show the
case of even numbers of backslashes entered at the lexical level.)

The problem with the historical approach is that there is no way to get
a literal @samp{\} followed by the matched text.

Several editions of the POSIX standard attempted to fix this problem
but weren't successful. The details are irrelevant at this point in time.

At one point, the @command{gawk} maintainer submitted
proposed text for a revised standard that
reverts to rules that correspond more closely to the original existing
practice. The proposed rules have special cases that make it possible
to produce a @samp{\} preceding the matched text.
This is shown in
@ref{table-sub-proposed}.

@float Table,table-sub-proposed
@caption{@command{gawk} rules for @code{sub()} and backslash}
@tex
\vbox{\bigskip
% We need more characters for escape and tab ...
\catcode`_ = 0
\catcode`! = 4
% ... since this table has lots of &'s and \'s, so we unspecialize them.
\catcode`\& = \other \catcode`\\ = \other
_halign{_hfil#!_qquad_hfil#!_qquad#_hfil_cr
    You type!@code{sub()} sees!@code{sub()} generates_cr
_hrulefill!_hrulefill!_hrulefill_cr
@code{\\\\\\&}!     @code{\\\&}!A literal @samp{\&}_cr
@code{\\\\&}!     @code{\\&}!A literal @samp{\}, followed by the matched text_cr
  @code{\\&}!      @code{\&}!A literal @samp{&}_cr
  @code{\\q}!      @code{\q}!A literal @samp{\q}_cr
 @code{\\\\}!      @code{\\}!@code{\\}_cr
}
_bigskip}
@end tex
@ifdocbook
@multitable @columnfractions .20 .20 .60
@headitem You type @tab @code{sub()} sees @tab @code{sub()} generates
@item @code{\\\\\\&} @tab @code{\\\&} @tab A literal @samp{\&}
@item @code{\\\\&}   @tab @code{\\&}  @tab A literal @samp{\}, followed by the matched text
@item @code{\\&}     @tab @code{\&}   @tab A literal @samp{&}
@item @code{\\q}     @tab @code{\q}   @tab A literal @samp{\q}
@item @code{\\\\}    @tab @code{\\}   @tab @code{\\}
@end multitable
@end ifdocbook
@ifnottex
@ifnotdocbook
@display
 You type         @code{sub()} sees         @code{sub()} generates
 --------         ----------         ---------------
@code{\\\\\\&}           @code{\\\&}            A literal @samp{\&}
  @code{\\\\&}            @code{\\&}            A literal @samp{\}, followed by the matched text
    @code{\\&}             @code{\&}            A literal @samp{&}
    @code{\\q}             @code{\q}            A literal @samp{\q}
   @code{\\\\}             @code{\\}            @code{\\}
@end display
@end ifnotdocbook
@end ifnottex
@end float

In a nutshell, at the runtime level, there are now three special sequences
of characters (@samp{\\\&}, @samp{\\&}, and @samp{\&}) whereas historically
there was only one.  However, as in the historical case, any @samp{\} that
is not part of one of these three sequences is not special and appears
in the output literally.

@command{gawk} 3.0 and 3.1 follow these rules for @code{sub()} and
@code{gsub()}.  The POSIX standard took much longer to be revised than
was expected.  In addition, the @command{gawk} maintainer's proposal was
lost during the standardization process.  The final rules are
somewhat simpler.  The results are similar except for one case.

@cindex POSIX @command{awk}, functions and, @code{gsub()}/@code{sub()}
The POSIX rules state that @samp{\&} in the replacement string produces
a literal @samp{&}, @samp{\\} produces a literal @samp{\}, and @samp{\} followed
by anything else is not special; the @samp{\} is placed straight into the output.
These rules are presented in @ref{table-posix-sub}.

@float Table,table-posix-sub
@caption{POSIX rules for @code{sub()} and @code{gsub()}}
@tex
\vbox{\bigskip
% We need more characters for escape and tab ...
\catcode`_ = 0
\catcode`! = 4
% ... since this table has lots of &'s and \'s, so we unspecialize them.
\catcode`\& = \other \catcode`\\ = \other
_halign{_hfil#!_qquad_hfil#!_qquad#_hfil_cr
    You type!@code{sub()} sees!@code{sub()} generates_cr
_hrulefill!_hrulefill!_hrulefill_cr
@code{\\\\\\&}!     @code{\\\&}!A literal @samp{\&}_cr
@code{\\\\&}!     @code{\\&}!A literal @samp{\}, followed by the matched text_cr
  @code{\\&}!      @code{\&}!A literal @samp{&}_cr
  @code{\\q}!      @code{\q}!A literal @samp{\q}_cr
 @code{\\\\}!      @code{\\}!@code{\}_cr
}
_bigskip}
@end tex
@ifdocbook
@multitable @columnfractions .20 .20 .60
@headitem You type @tab @code{sub()} sees @tab @code{sub()} generates
@item @code{\\\\\\&} @tab @code{\\\&} @tab A literal @samp{\&}
@item @code{\\\\&}   @tab @code{\\&}  @tab A literal @samp{\}, followed by the matched text
@item @code{\\&}     @tab @code{\&}   @tab A literal @samp{&}
@item @code{\\q}     @tab @code{\q}   @tab A literal @samp{\q}
@item @code{\\\\}    @tab @code{\\}   @tab @code{\}
@end multitable
@end ifdocbook
@ifnottex
@ifnotdocbook
@display
 You type         @code{sub()} sees         @code{sub()} generates
 --------         ----------         ---------------
@code{\\\\\\&}           @code{\\\&}            A literal @samp{\&}
  @code{\\\\&}            @code{\\&}            A literal @samp{\}, followed by the matched text
    @code{\\&}             @code{\&}            A literal @samp{&}
    @code{\\q}             @code{\q}            A literal @samp{\q}
   @code{\\\\}             @code{\\}            @code{\}
@end display
@end ifnotdocbook
@end ifnottex
@end float

The only case where the difference is noticeable is the last one: @samp{\\\\}
is seen as @samp{\\} and produces @samp{\} instead of @samp{\\}.

Starting with @value{PVERSION} 3.1.4, @command{gawk} followed the POSIX rules
when @option{--posix} was specified (@pxref{Options}). Otherwise,
it continued to follow the proposed rules, as
that had been its behavior for many years.

When @value{PVERSION} 4.0.0 was released, the @command{gawk} maintainer
made the POSIX rules the default, breaking well over a decade's worth
of backward compatibility.@footnote{This was rather naive of him, despite
there being a note in this @value{SECTION} indicating that the next major version
would move to the POSIX rules.} Needless to say, this was a bad idea,
and as of @value{PVERSION} 4.0.1, @command{gawk} resumed its historical
behavior, and only follows the POSIX rules when @option{--posix} is given.

The rules for @code{gensub()} are considerably simpler. At the runtime
level, whenever @command{gawk} sees a @samp{\}, if the following character
is a digit, then the text that matched the corresponding parenthesized
subexpression is placed in the generated output.  Otherwise,
no matter what character follows the @samp{\}, it
appears in the generated text and the @samp{\} does not,
as shown in @ref{table-gensub-escapes}.

@float Table,table-gensub-escapes
@caption{Escape sequence processing for @code{gensub()}}
@tex
\vbox{\bigskip
% We need more characters for escape and tab ...
\catcode`_ = 0
\catcode`! = 4
% ... since this table has lots of &'s and \'s, so we unspecialize them.
\catcode`\& = \other \catcode`\\ = \other
_halign{_hfil#!_qquad_hfil#!_qquad#_hfil_cr
    You type!@code{gensub()} sees!@code{gensub()} generates_cr
_hrulefill!_hrulefill!_hrulefill_cr
      @code{&}!           @code{&}!The matched text_cr
    @code{\\&}!          @code{\&}!A literal @samp{&}_cr
   @code{\\\\}!          @code{\\}!A literal @samp{\}_cr
  @code{\\\\&}!         @code{\\&}!A literal @samp{\}, then the matched text_cr
@code{\\\\\\&}!        @code{\\\&}!A literal @samp{\&}_cr
    @code{\\q}!          @code{\q}!A literal @samp{q}_cr
}
_bigskip}
@end tex
@ifdocbook
@multitable @columnfractions .20 .20 .60
@headitem You type @tab @code{gensub()} sees @tab @code{gensub()} generates
@item @code{&}       @tab @code{&}    @tab The matched text
@item @code{\\&}     @tab @code{\&}   @tab A literal @samp{&}
@item @code{\\\\}    @tab @code{\\}   @tab A literal @samp{\}
@item @code{\\\\&}   @tab @code{\\&}  @tab A literal @samp{\}, then the matched text
@item @code{\\\\\\&} @tab @code{\\\&} @tab A literal @samp{\&}
@item @code{\\q}     @tab  @code{\q}  @tab A literal @samp{q}
@end multitable
@end ifdocbook
@ifnottex
@ifnotdocbook
@display
  You type          @code{gensub()} sees         @code{gensub()} generates
  --------          -------------         ------------------
      @code{&}                    @code{&}            The matched text
    @code{\\&}                   @code{\&}            A literal @samp{&}
   @code{\\\\}                   @code{\\}            A literal @samp{\}
  @code{\\\\&}                  @code{\\&}            A literal @samp{\}, then the matched text
@code{\\\\\\&}                 @code{\\\&}            A literal @samp{\&}
    @code{\\q}                   @code{\q}            A literal @samp{q}
@end display
@end ifnotdocbook
@end ifnottex
@end float

Because of the complexity of the lexical- and runtime-level processing
and the special cases for @code{sub()} and @code{gsub()},
we recommend the use of @command{gawk} and @code{gensub()} when you have
to do substitutions.

@node I/O Functions
@subsection Input/Output Functions
@cindex input/output functions

The following functions relate to input/output (I/O).
Optional parameters are enclosed in square brackets ([ ]):

@table @asis
@item @code{close(}@var{filename} [@code{,} @var{how}]@code{)}
@cindexawkfunc{close}
@cindex files, closing
@cindex close file or coprocess
Close the file @var{filename} for input or output. Alternatively, the
argument may be a shell command that was used for creating a coprocess, or
for redirecting to or from a pipe; then the coprocess or pipe is closed.
@xref{Close Files And Pipes}
for more information.

When closing a coprocess, it is occasionally useful to first close
one end of the two-way pipe and then to close the other.  This is done
by providing a second argument to @code{close()}.  This second argument
(@var{how})
should be one of the two string values @code{"to"} or @code{"from"},
indicating which end of the pipe to close.  Case in the string does
not matter.
@xref{Two-way I/O},
which discusses this feature in more detail and gives an example.

Note that the second argument to @code{close()} is a @command{gawk}
extension; it is not available in compatibility mode (@pxref{Options}).

@item @code{fflush(}[@var{filename}]@code{)}
@cindexawkfunc{fflush}
@cindex flush buffered output
Flush any buffered output associated with @var{filename}, which is either a
file opened for writing or a shell command for redirecting output to
a pipe or coprocess.

@cindex buffers, flushing
@cindex output, buffering
Many utility programs @dfn{buffer} their output (i.e., they save information
to write to a disk file or the screen in memory until there is enough
for it to be worthwhile to send the data to the output device).
This is often more efficient than writing
every little bit of information as soon as it is ready.  However, sometimes
it is necessary to force a program to @dfn{flush} its buffers (i.e.,
write the information to its destination, even if a buffer is not full).
This is the purpose of the @code{fflush()} function---@command{gawk} also
buffers its output, and the @code{fflush()} function forces
@command{gawk} to flush its buffers.

@cindex extensions, common@comma{} @code{fflush()} function
@cindex Brian Kernighan's @command{awk}
Brian Kernighan added @code{fflush()} to his @command{awk} in April
1992.  For two decades, it was a common extension.  In December
2012, it was accepted for inclusion into the POSIX standard.
See @uref{http://austingroupbugs.net/view.php?id=634, the Austin Group website}.

POSIX standardizes @code{fflush()} as follows: if there
is no argument, or if the argument is the null string (@w{@code{""}}),
then @command{awk} flushes the buffers for @emph{all} open output files
and pipes.

@quotation NOTE
Prior to @value{PVERSION} 4.0.2, @command{gawk}
would flush only the standard output if there was no argument,
and flush all output files and pipes if the argument was the null
string. This was changed in order to be compatible with Brian
Kernighan's @command{awk}, in the hope that standardizing this
feature in POSIX would then be easier (which indeed proved to be the case).

With @command{gawk},
you can use @samp{fflush("/dev/stdout")} if you wish to flush
only the standard output.
@end quotation

@c @cindex automatic warnings
@c @cindex warnings, automatic
@cindex troubleshooting, @code{fflush()} function
@code{fflush()} returns zero if the buffer is successfully flushed;
otherwise, it returns a nonzero value. (@command{gawk} returns @minus{}1.)
In the case where all buffers are flushed, the return value is zero
only if all buffers were flushed successfully.  Otherwise, it is
@minus{}1, and @command{gawk} warns about the problem @var{filename}.

@command{gawk} also issues a warning message if you attempt to flush
a file or pipe that was opened for reading (such as with @code{getline}),
or if @var{filename} is not an open file, pipe, or coprocess.
In such a case, @code{fflush()} returns @minus{}1, as well.

@c end the table to let the sidebar take up the full width of the page.
@end table

@sidebar Interactive Versus Noninteractive Buffering
@cindex buffering, interactive vs.@: noninteractive

As a side point, buffering issues can be even more confusing if
your program is @dfn{interactive} (i.e., communicating
with a user sitting at a keyboard).@footnote{A program is interactive
if the standard output is connected to a terminal device. On modern
systems, this means your keyboard and screen.}

@c Thanks to Walter.Mecky@dresdnerbank.de for this example, and for
@c motivating me to write this section.
Interactive programs generally @dfn{line buffer} their output (i.e., they
write out every line).  Noninteractive programs wait until they have
a full buffer, which may be many lines of output.
Here is an example of the difference:

@example
$ @kbd{awk '@{ print $1 + $2 @}'}
@kbd{1 1}
@print{} 2
@kbd{2 3}
@print{} 5
@kbd{Ctrl-d}
@end example

@noindent
Each line of output is printed immediately. Compare that behavior
with this example:

@example
$ @kbd{awk '@{ print $1 + $2 @}' | cat}
@kbd{1 1}
@kbd{2 3}
@kbd{Ctrl-d}
@print{} 2
@print{} 5
@end example

@noindent
Here, no output is printed until after the @kbd{Ctrl-d} is typed, because
it is all buffered and sent down the pipe to @command{cat} in one shot.
@end sidebar

@table @asis
@item @code{system(@var{command})}
@cindexawkfunc{system}
@cindex invoke shell command
@cindex interacting with other programs
Execute the operating system
command @var{command} and then return to the @command{awk} program.
Return @var{command}'s exit status (see further on).

For example, if the following fragment of code is put in your @command{awk}
program:

@example
END @{
     system("date | mail -s 'awk run done' root")
@}
@end example

@noindent
the system administrator is sent mail when the @command{awk} program
finishes processing input and begins its end-of-input processing.

Note that redirecting @code{print} or @code{printf} into a pipe is often
enough to accomplish your task.  If you need to run many commands, it
is more efficient to simply print them down a pipeline to the shell:

@example
while (@var{more stuff to do})
    print @var{command} | "/bin/sh"
close("/bin/sh")
@end example

@noindent
@cindex troubleshooting, @code{system()} function
@cindex @option{--sandbox} option, disabling @code{system()} function
However, if your @command{awk}
program is interactive, @code{system()} is useful for running large
self-contained programs, such as a shell or an editor.
Some operating systems cannot implement the @code{system()} function.
@code{system()} causes a fatal error if it is not supported.

@quotation NOTE
When @option{--sandbox} is specified, the @code{system()} function is disabled
(@pxref{Options}).
@end quotation

On POSIX systems, a command's exit status is a 16-bit number. The exit
value passed to the C @code{exit()} function is held in the high-order
eight bits. The low-order bits indicate if the process was killed by a
signal (bit 7) and if so, the guilty signal number (bits 0--6).

Traditionally, @command{awk}'s @code{system()} function has simply
returned the exit status value divided by 256. In the normal case this
gives the exit status but in the case of death-by-signal it yields
a fractional floating-point value.@footnote{In private correspondence,
Dr.@: Kernighan has indicated to me that the way this was done
was probably a mistake.} POSIX states that @command{awk}'s
@code{system()} should return the full 16-bit value.

@command{gawk} steers a middle ground.
The return values are summarized in @ref{table-system-return-values}.

@float Table,table-system-return-values
@caption{Return values from @code{system()}}
@multitable @columnfractions .40 .60
@headitem Situation @tab Return value from @code{system()}
@item @option{--traditional} @tab C @code{system()}'s value divided by 256
@item @option{--posix} @tab C @code{system()}'s value
@item Normal exit of command @tab Command's exit status
@item Death by signal of command @tab 256 + number of murderous signal
@item Death by signal of command with core dump @tab 512 + number of murderous signal
@item Some kind of error @tab @minus{}1
@end multitable
@end float
@end table

@sidebar Controlling Output Buffering with @code{system()}
@cindex buffers, flushing
@cindex buffering, input/output
@cindex output, buffering

The @code{fflush()} function provides explicit control over output buffering for
individual files and pipes.  However, its use is not portable to many older
@command{awk} implementations.  An alternative method to flush output
buffers is to call @code{system()} with a null string as its argument:

@example
system("")   # flush output
@end example

@noindent
@command{gawk} treats this use of the @code{system()} function as a special
case and is smart enough not to run a shell (or other command
interpreter) with the empty command.  Therefore, with @command{gawk}, this
idiom is not only useful, it is also efficient.  Although this method should work
with other @command{awk} implementations, it does not necessarily avoid
starting an unnecessary shell.  (Other implementations may only
flush the buffer associated with the standard output and not necessarily
all buffered output.)

If you think about what a programmer expects, it makes sense that
@code{system()} should flush any pending output.  The following program:

@example
BEGIN @{
     print "first print"
     system("echo system echo")
     print "second print"
@}
@end example

@noindent
must print:

@example
first print
system echo
second print
@end example

@noindent
and not:

@example
system echo
first print
second print
@end example

If @command{awk} did not flush its buffers before calling @code{system()},
you would see the latter (undesirable) output.
@end sidebar

@node Time Functions
@subsection Time Functions
@cindex time functions

@cindex timestamps
@cindex log files, timestamps in
@cindex files, log@comma{} timestamps in
@cindex @command{gawk}, timestamps
@cindex POSIX @command{awk}, timestamps and
@command{awk} programs are commonly used to process log files
containing timestamp information, indicating when a
particular log record was written.  Many programs log their timestamps
in the form returned by the @code{time()} system call, which is the
number of seconds since a particular epoch.  On POSIX-compliant systems,
it is the number of seconds since
1970-01-01 00:00:00 UTC, not counting leap
@ifclear FOR_PRINT
seconds.@footnote{@xref{Glossary}, especially the entries ``Epoch'' and ``UTC.''}
@end ifclear
@ifset FOR_PRINT
seconds.
@end ifset
All known POSIX-compliant systems support timestamps from 0 through
@iftex
@math{2^{31} - 1},
@end iftex
@ifinfo
2^31 - 1,
@end ifinfo
@ifnottex
@ifnotinfo
2@sup{31} @minus{} 1,
@end ifnotinfo
@end ifnottex
which is sufficient to represent times through
2038-01-19 03:14:07 UTC.  Many systems support a wider range of timestamps,
including negative timestamps that represent times before the
epoch.

@cindex @command{date} utility, GNU
@cindex time, retrieving
In order to make it easier to process such log files and to produce
useful reports, @command{gawk} provides the following functions for
working with timestamps.  They are @command{gawk} extensions; they are
not specified in the POSIX standard.@footnote{The GNU @command{date} utility can
also do many of the things described here.  Its use may be preferable
for simple time-related operations in shell scripts.}
However, recent versions
of @command{mawk} (@pxref{Other Versions}) also support these functions.
Optional parameters are enclosed in square brackets ([ ]):

@c @asis for docbook
@table @asis
@item @code{mktime(@var{datespec}} [@code{, @var{utc-flag}} ]@code{)}
@cindexgawkfunc{mktime}
@cindex generate time values
Turn @var{datespec} into a timestamp in the same form
as is returned by @code{systime()}.  It is similar to the function of the
same name in ISO C.  The argument, @var{datespec}, is a string of the form
@w{@code{"@var{YYYY} @var{MM} @var{DD} @var{HH} @var{MM} @var{SS} [@var{DST}]"}}.
The string consists of six or seven numbers representing, respectively,
the full year including century, the month from 1 to 12, the day of the month
from 1 to 31, the hour of the day from 0 to 23, the minute from 0 to
59, the second from 0 to 60,@footnote{Occasionally there are
minutes in a year with a leap second, which is why the
seconds can go up to 60.}
and an optional daylight-savings flag.

The values of these numbers need not be within the ranges specified;
for example, an hour of @minus{}1 means 1 hour before midnight.
The origin-zero Gregorian calendar is assumed, with year 0 preceding
year 1 and year @minus{}1 preceding year 0.
If @var{utc-flag} is present and is either nonzero or non-null, the time
is assumed to be in the UTC time zone; otherwise, the
time is assumed to be in the local time zone.
If the @var{DST} daylight-savings flag is positive, the time is assumed to be
daylight savings time; if zero, the time is assumed to be standard
time; and if negative (the default), @code{mktime()} attempts to determine
whether daylight savings time is in effect for the specified time.

If @var{datespec} does not contain enough elements or if the resulting time
is out of range, @code{mktime()} returns @minus{}1.

@cindex @command{gawk}, @code{PROCINFO} array in
@cindex @code{PROCINFO} array
@item @code{strftime(}[@var{format} [@code{,} @var{timestamp} [@code{,} @var{utc-flag}] ] ]@code{)}
@cindexgawkfunc{strftime}
@cindex format time string
Format the time specified by @var{timestamp}
based on the contents of the @var{format} string and return the result.
It is similar to the function of the same name in ISO C.
If @var{utc-flag} is present and is either nonzero or non-null, the value
is formatted as UTC (Coordinated Universal Time, formerly GMT or Greenwich
Mean Time). Otherwise, the value is formatted for the local time zone.
The @var{timestamp} is in the same format as the value returned by the
@code{systime()} function.  If no @var{timestamp} argument is supplied,
@command{gawk} uses the current time of day as the timestamp.
Without a @var{format} argument, @code{strftime()} uses
the value of @code{PROCINFO["strftime"]} as the format string
(@pxref{Built-in Variables}).
The default string value is
@code{@w{"%a %b %e %H:%M:%S %Z %Y"}}.  This format string produces
output that is equivalent to that of the @command{date} utility.
You can assign a new value to @code{PROCINFO["strftime"]} to
change the default format; see the following list for the various format directives.

@item @code{systime()}
@cindexgawkfunc{systime}
@cindex timestamps
@cindex current system time
Return the current time as the number of seconds since
the system epoch.  On POSIX systems, this is the number of seconds
since 1970-01-01 00:00:00 UTC, not counting leap seconds.
It may be a different number on other systems.
@end table

The @code{systime()} function allows you to compare a timestamp from a
log file with the current time of day.  In particular, it is easy to
determine how long ago a particular record was logged.  It also allows
you to produce log records using the ``seconds since the epoch'' format.

@cindex converting, dates to timestamps
@cindex dates, converting to timestamps
@cindex timestamps, converting dates to
The @code{mktime()} function allows you to convert a textual representation
of a date and time into a timestamp.   This makes it easy to do before/after
comparisons of dates and times, particularly when dealing with date and
time data coming from an external source, such as a log file.

The @code{strftime()} function allows you to easily turn a timestamp
into human-readable information.  It is similar in nature to the @code{sprintf()}
function
(@pxref{String Functions}),
in that it copies nonformat specification characters verbatim to the
returned string, while substituting date and time values for format
specifications in the @var{format} string.

@cindex format specifiers, @code{strftime()} function (@command{gawk})
@code{strftime()} is guaranteed by the 1999 ISO C
standard@footnote{Unfortunately,
not every system's @code{strftime()} necessarily
supports all of the conversions listed here.}
to support the following date format specifications:

@table @code
@item %a
The locale's abbreviated weekday name.

@item %A
The locale's full weekday name.

@item %b
The locale's abbreviated month name.

@item %B
The locale's full month name.

@item %c
The locale's ``appropriate'' date and time representation.
(This is @samp{%A %B %d %T %Y} in the @code{"C"} locale.)

@item %C
The century part of the current year.
This is the year divided by 100 and truncated to the next
lower integer.

@item %d
The day of the month as a decimal number (01--31).

@item %D
Equivalent to specifying @samp{%m/%d/%y}.

@item %e
The day of the month, padded with a space if it is only one digit.

@item %F
Equivalent to specifying @samp{%Y-%m-%d}.
This is the ISO 8601 date format.

@item %g
The year modulo 100 of the ISO 8601 week number, as a decimal number (00--99).
For example, January 1, 2012, is in week 53 of 2011. Thus, the year
of its ISO 8601 week number is 2011, even though its year is 2012.
Similarly, December 31, 2012, is in week 1 of 2013. Thus, the year
of its ISO week number is 2013, even though its year is 2012.

@item %G
The full year of the ISO week number, as a decimal number.

@item %h
Equivalent to @samp{%b}.

@item %H
The hour (24-hour clock) as a decimal number (00--23).

@item %I
The hour (12-hour clock) as a decimal number (01--12).

@item %j
The day of the year as a decimal number (001--366).

@item %m
The month as a decimal number (01--12).

@item %M
The minute as a decimal number (00--59).

@item %n
A newline character (ASCII LF).

@item %p
The locale's equivalent of the AM/PM designations associated
with a 12-hour clock.

@item %r
The locale's 12-hour clock time.
(This is @samp{%I:%M:%S %p} in the @code{"C"} locale.)

@item %R
Equivalent to specifying @samp{%H:%M}.

@item %S
The second as a decimal number (00--60).

@item %t
A TAB character.

@item %T
Equivalent to specifying @samp{%H:%M:%S}.

@item %u
The weekday as a decimal number (1--7).  Monday is day one.

@item %U
The week number of the year (with the first Sunday as the first day of week one)
as a decimal number (00--53).

@c @cindex ISO 8601
@item %V
The week number of the year (with the first Monday as the first
day of week one) as a decimal number (01--53).
The method for determining the week number is as specified by ISO 8601.
(To wit: if the week containing January 1 has four or more days in the
new year, then it is week one; otherwise it is the last week
[52 or 53] of the previous year and the next week is week one.)

@item %w
The weekday as a decimal number (0--6).  Sunday is day zero.

@item %W
The week number of the year (with the first Monday as the first day of week one)
as a decimal number (00--53).

@item %x
The locale's ``appropriate'' date representation.
(This is @samp{%A %B %d %Y} in the @code{"C"} locale.)

@item %X
The locale's ``appropriate'' time representation.
(This is @samp{%T} in the @code{"C"} locale.)

@item %y
The year modulo 100 as a decimal number (00--99).

@item %Y
The full year as a decimal number (e.g., 2015).

@c @cindex RFC 822
@c @cindex RFC 1036
@item %z
The time zone offset in a @samp{+@var{HHMM}} format (e.g., the format
necessary to produce RFC 822/RFC 1036 date headers).

@item %Z
The time zone name or abbreviation; no characters if
no time zone is determinable.

@item %Ec %EC %Ex %EX %Ey %EY %Od %Oe %OH
@itemx %OI %Om %OM %OS %Ou %OU %OV %Ow %OW %Oy
``Alternative representations'' for the specifications
that use only the second letter (@samp{%c}, @samp{%C},
and so on).@footnote{If you don't understand any of this, don't worry about
it; these facilities are meant to make it easier to ``internationalize''
programs.
Other internationalization features are described in
@ref{Internationalization}.}
(These facilitate compliance with the POSIX @command{date} utility.)

@item %%
A literal @samp{%}.
@end table

If a conversion specifier is not one of those just listed, the behavior is
undefined.@footnote{This is because ISO C leaves the
behavior of the C version of @code{strftime()} undefined and @command{gawk}
uses the system's version of @code{strftime()} if it's there.
Typically, the conversion specifier either does not appear in the
returned string or appears literally.}

For systems that are not yet fully standards-compliant,
@command{gawk} supplies a copy of
@code{strftime()} from the GNU C Library.
It supports all of the just-listed format specifications.
If that version is
used to compile @command{gawk} (@pxref{Installation}),
then the following additional format specifications are available:

@table @code
@item %k
The hour (24-hour clock) as a decimal number (0--23).
Single-digit numbers are padded with a space.

@item %l
The hour (12-hour clock) as a decimal number (1--12).
Single-digit numbers are padded with a space.

@ignore
@item %N
The ``Emperor/Era'' name.
Equivalent to @samp{%C}.

@item %o
The ``Emperor/Era'' year.
Equivalent to @samp{%y}.
@end ignore

@item %s
The time as a decimal timestamp in seconds since the epoch.

@ignore
@item %v
The date in VMS format (e.g., @samp{20-JUN-1991}).
@end ignore
@end table

Additionally, the alternative representations are recognized but their
normal representations are used.

@cindex @code{date} utility, POSIX
@cindex POSIX @command{awk}, @code{date} utility and
The following example is an @command{awk} implementation of the POSIX
@command{date} utility.  Normally, the @command{date} utility prints the
current date and time of day in a well-known format.  However, if you
provide an argument to it that begins with a @samp{+}, @command{date}
copies nonformat specifier characters to the standard output and
interprets the current time according to the format specifiers in
the string.  For example:

@example
$ @kbd{date '+Today is %A, %B %d, %Y.'}
@print{} Today is Monday, September 22, 2014.
@end example

Here is the @command{gawk} version of the @command{date} utility.
It has a shell ``wrapper'' to handle the @option{-u} option,
which requires that @command{date} run as if the time zone
is set to UTC:

@example
#! /bin/sh
#
# date --- approximate the POSIX 'date' command

case $1 in
-u)  TZ=UTC0     # use UTC
     export TZ
     shift ;;
esac

gawk 'BEGIN  @{
    format = PROCINFO["strftime"]
    exitval = 0

    if (ARGC > 2)
        exitval = 1
    else if (ARGC == 2) @{
        format = ARGV[1]
        if (format ~ /^\+/)
            format = substr(format, 2)   # remove leading +
    @}
    print strftime(format)
    exit exitval
@}' "$@@"
@end example

@node Bitwise Functions
@subsection Bit-Manipulation Functions
@cindex bit-manipulation functions
@cindex bitwise, operations
@cindex AND bitwise operation
@cindex OR bitwise operation
@cindex XOR bitwise operation
@cindex operations, bitwise
@quotation
@i{I can explain it for you, but I can't understand it for you.}
@author Anonymous
@end quotation

Many languages provide the ability to perform @dfn{bitwise} operations
on two integer numbers.  In other words, the operation is performed on
each successive pair of bits in the operands.
Three common operations are bitwise AND, OR, and XOR.
The operations are described in @ref{table-bitwise-ops}.

@c 11/2014: Postprocessing turns the docbook informaltable
@c into a table. Hurray for scripting!
@float Table,table-bitwise-ops
@caption{Bitwise operations}
@ifnottex
@ifnotdocbook
@verbatim
                Bit operator
          |  AND  |   OR  |  XOR
          |---+---+---+---+---+---
Operands  | 0 | 1 | 0 | 1 | 0 | 1
----------+---+---+---+---+---+---
    0     | 0   0 | 0   1 | 0   1
    1     | 0   1 | 1   1 | 1   0
@end verbatim
@end ifnotdocbook
@end ifnottex
@tex
\centerline{
\vbox{\bigskip % space above the table (about 1 linespace)
% Because we have vertical rules, we can't let TeX insert interline space
% in its usual way.
\offinterlineskip
\halign{\strut\hfil#\quad\hfil  % operands
        &\vrule#&\quad#\quad    % rule, 0 (of and)
        &\vrule#&\quad#\quad    % rule, 1 (of and)
        &\vrule#                % rule between and and or
        &\quad#\quad            % 0 (of or)
        &\vrule#&\quad#\quad    % rule, 1 (of of)
        &\vrule#                % rule between or and xor
        &\quad#\quad            % 0 of xor
        &\vrule#&\quad#\quad    % rule, 1 of xor
        \cr
&\omit&\multispan{11}\hfil\bf Bit operator\hfil\cr
\noalign{\smallskip}
&     &\multispan3\hfil AND\hfil&&\multispan3\hfil  OR\hfil
                           &&\multispan3\hfil XOR\hfil\cr
\bf Operands&&0&&1&&0&&1&&0&&1\cr
\noalign{\hrule}
\omit&height 2pt&&\omit&&&&\omit&&&&\omit\cr
\noalign{\hrule height0pt}% without this the rule does not extend; why?
0&&0&\omit&0&&0&\omit&1&&0&\omit&1\cr
1&&0&\omit&1&&1&\omit&1&&1&\omit&0\cr
}}}
@end tex

@docbook
<informaltable>

<tgroup cols="7" colsep="1">
<colspec colname="c1"/>
<colspec colname="c2"/>
<colspec colname="c3"/>
<colspec colname="c4"/>
<colspec colname="c5"/>
<colspec colname="c6"/>
<colspec colname="c7"/>
<spanspec spanname="optitle" namest="c2" nameend="c7" align="center"/>
<spanspec spanname="andspan" namest="c2" nameend="c3" align="center"/>
<spanspec spanname="orspan" namest="c4" nameend="c5" align="center"/>
<spanspec spanname="xorspan" namest="c6" nameend="c7" align="center"/>

<tbody>
<row>
<entry colsep="0"></entry>
<entry spanname="optitle"><emphasis role="bold">Bit operator</emphasis></entry>
</row>

<row rowsep="1">
<entry rowsep="0"></entry>
<entry spanname="andspan">AND</entry>
<entry spanname="orspan">OR</entry>
<entry spanname="xorspan">XOR</entry>
</row>

<row rowsep="1">
<entry ><emphasis role="bold">Operands</emphasis></entry>
<entry colsep="0">0</entry>
<entry colsep="1">1</entry>
<entry colsep="0">0</entry>
<entry colsep="1">1</entry>
<entry colsep="0">0</entry>
<entry colsep="1">1</entry>
</row>

<row>
<entry align="center">0</entry>
<entry colsep="0">0</entry>
<entry>0</entry>
<entry colsep="0">0</entry>
<entry>1</entry>
<entry colsep="0">0</entry>
<entry>1</entry>
</row>

<row>
<entry align="center">1</entry>
<entry colsep="0">0</entry>
<entry>1</entry>
<entry colsep="0">1</entry>
<entry>1</entry>
<entry colsep="0">1</entry>
<entry>0</entry>
</row>

</tbody>
</tgroup>
</informaltable>
@end docbook
@end float

@cindex bitwise, complement
@cindex complement, bitwise
As you can see, the result of an AND operation is 1 only when @emph{both}
bits are 1.
The result of an OR operation is 1 if @emph{either} bit is 1.
The result of an XOR operation is 1 if either bit is 1,
but not both.
The next operation is the @dfn{complement}; the complement of 1 is 0 and
the complement of 0 is 1. Thus, this operation ``flips'' all the bits
of a given value.

@cindex bitwise, shift
@cindex left shift, bitwise
@cindex right shift, bitwise
@cindex shift, bitwise
Finally, two other common operations are to shift the bits left or right.
For example, if you have a bit string @samp{10111001} and you shift it
right by three bits, you end up with @samp{00010111}.@footnote{This example
shows that zeros come in on the left side. For @command{gawk}, this is
always true, but in some languages, it's possible to have the left side
fill with ones.}
If you start over again with @samp{10111001} and shift it left by three
bits, you end up with @samp{11001000}.  The following list describes
@command{gawk}'s built-in functions that implement the bitwise operations.
Optional parameters are enclosed in square brackets ([ ]):

@cindex @command{gawk}, bitwise operations in
@table @code
@cindexgawkfunc{and}
@cindex bitwise AND
@item @code{and(}@var{v1}@code{,} @var{v2} [@code{,} @dots{}]@code{)}
Return the bitwise AND of the arguments. There must be at least two.

@cindexgawkfunc{compl}
@cindex bitwise complement
@item @code{compl(@var{val})}
Return the bitwise complement of @var{val}.

@cindexgawkfunc{lshift}
@cindex left shift
@item @code{lshift(@var{val}, @var{count})}
Return the value of @var{val}, shifted left by @var{count} bits.

@cindexgawkfunc{or}
@cindex bitwise OR
@item @code{or(}@var{v1}@code{,} @var{v2} [@code{,} @dots{}]@code{)}
Return the bitwise OR of the arguments. There must be at least two.

@cindexgawkfunc{rshift}
@cindex right shift
@item @code{rshift(@var{val}, @var{count})}
Return the value of @var{val}, shifted right by @var{count} bits.

@cindexgawkfunc{xor}
@cindex bitwise XOR
@item @code{xor(}@var{v1}@code{,} @var{v2} [@code{,} @dots{}]@code{)}
Return the bitwise XOR of the arguments. There must be at least two.
@end table

@quotation CAUTION
Beginning with @command{gawk} @value{PVERSION} 4.2, negative
operands are not allowed for any of these functions. A negative
operand produces a fatal error.  See the sidebar
``Beware The Smoke and Mirrors!'' for more information as to why.
@end quotation

Here is a user-defined function (@pxref{User-defined})
that illustrates the use of these functions:

@cindex @code{bits2str()} user-defined function
@cindex @code{testbits.awk} program
@example
@group
@c file eg/lib/bits2str.awk
# bits2str --- turn an integer into readable ones and zeros

function bits2str(bits,        data, mask)
@{
    if (bits == 0)
        return "0"

    mask = 1
    for (; bits != 0; bits = rshift(bits, 1))
        data = (and(bits, mask) ? "1" : "0") data

    while ((length(data) % 8) != 0)
        data = "0" data

    return data
@}
@c endfile
@end group

@c this is a hack to make testbits.awk self-contained
@ignore
@c file eg/prog/testbits.awk
# bits2str --- turn an integer into readable ones and zeros

function bits2str(bits,        data, mask)
@{
    if (bits == 0)
        return "0"

    mask = 1
    for (; bits != 0; bits = rshift(bits, 1))
        data = (and(bits, mask) ? "1" : "0") data

    while ((length(data) % 8) != 0)
        data = "0" data

    return data
@}
@c endfile
@end ignore
@c file eg/prog/testbits.awk
BEGIN @{
    printf "123 = %s\n", bits2str(123)
    printf "0123 = %s\n", bits2str(0123)
    printf "0x99 = %s\n", bits2str(0x99)
    comp = compl(0x99)
    printf "compl(0x99) = %#x = %s\n", comp, bits2str(comp)
    shift = lshift(0x99, 2)
    printf "lshift(0x99, 2) = %#x = %s\n", shift, bits2str(shift)
    shift = rshift(0x99, 2)
    printf "rshift(0x99, 2) = %#x = %s\n", shift, bits2str(shift)
@}
@c endfile
@end example

@noindent
This program produces the following output when run:

@example
$ @kbd{gawk -f testbits.awk}
@print{} 123 = 01111011
@print{} 0123 = 01010011
@print{} 0x99 = 10011001
@print{} compl(0x99) = 0x3fffffffffff66 =
@print{} 00111111111111111111111111111111111111111111111101100110
@print{} lshift(0x99, 2) = 0x264 = 0000001001100100
@print{} rshift(0x99, 2) = 0x26 = 00100110
@end example

@cindex converting, strings to numbers
@cindex strings, converting
@cindex numbers, converting
@cindex converting, numbers to strings
@cindex number as string of bits
The @code{bits2str()} function turns a binary number into a string.
Initializing @code{mask} to one creates
a binary value where the rightmost bit
is set to one.  Using this mask,
the function repeatedly checks the rightmost bit.
ANDing the mask with the value indicates whether the
rightmost bit is one or not. If so, a @code{"1"} is concatenated onto the front
of the string.
Otherwise, a @code{"0"} is added.
The value is then shifted right by one bit and the loop continues
until there are no more one bits.

If the initial value is zero, it returns a simple @code{"0"}.
Otherwise, at the end, it pads the value with zeros to represent multiples
of 8-bit quantities. This is typical in modern computers.

The main code in the @code{BEGIN} rule shows the difference between the
decimal and octal values for the same numbers
(@pxref{Nondecimal-numbers}),
and then demonstrates the
results of the @code{compl()}, @code{lshift()}, and @code{rshift()} functions.

@sidebar Beware The Smoke and Mirrors!

It other languages, bitwise operations are performed on integer values,
not floating-point values.  As a general statement, such operations work
best when performed on unsigned integers.

@command{gawk} attempts to treat the arguments to the bitwise functions
as unsigned integers.  For this reason, negative arguments produce a
fatal error.

In normal operation, for all of these functions, first the
double-precision floating-point value is converted to the widest C
unsigned integer type, then the bitwise operation is performed.  If the
result cannot be represented exactly as a C @code{double}, leading
nonzero bits are removed one by one until it can be represented exactly.
The result is then converted back into a C @code{double}.@footnote{If you don't
understand this paragraph, the upshot is that @command{gawk} can only
store a particular range of integer values; numbers outside that range
are reduced to fit within the range.}

However, when using arbitrary precision arithmetic with the @option{-M}
option (@pxref{Arbitrary Precision Arithmetic}), the results may differ.
This is particularly noticeable with the @code{compl()} function:

@example
$ @kbd{gawk 'BEGIN @{ print compl(42) @}'}
@print{} 9007199254740949
$ @kbd{gawk -M 'BEGIN @{ print compl(42) @}'}
@print{} -43
@end example

What's going on becomes clear when printing the results
in hexadecimal:

@example
$ @kbd{gawk 'BEGIN @{ printf "%#x\n", compl(42) @}'}
@print{} 0x1fffffffffffd5
$ @kbd{gawk -M 'BEGIN @{ printf "%#x\n", compl(42) @}'}
@print{} 0xffffffffffffffd5
@end example

When using the @option{-M} option, under the hood, @command{gawk} uses
GNU MP arbitrary precision integers which have at least 64 bits of precision.
When not using @option{-M}, @command{gawk} stores integral values in
regular double-precision floating point, which only maintain 53 bits of
precision.  Furthermore, the GNU MP library treats (or at least seems to treat)
the leading bit as a sign bit; thus the result with @option{-M} in this case is
a negative number.

In short, using @command{gawk} for any but the simplest kind of bitwise
operations is probably a bad idea; caveat emptor!

@end sidebar

@node Type Functions
@subsection Getting Type Information

@command{gawk} provides two functions that let you distinguish
the type of a variable.
This is necessary for writing code
that traverses every element of an array of arrays
(@pxref{Arrays of Arrays}), and in other contexts.

@table @code
@cindexgawkfunc{isarray}
@cindex scalar or array
@item isarray(@var{x})
Return a true value if @var{x} is an array. Otherwise, return false.

@cindexgawkfunc{typeof}
@cindex variable type
@cindex type, of variable
@item typeof(@var{x})
Return one of the following strings, depending upon the type of @var{x}:

@c nested table
@table @code
@item "array"
@var{x} is an array.

@item "regexp"
@var{x} is a strongly typed regexp (@pxref{Strong Regexp Constants}).

@item "number"
@var{x} is a number.

@item "string"
@var{x} is a string.

@item "strnum"
@var{x} is a number that started life as user input, such as a field or
the result of calling @code{split()}. (I.e., @var{x} has the strnum
attribute; @pxref{Variable Typing}.)

@item "unassigned"
@var{x} is a scalar variable that has not been assigned a value yet.
For example:

@example
BEGIN @{
    # creates a[1] but it has no assigned value
    a[1]
    print typeof(a[1])  # unassigned
@}
@end example

@item "untyped"
@var{x} has not yet been used yet at all; it can become a scalar or an
array.  The typing could even conceivably differ from run to run of
the same program! For example:

@example
BEGIN @{
    print "initially, typeof(v) = ", typeof(v)

    if ("FOO" in ENVIRON)
        make_scalar(v)
    else
        make_array(v)

    print "typeof(v) =", typeof(v)
@}

function make_scalar(p,    l) @{ l = p @}

function make_array(p) @{ p[1] = 1 @}
@end example

@end table
@end table

@code{isarray()} is meant for use in two circumstances. The first is when
traversing a multidimensional array: you can test if an element is itself
an array or not.  The second is inside the body of a user-defined function
(not discussed yet; @pxref{User-defined}), to test if a parameter is an
array or not.

@quotation NOTE
Using @code{isarray()} at the global level to test
variables makes no sense. Because you are the one writing the program, you
are supposed to know if your variables are arrays or not. And in fact,
due to the way @command{gawk} works, if you pass the name of a variable
that has not been previously used to @code{isarray()}, @command{gawk}
ends up turning it into a scalar.
@end quotation

The @code{typeof()} function is general; it allows you to determine
if a variable or function parameter is a scalar, an array, or a strongly
typed regexp.

@node I18N Functions
@subsection String-Translation Functions
@cindex @command{gawk}, string-translation functions
@cindex functions, string-translation
@cindex string-translation functions
@cindex internationalization
@cindex @command{awk} programs, internationalizing

@command{gawk} provides facilities for internationalizing @command{awk} programs.
These include the functions described in the following list.
The descriptions here are purposely brief.
@xref{Internationalization},
for the full story.
Optional parameters are enclosed in square brackets ([ ]):

@table @asis
@cindexgawkfunc{bindtextdomain}
@cindex set directory of message catalogs
@item @code{bindtextdomain(@var{directory}} [@code{,} @var{domain}]@code{)}
Set the directory in which
@command{gawk} will look for message translation files, in case they
will not or cannot be placed in the ``standard'' locations
(e.g., during testing).
It returns the directory in which @var{domain} is ``bound.''

The default @var{domain} is the value of @code{TEXTDOMAIN}.
If @var{directory} is the null string (@code{""}), then
@code{bindtextdomain()} returns the current binding for the
given @var{domain}.

@cindexgawkfunc{dcgettext}
@cindex translate string
@item @code{dcgettext(@var{string}} [@code{,} @var{domain} [@code{,} @var{category}] ]@code{)}
Return the translation of @var{string} in
text domain @var{domain} for locale category @var{category}.
The default value for @var{domain} is the current value of @code{TEXTDOMAIN}.
The default value for @var{category} is @code{"LC_MESSAGES"}.

@cindexgawkfunc{dcngettext}
@item @code{dcngettext(@var{string1}, @var{string2}, @var{number}} [@code{,} @var{domain} [@code{,} @var{category}] ]@code{)}
Return the plural form used for @var{number} of the
translation of @var{string1} and @var{string2} in text domain
@var{domain} for locale category @var{category}. @var{string1} is the
English singular variant of a message, and @var{string2} is the English plural
variant of the same message.
The default value for @var{domain} is the current value of @code{TEXTDOMAIN}.
The default value for @var{category} is @code{"LC_MESSAGES"}.
@end table

@node User-defined
@section User-Defined Functions

@cindex user-defined functions
@cindex functions, user-defined
Complicated @command{awk} programs can often be simplified by defining
your own functions.  User-defined functions can be called just like
built-in ones (@pxref{Function Calls}), but it is up to you to define
them (i.e., to tell @command{awk} what they should do).

@menu
* Definition Syntax::           How to write definitions and what they mean.
* Function Example::            An example function definition and what it
                                does.
* Function Caveats::            Things to watch out for.
* Return Statement::            Specifying the value a function returns.
* Dynamic Typing::              How variable types can change at runtime.
@end menu

@node Definition Syntax
@subsection Function Definition Syntax

@quotation
@i{It's entirely fair to say that the awk syntax for local
variable definitions is appallingly awful.}
@author Brian Kernighan
@end quotation

@cindex functions, defining
Definitions of functions can appear anywhere between the rules of an
@command{awk} program.  Thus, the general form of an @command{awk} program is
extended to include sequences of rules @emph{and} user-defined function
definitions.
There is no need to put the definition of a function
before all uses of the function.  This is because @command{awk} reads the
entire program before starting to execute any of it.

The definition of a function named @var{name} looks like this:

@display
@group
@code{function} @var{name}@code{(}[@var{parameter-list}]@code{)}
@code{@{}
     @var{body-of-function}
@code{@}}
@end group
@end display

@cindex names, functions
@cindex functions, names of
@cindex namespace issues, functions
@noindent
Here, @var{name} is the name of the function to define.  A valid function
name is like a valid variable name: a sequence of letters, digits, and
underscores that doesn't start with a digit.
Here too, only the 52 upper- and lowercase English letters may
be used in a function name.
Within a single @command{awk} program, any particular name can only be
used as a variable, array, or function.

@var{parameter-list} is an optional list of the function's arguments and local
variable names, separated by commas.  When the function is called,
the argument names are used to hold the argument values given in
the call.

A function cannot have two parameters with the same name, nor may it
have a parameter with the same name as the function itself.

@quotation CAUTION
According to the POSIX standard, function parameters
cannot have the same name as one of the special predefined variables
(@pxref{Built-in Variables}), nor may a function parameter have the
same name as another function.

Not all versions of @command{awk} enforce
these restrictions.
@command{gawk} always enforces the first restriction.
With @option{--posix} (@pxref{Options}),
it also enforces the second restriction.
@end quotation

Local variables act like the empty string if referenced where a string
value is required, and like zero if referenced where a numeric value
is required. This is the same as the behavior of regular variables that have never been
assigned a value.  (There is more to understand about local variables;
@pxref{Dynamic Typing}.)

The @var{body-of-function} consists of @command{awk} statements.  It is the
most important part of the definition, because it says what the function
should actually @emph{do}.  The argument names exist to give the body a
way to talk about the arguments; local variables exist to give the body
places to keep temporary values.

Argument names are not distinguished syntactically from local variable
names. Instead, the number of arguments supplied when the function is
called determines how many argument variables there are.  Thus, if three
argument values are given, the first three names in @var{parameter-list}
are arguments and the rest are local variables.

It follows that if the number of arguments is not the same in all calls
to the function, some of the names in @var{parameter-list} may be
arguments on some occasions and local variables on others.  Another
way to think of this is that omitted arguments default to the
null string.

@cindex programming conventions, functions, writing
Usually when you write a function, you know how many names you intend to
use for arguments and how many you intend to use as local variables.  It is
conventional to place some extra space between the arguments and
the local variables, in order to document how your function is supposed to be used.

@cindex variables, shadowing
@cindex shadowing of variable values
During execution of the function body, the arguments and local variable
values hide, or @dfn{shadow}, any variables of the same names used in the
rest of the program.  The shadowed variables are not accessible in the
function definition, because there is no way to name them while their
names have been taken away for the arguments and local variables.  All other variables
used in the @command{awk} program can be referenced or set normally in the
function's body.

The arguments and local variables last only as long as the function body
is executing.  Once the body finishes, you can once again access the
variables that were shadowed while the function was running.

@cindex recursive functions
@cindex functions, recursive
The function body can contain expressions that call functions.  They
can even call this function, either directly or by way of another
function.  When this happens, we say the function is @dfn{recursive}.
The act of a function calling itself is called @dfn{recursion}.

All the built-in functions return a value to their caller.
User-defined functions can do so also, using the @code{return} statement,
which is described in detail in @ref{Return Statement}.
Many of the subsequent examples in this @value{SECTION} use
the @code{return} statement.

@cindex common extensions, @code{func} keyword
@cindex extensions, common@comma{} @code{func} keyword
@c @cindex @command{awk} language, POSIX version
@c @cindex POSIX @command{awk}
@cindex POSIX @command{awk}, @code{function} keyword in
In many @command{awk} implementations, including @command{gawk},
the keyword @code{function} may be
abbreviated @code{func}. @value{COMMONEXT}
However, POSIX only specifies the use of
the keyword @code{function}.  This actually has some practical implications.
If @command{gawk} is in POSIX-compatibility mode
(@pxref{Options}), then the following
statement does @emph{not} define a function:

@example
func foo() @{ a = sqrt($1) ; print a @}
@end example

@noindent
Instead, it defines a rule that, for each record, concatenates the value
of the variable @samp{func} with the return value of the function @samp{foo}.
If the resulting string is non-null, the action is executed.
This is probably not what is desired.  (@command{awk} accepts this input as
syntactically valid, because functions may be used before they are defined
in @command{awk} programs.@footnote{This program won't actually run,
because @code{foo()} is undefined.})

@cindex portability, functions@comma{} defining
To ensure that your @command{awk} programs are portable, always use the
keyword @code{function} when defining a function.

@node Function Example
@subsection Function Definition Examples
@cindex function definition example

Here is an example of a user-defined function, called @code{myprint()}, that
takes a number and prints it in a specific format:

@example
function myprint(num)
@{
     printf "%6.3g\n", num
@}
@end example

@noindent
To illustrate, here is an @command{awk} rule that uses our @code{myprint()}
function:

@example
$3 > 0     @{ myprint($3) @}
@end example

@noindent
This program prints, in our special format, all the third fields that
contain a positive number in our input.  Therefore, when given the following input:

@example
 1.2   3.4    5.6   7.8
 9.10 11.12 -13.14 15.16
17.18 19.20  21.22 23.24
@end example

@noindent
this program, using our function to format the results, prints:

@example
   5.6
  21.2
@end example

This function deletes all the elements in an array (recall that the
extra whitespace signifies the start of the local variable list):

@example
@group
function delarray(a,    i)
@{
    for (i in a)
        delete a[i]
@}
@end group
@end example

When working with arrays, it is often necessary to delete all the elements
in an array and start over with a new list of elements
(@pxref{Delete}).
Instead of having
to repeat this loop everywhere that you need to clear out
an array, your program can just call @code{delarray()}.
(This guarantees portability.  The use of @samp{delete @var{array}} to delete
the contents of an entire array is a relatively recent@footnote{Late in 2012.}
addition to the POSIX standard.)

The following is an example of a recursive function.  It takes a string
as an input parameter and returns the string in reverse order.
Recursive functions must always have a test that stops the recursion.
In this case, the recursion terminates when the input string is
already empty:

@c 8/2014: Thanks to Mike Brennan for the improved formulation
@cindex @code{rev()} user-defined function
@example
function rev(str)
@{
    if (str == "")
        return ""

    return (rev(substr(str, 2)) substr(str, 1, 1))
@}
@end example

If this function is in a file named @file{rev.awk}, it can be tested
this way:

@example
$ @kbd{echo "Don't Panic!" |}
> @kbd{gawk -e '@{ print rev($0) @}' -f rev.awk}
@print{} !cinaP t'noD
@end example

The C @code{ctime()} function takes a timestamp and returns it as a string,
formatted in a well-known fashion.
The following example uses the built-in @code{strftime()} function
(@pxref{Time Functions})
to create an @command{awk} version of @code{ctime()}:

@cindex @code{ctime()} user-defined function
@example
@c file eg/lib/ctime.awk
# ctime.awk
#
# awk version of C ctime(3) function

function ctime(ts,    format)
@{
    format = "%a %b %e %H:%M:%S %Z %Y"

    if (ts == 0)
        ts = systime()       # use current time as default
    return strftime(format, ts)
@}
@c endfile
@end example

You might think that @code{ctime()} could use @code{PROCINFO["strftime"]}
for its format string. That would be a mistake, because @code{ctime()} is
supposed to return the time formatted in a standard fashion, and user-level
code could have changed @code{PROCINFO["strftime"]}.

@node Function Caveats
@subsection Calling User-Defined Functions

@cindex functions, user-defined, calling
@dfn{Calling a function} means causing the function to run and do its job.
A function call is an expression and its value is the value returned by
the function.

@menu
* Calling A Function::          Don't use spaces.
* Variable Scope::              Controlling variable scope.
* Pass By Value/Reference::     Passing parameters.
@end menu

@node Calling A Function
@subsubsection Writing a Function Call

A function call consists of the function name followed by the arguments
in parentheses.  @command{awk} expressions are what you write in the
call for the arguments.  Each time the call is executed, these
expressions are evaluated, and the values become the actual arguments.  For
example, here is a call to @code{foo()} with three arguments (the first
being a string concatenation):

@example
foo(x y, "lose", 4 * z)
@end example

@quotation CAUTION
Whitespace characters (spaces and TABs) are not allowed
between the function name and the opening parenthesis of the argument list.
If you write whitespace by mistake, @command{awk} might think that you mean
to concatenate a variable with an expression in parentheses.  However, it
notices that you used a function name and not a variable name, and reports
an error.
@end quotation

@node Variable Scope
@subsubsection Controlling Variable Scope

@cindex local variables, in a function
@cindex variables, local to a function
Unlike in many languages,
there is no way to make a variable local to a @code{@{} @dots{} @code{@}} block in
@command{awk}, but you can make a variable local to a function. It is
good practice to do so whenever a variable is needed only in that
function.

To make a variable local to a function, simply declare the variable as
an argument after the actual function arguments
(@pxref{Definition Syntax}).
Look at the following example, where variable
@code{i} is a global variable used by both functions @code{foo()} and
@code{bar()}:

@example
function bar()
@{
    for (i = 0; i < 3; i++)
        print "bar's i=" i
@}

function foo(j)
@{
    i = j + 1
    print "foo's i=" i
    bar()
    print "foo's i=" i
@}

BEGIN @{
      i = 10
      print "top's i=" i
      foo(0)
      print "top's i=" i
@}
@end example

Running this script produces the following, because the @code{i} in
functions @code{foo()} and @code{bar()} and at the top level refer to the same
variable instance:

@example
top's i=10
foo's i=1
bar's i=0
bar's i=1
bar's i=2
foo's i=3
top's i=3
@end example

If you want @code{i} to be local to both @code{foo()} and @code{bar()}, do as
follows (the extra space before @code{i} is a coding convention to
indicate that @code{i} is a local variable, not an argument):

@example
function bar(    i)
@{
    for (i = 0; i < 3; i++)
        print "bar's i=" i
@}

function foo(j,    i)
@{
    i = j + 1
    print "foo's i=" i
    bar()
    print "foo's i=" i
@}

BEGIN @{
      i = 10
      print "top's i=" i
      foo(0)
      print "top's i=" i
@}
@end example

Running the corrected script produces the following:

@example
top's i=10
foo's i=1
bar's i=0
bar's i=1
bar's i=2
foo's i=1
top's i=10
@end example

Besides scalar values (strings and numbers), you may also have
local arrays.  By using a parameter name as an array, @command{awk}
treats it as an array, and it is local to the function.
In addition, recursive calls create new arrays.
Consider this example:

@example
@group
function some_func(p1,      a)
@{
    if (p1++ > 3)
        return
@end group

    a[p1] = p1

    some_func(p1)

    printf("At level %d, index %d %s found in a\n",
         p1, (p1 - 1), (p1 - 1) in a ? "is" : "is not")
    printf("At level %d, index %d %s found in a\n",
         p1, p1, p1 in a ? "is" : "is not")
    print ""
@}

BEGIN @{
    some_func(1)
@}
@end example

When run, this program produces the following output:

@example
At level 4, index 3 is not found in a
At level 4, index 4 is found in a

At level 3, index 2 is not found in a
At level 3, index 3 is found in a

At level 2, index 1 is not found in a
At level 2, index 2 is found in a
@end example

@node Pass By Value/Reference
@subsubsection Passing Function Arguments by Value Or by Reference

In @command{awk}, when you declare a function, there is no way to
declare explicitly whether the arguments are passed @dfn{by value} or
@dfn{by reference}.

Instead, the passing convention is determined at runtime when
the function is called, according to the following rule:
if the argument is an array variable, then it is passed by reference.
Otherwise, the argument is passed by value.

@cindex call by value
Passing an argument by value means that when a function is called, it
is given a @emph{copy} of the value of this argument.
The caller may use a variable as the expression for the argument, but
the called function does not know this---it only knows what value the
argument had.  For example, if you write the following code:

@example
foo = "bar"
z = myfunc(foo)
@end example

@noindent
then you should not think of the argument to @code{myfunc()} as being
``the variable @code{foo}.''  Instead, think of the argument as the
string value @code{"bar"}.
If the function @code{myfunc()} alters the values of its local variables,
this has no effect on any other variables.  Thus, if @code{myfunc()}
does this:

@example
@group
function myfunc(str)
@{
   print str
   str = "zzz"
   print str
@}
@end group
@end example

@noindent
to change its first argument variable @code{str}, it does @emph{not}
change the value of @code{foo} in the caller.  The role of @code{foo} in
calling @code{myfunc()} ended when its value (@code{"bar"}) was computed.
If @code{str} also exists outside of @code{myfunc()}, the function body
cannot alter this outer value, because it is shadowed during the
execution of @code{myfunc()} and cannot be seen or changed from there.

@cindex call by reference
@cindex arrays, as parameters to functions
@cindex functions, arrays as parameters to
However, when arrays are the parameters to functions, they are @emph{not}
copied.  Instead, the array itself is made available for direct manipulation
by the function.  This is usually termed @dfn{call by reference}.
Changes made to an array parameter inside the body of a function @emph{are}
visible outside that function.

@quotation NOTE
Changing an array parameter inside a function
can be very dangerous if you do not watch what you are doing.
For example:

@example
function changeit(array, ind, nvalue)
@{
     array[ind] = nvalue
@}

BEGIN @{
    a[1] = 1; a[2] = 2; a[3] = 3
    changeit(a, 2, "two")
    printf "a[1] = %s, a[2] = %s, a[3] = %s\n",
            a[1], a[2], a[3]
@}
@end example

@noindent
prints @samp{a[1] = 1, a[2] = two, a[3] = 3}, because
@code{changeit()} stores @code{"two"} in the second element of @code{a}.
@end quotation

@cindex undefined functions
@cindex functions, undefined
Some @command{awk} implementations allow you to call a function that
has not been defined. They only report a problem at runtime, when the
program actually tries to call the function. For example:

@example
BEGIN @{
    if (0)
        foo()
    else
        bar()
@}
function bar() @{ @dots{} @}
# note that `foo' is not defined
@end example

@noindent
Because the @samp{if} statement will never be true, it is not really a
problem that @code{foo()} has not been defined.  Usually, though, it is a
problem if a program calls an undefined function.

@cindex lint checking, undefined functions
If @option{--lint} is specified
(@pxref{Options}),
@command{gawk} reports calls to undefined functions.

@cindex portability, @code{next} statement in user-defined functions
Some @command{awk} implementations generate a runtime
error if you use either the @code{next} statement
or the @code{nextfile} statement
(@pxref{Next Statement}, and
@ifdocbook
@ref{Nextfile Statement})
@end ifdocbook
@ifnotdocbook
@pxref{Nextfile Statement})
@end ifnotdocbook
inside a user-defined function.
@command{gawk} does not have this limitation.

@node Return Statement
@subsection The @code{return} Statement
@cindex @code{return} statement@comma{} user-defined functions

As seen in several earlier examples,
the body of a user-defined function can contain a @code{return} statement.
This statement returns control to the calling part of the @command{awk} program.  It
can also be used to return a value for use in the rest of the @command{awk}
program.  It looks like this:

@display
@code{return} [@var{expression}]
@end display

The @var{expression} part is optional.
Due most likely to an oversight, POSIX does not define what the return
value is if you omit the @var{expression}.  Technically speaking, this
makes the returned value undefined, and therefore, unpredictable.
In practice, though, all versions of @command{awk} simply return the
null string, which acts like zero if used in a numeric context.

A @code{return} statement without an @var{expression} is assumed at the end of
every function definition.  So, if control reaches the end of the function
body, then technically the function returns an unpredictable value.
In practice, it returns the empty string.  @command{awk}
does @emph{not} warn you if you use the return value of such a function.

Sometimes, you want to write a function for what it does, not for
what it returns.  Such a function corresponds to a @code{void} function
in C, C++, or Java, or to a @code{procedure} in Ada.  Thus, it may be appropriate to not
return any value; simply bear in mind that you should not be using the
return value of such a function.

The following is an example of a user-defined function that returns a value
for the largest number among the elements of an array:

@example
function maxelt(vec,   i, ret)
@{
     for (i in vec) @{
          if (ret == "" || vec[i] > ret)
               ret = vec[i]
     @}
     return ret
@}
@end example

@cindex programming conventions, function parameters
@noindent
You call @code{maxelt()} with one argument, which is an array name.  The local
variables @code{i} and @code{ret} are not intended to be arguments;
there is nothing to stop you from passing more than one argument
to @code{maxelt()} but the results would be strange.  The extra space before
@code{i} in the function parameter list indicates that @code{i} and
@code{ret} are local variables.
You should follow this convention when defining functions.

The following program uses the @code{maxelt()} function.  It loads an
array, calls @code{maxelt()}, and then reports the maximum number in that
array:

@example
function maxelt(vec,   i, ret)
@{
     for (i in vec) @{
          if (ret == "" || vec[i] > ret)
               ret = vec[i]
     @}
     return ret
@}

@group
# Load all fields of each record into nums.
@{
     for(i = 1; i <= NF; i++)
          nums[NR, i] = $i
@}
@end group

END @{
     print maxelt(nums)
@}
@end example

Given the following input:

@example
 1 5 23 8 16
44 3 5 2 8 26
256 291 1396 2962 100
-6 467 998 1101
99385 11 0 225
@end example

@noindent
the program reports (predictably) that 99,385 is the largest value
in the array.

@node Dynamic Typing
@subsection Functions and Their Effects on Variable Typing

@command{awk} is a very fluid language.
It is possible that @command{awk} can't tell if an identifier
represents a scalar variable or an array until runtime.
Here is an annotated sample program:

@example
function foo(a)
@{
    a[1] = 1   # parameter is an array
@}

BEGIN @{
    b = 1
    foo(b)  # invalid: fatal type mismatch

    foo(x)  # x uninitialized, becomes an array dynamically
    x = 1   # now not allowed, runtime error
@}
@end example

In this example, the first call to @code{foo()} generates
a fatal error, so @command{awk} will not report the second
error. If you comment out that call, though, then @command{awk}
does report the second error.

Usually, such things aren't a big issue, but it's worth
being aware of them.

@node Indirect Calls
@section Indirect Function Calls

@cindex indirect function calls
@cindex function calls, indirect
@cindex function pointers
@cindex pointers to functions
@cindex differences in @command{awk} and @command{gawk}, indirect function calls

This section describes an advanced, @command{gawk}-specific extension.

Often, you may wish to defer the choice of function to call until runtime.
For example, you may have different kinds of records, each of which
should be processed differently.

Normally, you would have to use a series of @code{if}-@code{else}
statements to decide which function to call.  By using @dfn{indirect}
function calls, you can specify the name of the function to call as a
string variable, and then call the function.  Let's look at an example.

Suppose you have a file with your test scores for the classes you
are taking, and
you wish to get the sum and the average of
your test scores.
The first field is the class name. The following fields
are the functions to call to process the data, up to a ``marker''
field @samp{data:}.  Following the marker, to the end of the record,
are the various numeric test scores.

Here is the initial file:

@example
@c file eg/data/class_data1
Biology_101 sum average data: 87.0 92.4 78.5 94.9
Chemistry_305 sum average data: 75.2 98.3 94.7 88.2
English_401 sum average data: 100.0 95.6 87.1 93.4
@c endfile
@end example

To process the data, you might write initially:

@example
@{
    class = $1
    for (i = 2; $i != "data:"; i++) @{
        if ($i == "sum")
            sum()   # processes the whole record
        else if ($i == "average")
            average()
        @dots{}           # and so on
    @}
@}
@end example

@noindent
This style of programming works, but can be awkward.  With @dfn{indirect}
function calls, you tell @command{gawk} to use the @emph{value} of a
variable as the @emph{name} of the function to call.

@cindex @code{@@}-notation for indirect function calls
@cindex indirect function calls, @code{@@}-notation
@cindex function calls, indirect, @code{@@}-notation for
The syntax is similar to that of a regular function call: an identifier
immediately followed by an opening parenthesis, any arguments, and then
a closing parenthesis, with the addition of a leading @samp{@@}
character:

@example
the_func = "sum"
result = @@the_func()   # calls the sum() function
@end example

Here is a full program that processes the previously shown data,
using indirect function calls:

@example
@c file eg/prog/indirectcall.awk
# indirectcall.awk --- Demonstrate indirect function calls
@c endfile
@ignore
@c file eg/prog/indirectcall.awk
#
# Arnold Robbins, arnold@skeeve.com, Public Domain
# January 2009
@c endfile
@end ignore

@c file eg/prog/indirectcall.awk
# average --- return the average of the values in fields $first - $last

function average(first, last,   sum, i)
@{
    sum = 0;
    for (i = first; i <= last; i++)
        sum += $i

    return sum / (last - first + 1)
@}

# sum --- return the sum of the values in fields $first - $last

function sum(first, last,   ret, i)
@{
    ret = 0;
    for (i = first; i <= last; i++)
        ret += $i

    return ret
@}
@c endfile
@end example

These two functions expect to work on fields; thus, the parameters
@code{first} and @code{last} indicate where in the fields to start and end.
Otherwise, they perform the expected computations and are not unusual:

@example
@c file eg/prog/indirectcall.awk
# For each record, print the class name and the requested statistics
@{
    class_name = $1
    gsub(/_/, " ", class_name)  # Replace _ with spaces

    # find start
    for (i = 1; i <= NF; i++) @{
        if ($i == "data:") @{
            start = i + 1
            break
        @}
    @}

    printf("%s:\n", class_name)
    for (i = 2; $i != "data:"; i++) @{
        the_function = $i
        printf("\t%s: <%s>\n", $i, @@the_function(start, NF) "")
    @}
    print ""
@}
@c endfile
@end example

This is the main processing for each record. It prints the class name (with
underscores replaced with spaces). It then finds the start of the actual data,
saving it in @code{start}.
The last part of the code loops through each function name (from @code{$2} up to
the marker, @samp{data:}), calling the function named by the field. The indirect
function call itself occurs as a parameter in the call to @code{printf}.
(The @code{printf} format string uses @samp{%s} as the format specifier so that we
can use functions that return strings, as well as numbers. Note that the result
from the indirect call is concatenated with the empty string, in order to force
it to be a string value.)

Here is the result of running the program:

@example
$ @kbd{gawk -f indirectcall.awk class_data1}
@print{} Biology 101:
@print{}     sum: <352.8>
@print{}     average: <88.2>
@print{}
@print{} Chemistry 305:
@print{}     sum: <356.4>
@print{}     average: <89.1>
@print{}
@print{} English 401:
@print{}     sum: <376.1>
@print{}     average: <94.025>
@end example

The ability to use indirect function calls is more powerful than you may
think at first.  The C and C++ languages provide ``function pointers,'' which
are a mechanism for calling a function chosen at runtime.  One of the most
well-known uses of this ability is the C @code{qsort()} function, which sorts
an array using the famous ``quicksort'' algorithm
(see @uref{https://en.wikipedia.org/wiki/Quicksort, the Wikipedia article}
for more information).  To use this function, you supply a pointer to a comparison
function.  This mechanism allows you to sort arbitrary data in an arbitrary
fashion.

We can do something similar using @command{gawk}, like this:

@example
@c file eg/lib/quicksort.awk
# quicksort.awk --- Quicksort algorithm, with user-supplied
#                   comparison function
@c endfile
@ignore
@c file eg/lib/quicksort.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# January 2009

@c endfile
@end ignore
@c file eg/lib/quicksort.awk

# quicksort --- C.A.R. Hoare's quicksort algorithm. See Wikipedia
#               or almost any algorithms or computer science text.
@c endfile
@ignore
@c file eg/lib/quicksort.awk
#
# Adapted from K&R-II, page 110
@c endfile
@end ignore
@c file eg/lib/quicksort.awk

function quicksort(data, left, right, less_than,    i, last)
@{
    if (left >= right)  # do nothing if array contains fewer
        return          # than two elements

    quicksort_swap(data, left, int((left + right) / 2))
    last = left
    for (i = left + 1; i <= right; i++)
        if (@@less_than(data[i], data[left]))
            quicksort_swap(data, ++last, i)
    quicksort_swap(data, left, last)
    quicksort(data, left, last - 1, less_than)
    quicksort(data, last + 1, right, less_than)
@}

# quicksort_swap --- helper function for quicksort, should really be inline

function quicksort_swap(data, i, j,      temp)
@{
    temp = data[i]
    data[i] = data[j]
    data[j] = temp
@}
@c endfile
@end example

The @code{quicksort()} function receives the @code{data} array, the starting and ending
indices to sort (@code{left} and @code{right}), and the name of a function that
performs a ``less than'' comparison.  It then implements the quicksort algorithm.

To make use of the sorting function, we return to our previous example. The
first thing to do is write some comparison functions:

@example
@c file eg/prog/indirectcall.awk
@group
# num_lt --- do a numeric less than comparison

function num_lt(left, right)
@{
    return ((left + 0) < (right + 0))
@}
@end group

# num_ge --- do a numeric greater than or equal to comparison

function num_ge(left, right)
@{
    return ((left + 0) >= (right + 0))
@}
@c endfile
@end example

The @code{num_ge()} function is needed to perform a descending sort; when used
to perform a ``less than'' test, it actually does the opposite (greater than
or equal to), which yields data sorted in descending order.

Next comes a sorting function.  It is parameterized with the starting and
ending field numbers and the comparison function. It builds an array with
the data and calls @code{quicksort()} appropriately, and then formats the
results as a single string:

@example
@c file eg/prog/indirectcall.awk
# do_sort --- sort the data according to `compare'
#             and return it as a string

function do_sort(first, last, compare,      data, i, retval)
@{
    delete data
    for (i = 1; first <= last; first++) @{
        data[i] = $first
        i++
    @}

    quicksort(data, 1, i-1, compare)

    retval = data[1]
    for (i = 2; i in data; i++)
        retval = retval " " data[i]

    return retval
@}
@c endfile
@end example

Finally, the two sorting functions call @code{do_sort()}, passing in the
names of the two comparison functions:

@example
@c file eg/prog/indirectcall.awk
@group
# sort --- sort the data in ascending order and return it as a string

function sort(first, last)
@{
    return do_sort(first, last, "num_lt")
@}
@end group

@group
# rsort --- sort the data in descending order and return it as a string

function rsort(first, last)
@{
    return do_sort(first, last, "num_ge")
@}
@end group
@c endfile
@end example

Here is an extended version of the @value{DF}:

@example
@c file eg/data/class_data2
Biology_101 sum average sort rsort data: 87.0 92.4 78.5 94.9
Chemistry_305 sum average sort rsort data: 75.2 98.3 94.7 88.2
English_401 sum average sort rsort data: 100.0 95.6 87.1 93.4
@c endfile
@end example

Finally, here are the results when the enhanced program is run:

@example
$ @kbd{gawk -f quicksort.awk -f indirectcall.awk class_data2}
@print{} Biology 101:
@print{}     sum: <352.8>
@print{}     average: <88.2>
@print{}     sort: <78.5 87.0 92.4 94.9>
@print{}     rsort: <94.9 92.4 87.0 78.5>
@print{}
@print{} Chemistry 305:
@print{}     sum: <356.4>
@print{}     average: <89.1>
@print{}     sort: <75.2 88.2 94.7 98.3>
@print{}     rsort: <98.3 94.7 88.2 75.2>
@print{}
@print{} English 401:
@print{}     sum: <376.1>
@print{}     average: <94.025>
@print{}     sort: <87.1 93.4 95.6 100.0>
@print{}     rsort: <100.0 95.6 93.4 87.1>
@end example

Another example where indirect functions calls are useful can be found in
processing arrays. This is described in @ref{Walking Arrays}.

Remember that you must supply a leading @samp{@@} in front of an indirect function call.

Starting with @value{PVERSION} 4.1.2 of @command{gawk}, indirect function
calls may also be used with built-in functions and with extension functions
(@pxref{Dynamic Extensions}). There are some limitations when calling
built-in functions indirectly, as follows.

@itemize @value{BULLET}
@item
You cannot pass a regular expression constant to a built-in function
through an indirect function call.@footnote{This may change in a future
version; recheck the documentation that comes with your version of
@command{gawk} to see if it has.} This applies to the @code{sub()},
@code{gsub()}, @code{gensub()}, @code{match()}, @code{split()} and
@code{patsplit()} functions.

@item
If calling @code{sub()} or @code{gsub()}, you may only pass two arguments,
since those functions are unusual in that they update their third argument.
This means that @code{$0} will be updated.
@end itemize

@command{gawk} does its best to make indirect function calls efficient.
For example, in the following case:

@example
for (i = 1; i <= n; i++)
    @@the_func()
@end example

@noindent
@command{gawk} looks up the actual function to call only once.

@node Functions Summary
@section Summary

@itemize @value{BULLET}
@item
@command{awk} provides built-in functions and lets you define your own
functions.

@item
POSIX @command{awk} provides three kinds of built-in functions: numeric,
string, and I/O.  @command{gawk} provides functions that sort arrays, work
with values representing time, do bit manipulation, determine variable
type (array versus scalar), and internationalize and localize programs.
@command{gawk} also provides several extensions to some of standard
functions, typically in the form of additional arguments.

@item
Functions accept zero or more arguments and return a value.  The
expressions that provide the argument values are completely evaluated
before the function is called. Order of evaluation is not defined.
The return value can be ignored.

@item
The handling of backslash in @code{sub()} and @code{gsub()} is not simple.
It is more straightforward in @command{gawk}'s @code{gensub()} function,
but that function still requires care in its use.

@item
User-defined functions provide important capabilities but come with
some syntactic inelegancies. In a function call, there cannot be any
space between the function name and the opening left parenthesis of the
argument list.  Also, there is no provision for local variables, so the
convention is to add extra parameters, and to separate them visually
from the real parameters by extra whitespace.

@item
User-defined functions may call other user-defined (and built-in)
functions and may call themselves recursively. Function parameters
``hide'' any global variables of the same names.
You cannot use the name of a reserved variable (such as @code{ARGC})
as the name of a parameter in user-defined functions.

@item
Scalar values are passed to user-defined functions by value. Array
parameters are passed by reference; any changes made by the function to
array parameters are thus visible after the function has returned.

@item
Use the @code{return} statement to return from a user-defined function.
An optional expression becomes the function's return value.  Only scalar
values may be returned by a function.

@item
If a variable that has never been used is passed to a user-defined
function, how that function treats the variable can set its nature:
either scalar or array.

@item
@command{gawk} provides indirect function calls using a special syntax.
By setting a variable to the name of a function, you can
determine at runtime what function will be called at that point in the
program. This is equivalent to function pointers in C and C++.

@end itemize


@ifnotinfo
@part @value{PART2}Problem Solving with @command{awk}
@end ifnotinfo

@ifdocbook
Part II shows how to use @command{awk} and @command{gawk} for problem solving.
There is lots of code here for you to read and learn from.
It contains the following chapters:

@itemize @value{BULLET}
@item
@ref{Library Functions}

@item
@ref{Sample Programs}
@end itemize
@end ifdocbook

@node Library Functions
@chapter A Library of @command{awk} Functions
@cindex libraries of @command{awk} functions
@cindex functions, library
@cindex functions, user-defined, library of

@ref{User-defined} describes how to write
your own @command{awk} functions.  Writing functions is important, because
it allows you to encapsulate algorithms and program tasks in a single
place.  It simplifies programming, making program development more
manageable and making programs more readable.

@cindex Kernighan, Brian
@cindex Plauger, P.J.@:
In their seminal 1976 book, @cite{Software Tools},@footnote{Sadly, over 35
years later, many of the lessons taught by this book have yet to be
learned by a vast number of practicing programmers.} Brian Kernighan
and P.J.@: Plauger wrote:

@quotation
Good Programming is not learned from generalities, but by seeing how
significant programs can be made clean, easy to read, easy to maintain and
modify, human-engineered, efficient and reliable, by the application of
common sense and good programming practices.  Careful study and imitation
of good programs leads to better writing.
@end quotation

In fact, they felt this idea was so important that they placed this
statement on the cover of their book.  Because we believe strongly
that their statement is correct, this @value{CHAPTER} and @ref{Sample
Programs}, provide a good-sized body of code for you to read and, we hope,
to learn from.

This @value{CHAPTER} presents a library of useful @command{awk} functions.
Many of the sample programs presented later in this @value{DOCUMENT}
use these functions.
The functions are presented here in a progression from simple to complex.

@cindex Texinfo
@ref{Extract Program}
presents a program that you can use to extract the source code for
these example library functions and programs from the Texinfo source
for this @value{DOCUMENT}.
(This has already been done as part of the @command{gawk} distribution.)

@ifclear FOR_PRINT
If you have written one or more useful, general-purpose @command{awk} functions
and would like to contribute them to the @command{awk} user community, see
@ref{How To Contribute}, for more information.
@end ifclear

@cindex portability, example programs
The programs in this @value{CHAPTER} and in
@ref{Sample Programs},
freely use @command{gawk}-specific features.
Rewriting these programs for different implementations of @command{awk}
is pretty straightforward:

@itemize @value{BULLET}
@item
Diagnostic error messages are sent to @file{/dev/stderr}.
Use @samp{| "cat 1>&2"} instead of @samp{> "/dev/stderr"} if your system
does not have a @file{/dev/stderr}, or if you cannot use @command{gawk}.

@item
A number of programs use @code{nextfile}
(@pxref{Nextfile Statement})
to skip any remaining input in the input file.

@item
@c 12/2000: Thanks to Nelson Beebe for pointing out the output issue.
@cindex case sensitivity, example programs
@cindex @code{IGNORECASE} variable, in example programs
Finally, some of the programs choose to ignore upper- and lowercase
distinctions in their input. They do so by assigning one to @code{IGNORECASE}.
You can achieve almost the same effect@footnote{The effects are
not identical.  Output of the transformed
record will be in all lowercase, while @code{IGNORECASE} preserves the original
contents of the input record.} by adding the following rule to the
beginning of the program:

@example
# ignore case
@{ $0 = tolower($0) @}
@end example

@noindent
Also, verify that all regexp and string constants used in
comparisons use only lowercase letters.
@end itemize

@menu
* Library Names::               How to best name private global variables in
                                library functions.
* General Functions::           Functions that are of general use.
* Data File Management::        Functions for managing command-line data
                                files.
* Getopt Function::             A function for processing command-line
                                arguments.
* Passwd Functions::            Functions for getting user information.
* Group Functions::             Functions for getting group information.
* Walking Arrays::              A function to walk arrays of arrays.
* Library Functions Summary::   Summary of library functions.
* Library Exercises::           Exercises.
@end menu

@node Library Names
@section Naming Library Function Global Variables

@cindex names, arrays/variables
@cindex names, functions
@cindex namespace issues
@cindex @command{awk} programs, documenting
@cindex documentation, of @command{awk} programs
Due to the way the @command{awk} language evolved, variables are either
@dfn{global} (usable by the entire program) or @dfn{local} (usable just by
a specific function).  There is no intermediate state analogous to
@code{static} variables in C.

@cindex variables, global, for library functions
@cindex private variables
@cindex variables, private
Library functions often need to have global variables that they can use to
preserve state information between calls to the function---for example,
@code{getopt()}'s variable @code{_opti}
(@pxref{Getopt Function}).
Such variables are called @dfn{private}, as the only functions that need to
use them are the ones in the library.

When writing a library function, you should try to choose names for your
private variables that will not conflict with any variables used by
either another library function or a user's main program.  For example, a
name like @code{i} or @code{j} is not a good choice, because user programs
often use variable names like these for their own purposes.

@cindex programming conventions, private variable names
The example programs shown in this @value{CHAPTER} all start the names of their
private variables with an underscore (@samp{_}).  Users generally don't use
leading underscores in their variable names, so this convention immediately
decreases the chances that the variable names will be accidentally shared
with the user's program.

@cindex @code{_} (underscore), in names of private variables
@cindex underscore (@code{_}), in names of private variables
In addition, several of the library functions use a prefix that helps
indicate what function or set of functions use the variables---for example,
@code{_pw_byname()} in the user database routines
(@pxref{Passwd Functions}).
This convention is recommended, as it even further decreases the
chance of inadvertent conflict among variable names.  Note that this
convention is used equally well for variable names and for private
function names.@footnote{Although all the library routines could have
been rewritten to use this convention, this was not done, in order to
show how our own @command{awk} programming style has evolved and to
provide some basis for this discussion.}

As a final note on variable naming, if a function makes global variables
available for use by a main program, it is a good convention to start those
variables' names with a capital letter---for
example, @code{getopt()}'s @code{Opterr} and @code{Optind} variables
(@pxref{Getopt Function}).
The leading capital letter indicates that it is global, while the fact that
the variable name is not all capital letters indicates that the variable is
not one of @command{awk}'s predefined variables, such as @code{FS}.

@cindex @option{--dump-variables} option, using for library functions
It is also important that @emph{all} variables in library
functions that do not need to save state are, in fact, declared
local.@footnote{@command{gawk}'s @option{--dump-variables} command-line
option is useful for verifying this.} If this is not done, the variables
could accidentally be used in the user's program, leading to bugs that
are very difficult to track down:

@example
function lib_func(x, y,    l1, l2)
@{
    @dots{}
    # some_var should be local but by oversight is not
    @var{use variable} some_var
    @dots{}
@}
@end example

@cindex arrays, associative, library functions and
@cindex libraries of @command{awk} functions, associative arrays and
@cindex functions, library, associative arrays and
@cindex Tcl
A different convention, common in the Tcl community, is to use a single
associative array to hold the values needed by the library function(s), or
``package.''  This significantly decreases the number of actual global names
in use.  For example, the functions described in
@ref{Passwd Functions}
might have used array elements @code{@w{PW_data["inited"]}}, @code{@w{PW_data["total"]}},
@code{@w{PW_data["count"]}}, and @code{@w{PW_data["awklib"]}}, instead of
@code{@w{_pw_inited}}, @code{@w{_pw_awklib}}, @code{@w{_pw_total}},
and @code{@w{_pw_count}}.

The conventions presented in this @value{SECTION} are exactly
that: conventions. You are not required to write your programs this
way---we merely recommend that you do so.

@node General Functions
@section General Programming

This @value{SECTION} presents a number of functions that are of general
programming use.

@menu
* Strtonum Function::           A replacement for the built-in
                                @code{strtonum()} function.
* Assert Function::             A function for assertions in @command{awk}
                                programs.
* Round Function::              A function for rounding if @code{sprintf()}
                                does not do it correctly.
* Cliff Random Function::       The Cliff Random Number Generator.
* Ordinal Functions::           Functions for using characters as numbers and
                                vice versa.
* Join Function::               A function to join an array into a string.
* Getlocaltime Function::       A function to get formatted times.
* Readfile Function::           A function to read an entire file at once.
* Shell Quoting::               A function to quote strings for the shell.
@end menu

@node Strtonum Function
@subsection Converting Strings to Numbers

The @code{strtonum()} function (@pxref{String Functions})
is a @command{gawk} extension.  The following function
provides an implementation for other versions of @command{awk}:

@example
@c file eg/lib/strtonum.awk
# mystrtonum --- convert string to number

@c endfile
@ignore
@c file eg/lib/strtonum.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# February, 2004
# Revised June, 2014

@c endfile
@end ignore
@c file eg/lib/strtonum.awk
function mystrtonum(str,        ret, n, i, k, c)
@{
    if (str ~ /^0[0-7]*$/) @{
        # octal
        n = length(str)
        ret = 0
        for (i = 1; i <= n; i++) @{
            c = substr(str, i, 1)
            # index() returns 0 if c not in string,
            # includes c == "0"
            k = index("1234567", c)

            ret = ret * 8 + k
        @}
    @} else if (str ~ /^0[xX][[:xdigit:]]+$/) @{
        # hexadecimal
        str = substr(str, 3)    # lop off leading 0x
        n = length(str)
        ret = 0
        for (i = 1; i <= n; i++) @{
            c = substr(str, i, 1)
            c = tolower(c)
            # index() returns 0 if c not in string,
            # includes c == "0"
            k = index("123456789abcdef", c)

            ret = ret * 16 + k
        @}
    @} else if (str ~ \
  /^[-+]?([0-9]+([.][0-9]*([Ee][0-9]+)?)?|([.][0-9]+([Ee][-+]?[0-9]+)?))$/) @{
        # decimal number, possibly floating point
        ret = str + 0
    @} else
        ret = "NOT-A-NUMBER"

    return ret
@}

# BEGIN @{     # gawk test harness
#     a[1] = "25"
#     a[2] = ".31"
#     a[3] = "0123"
#     a[4] = "0xdeadBEEF"
#     a[5] = "123.45"
#     a[6] = "1.e3"
#     a[7] = "1.32"
#     a[8] = "1.32E2"
#
#     for (i = 1; i in a; i++)
#         print a[i], strtonum(a[i]), mystrtonum(a[i])
# @}
@c endfile
@end example

The function first looks for C-style octal numbers (base 8).
If the input string matches a regular expression describing octal
numbers, then @code{mystrtonum()} loops through each character in the
string.  It sets @code{k} to the index in @code{"1234567"} of the current
octal digit.
The return value will either be the same number as the digit, or zero
if the character is not there, which will be true for a @samp{0}.
This is safe, because the regexp test in the @code{if} ensures that
only octal values are converted.

Similar logic applies to the code that checks for and converts a
hexadecimal value, which starts with @samp{0x} or @samp{0X}.
The use of @code{tolower()} simplifies the computation for finding
the correct numeric value for each hexadecimal digit.

Finally, if the string matches the (rather complicated) regexp for a
regular decimal integer or floating-point number, the computation
@samp{ret = str + 0} lets @command{awk} convert the value to a
number.

A commented-out test program is included, so that the function can
be tested with @command{gawk} and the results compared to the built-in
@code{strtonum()} function.

@node Assert Function
@subsection Assertions

@cindex assertions
@cindex @code{assert()} function (C library)
@cindex libraries of @command{awk} functions, assertions
@cindex functions, library, assertions
@cindex @command{awk} programs, lengthy, assertions
When writing large programs, it is often useful to know
that a condition or set of conditions is true.  Before proceeding with a
particular computation, you make a statement about what you believe to be
the case.  Such a statement is known as an
@dfn{assertion}.  The C language provides an @code{<assert.h>} header file
and corresponding @code{assert()} macro that a programmer can use to make
assertions.  If an assertion fails, the @code{assert()} macro arranges to
print a diagnostic message describing the condition that should have
been true but was not, and then it kills the program.  In C, using
@code{assert()} looks this:

@example
@group
#include <assert.h>

int myfunc(int a, double b)
@{
     assert(a <= 5 && b >= 17.1);
     @dots{}
@}
@end group
@end example

If the assertion fails, the program prints a message similar to this:

@example
prog.c:5: assertion failed: a <= 5 && b >= 17.1
@end example

@cindex @code{assert()} user-defined function
The C language makes it possible to turn the condition into a string for use
in printing the diagnostic message.  This is not possible in @command{awk}, so
this @code{assert()} function also requires a string version of the condition
that is being tested.
Following is the function:

@example
@c file eg/lib/assert.awk
# assert --- assert that a condition is true. Otherwise, exit.

@c endfile
@ignore
@c file eg/lib/assert.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May, 1993

@c endfile
@end ignore
@c file eg/lib/assert.awk
function assert(condition, string)
@{
    if (! condition) @{
        printf("%s:%d: assertion failed: %s\n",
            FILENAME, FNR, string) > "/dev/stderr"
        _assert_exit = 1
        exit 1
    @}
@}

@group
END @{
    if (_assert_exit)
        exit 1
@}
@end group
@c endfile
@end example

The @code{assert()} function tests the @code{condition} parameter. If it
is false, it prints a message to standard error, using the @code{string}
parameter to describe the failed condition.  It then sets the variable
@code{_assert_exit} to one and executes the @code{exit} statement.
The @code{exit} statement jumps to the @code{END} rule. If the @code{END}
rule finds @code{_assert_exit} to be true, it exits immediately.

The purpose of the test in the @code{END} rule is to
keep any other @code{END} rules from running.  When an assertion fails, the
program should exit immediately.
If no assertions fail, then @code{_assert_exit} is still
false when the @code{END} rule is run normally, and the rest of the
program's @code{END} rules execute.
For all of this to work correctly, @file{assert.awk} must be the
first source file read by @command{awk}.
The function can be used in a program in the following way:

@example
function myfunc(a, b)
@{
     assert(a <= 5 && b >= 17.1, "a <= 5 && b >= 17.1")
     @dots{}
@}
@end example

@noindent
If the assertion fails, you see a message similar to the following:

@example
mydata:1357: assertion failed: a <= 5 && b >= 17.1
@end example

@cindex @code{END} pattern, @code{assert()} user-defined function and
There is a small problem with this version of @code{assert()}.
An @code{END} rule is automatically added
to the program calling @code{assert()}.  Normally, if a program consists
of just a @code{BEGIN} rule, the input files and/or standard input are
not read. However, now that the program has an @code{END} rule, @command{awk}
attempts to read the input @value{DF}s or standard input
(@pxref{Using BEGIN/END}),
most likely causing the program to hang as it waits for input.

@cindex @code{BEGIN} pattern, @code{assert()} user-defined function and
There is a simple workaround to this:
make sure that such a @code{BEGIN} rule always ends
with an @code{exit} statement.

@node Round Function
@subsection Rounding Numbers

@cindex rounding numbers
@cindex numbers, rounding
@cindex libraries of @command{awk} functions, rounding numbers
@cindex functions, library, rounding numbers
@cindex @code{print} statement, @code{sprintf()} function and
@cindex @code{printf} statement, @code{sprintf()} function and
@cindex @code{sprintf()} function, @code{print}/@code{printf} statements and
The way @code{printf} and @code{sprintf()}
(@pxref{Printf})
perform rounding often depends upon the system's C @code{sprintf()}
subroutine.  On many machines, @code{sprintf()} rounding is @dfn{unbiased},
which means it doesn't always round a trailing .5 up, contrary
to naive expectations.  In unbiased rounding, .5 rounds to even,
rather than always up, so 1.5 rounds to 2 but 4.5 rounds to 4.  This means
that if you are using a format that does rounding (e.g., @code{"%.0f"}),
you should check what your system does.  The following function does
traditional rounding; it might be useful if your @command{awk}'s @code{printf}
does unbiased rounding:

@cindex @code{round()} user-defined function
@example
@c file eg/lib/round.awk
# round.awk --- do normal rounding
@c endfile
@ignore
@c file eg/lib/round.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# August, 1996
@c endfile
@end ignore
@c file eg/lib/round.awk

function round(x,   ival, aval, fraction)
@{
   ival = int(x)    # integer part, int() truncates

   # see if fractional part
   if (ival == x)   # no fraction
      return ival   # ensure no decimals

   if (x < 0) @{
      aval = -x     # absolute value
      ival = int(aval)
      fraction = aval - ival
      if (fraction >= .5)
         return int(x) - 1   # -2.5 --> -3
      else
         return int(x)       # -2.3 --> -2
   @} else @{
      fraction = x - ival
      if (fraction >= .5)
         return ival + 1
      else
         return ival
   @}
@}
@c endfile
@c don't include test harness in the file that gets installed
@group
# test harness
# @{ print $0, round($0) @}
@end group
@end example

@node Cliff Random Function
@subsection The Cliff Random Number Generator
@cindex random numbers, Cliff
@cindex Cliff random numbers
@cindex numbers, Cliff random
@cindex functions, library, Cliff random numbers

The
@uref{http://mathworld.wolfram.com/CliffRandomNumberGenerator.html, Cliff random number generator}
is a very simple random number generator that ``passes the noise sphere test
for randomness by showing no structure.''
It is easily programmed, in less than 10 lines of @command{awk} code:

@cindex @code{cliff_rand()} user-defined function
@example
@c file eg/lib/cliff_rand.awk
# cliff_rand.awk --- generate Cliff random numbers
@c endfile
@ignore
@c file eg/lib/cliff_rand.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# December 2000
@c endfile
@end ignore
@c file eg/lib/cliff_rand.awk

BEGIN @{ _cliff_seed = 0.1 @}

function cliff_rand()
@{
    _cliff_seed = (100 * log(_cliff_seed)) % 1
    if (_cliff_seed < 0)
        _cliff_seed = - _cliff_seed
    return _cliff_seed
@}
@c endfile
@end example

This algorithm requires an initial ``seed'' of 0.1.  Each new value
uses the current seed as input for the calculation.
If the built-in @code{rand()} function
(@pxref{Numeric Functions})
isn't random enough, you might try using this function instead.

@node Ordinal Functions
@subsection Translating Between Characters and Numbers

@cindex libraries of @command{awk} functions, character values as numbers
@cindex functions, library, character values as numbers
@cindex characters, values of as numbers
@cindex numbers, as values of characters
One commercial implementation of @command{awk} supplies a built-in function,
@code{ord()}, which takes a character and returns the numeric value for that
character in the machine's character set.  If the string passed to
@code{ord()} has more than one character, only the first one is used.

The inverse of this function is @code{chr()} (from the function of the same
name in Pascal), which takes a number and returns the corresponding character.
Both functions are written very nicely in @command{awk}; there is no real
reason to build them into the @command{awk} interpreter:

@cindex @code{ord()} user-defined function
@cindex @code{chr()} user-defined function
@cindex @code{_ord_init()} user-defined function
@example
@c file eg/lib/ord.awk
# ord.awk --- do ord and chr

# Global identifiers:
#    _ord_:        numerical values indexed by characters
#    _ord_init:    function to initialize _ord_
@c endfile
@ignore
@c file eg/lib/ord.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# 16 January, 1992
# 20 July, 1992, revised
@c endfile
@end ignore
@c file eg/lib/ord.awk

BEGIN    @{ _ord_init() @}

function _ord_init(    low, high, i, t)
@{
    low = sprintf("%c", 7) # BEL is ascii 7
    if (low == "\a") @{    # regular ascii
        low = 0
        high = 127
    @} else if (sprintf("%c", 128 + 7) == "\a") @{
        # ascii, mark parity
        low = 128
        high = 255
    @} else @{        # ebcdic(!)
        low = 0
        high = 255
    @}

    for (i = low; i <= high; i++) @{
        t = sprintf("%c", i)
        _ord_[t] = i
    @}
@}
@c endfile
@end example

@cindex character sets (machine character encodings)
@cindex ASCII
@cindex EBCDIC
@cindex Unicode
@cindex mark parity
Some explanation of the numbers used by @code{_ord_init()} is worthwhile.
The most prominent character set in use today is ASCII.@footnote{This
is changing; many systems use Unicode, a very large character set
that includes ASCII as a subset.  On systems with full Unicode support,
a character can occupy up to 32 bits, making simple tests such as
used here prohibitively expensive.}
Although an
8-bit byte can hold 256 distinct values (from 0 to 255), ASCII only
defines characters that use the values from 0 to 127.@footnote{ASCII
has been extended in many countries to use the values from 128 to 255
for country-specific characters.  If your  system uses these extensions,
you can simplify @code{_ord_init()} to loop from 0 to 255.}
In the now distant past,
at least one minicomputer manufacturer
@c Pr1me, blech
used ASCII, but with mark parity, meaning that the leftmost bit in the byte
is always 1.  This means that on those systems, characters
have numeric values from 128 to 255.
Finally, large mainframe systems use the EBCDIC character set, which
uses all 256 values.
There are other character sets in use on some older systems, but
they are not really worth worrying about:

@example
@c file eg/lib/ord.awk
function ord(str,    c)
@{
    # only first character is of interest
    c = substr(str, 1, 1)
    return _ord_[c]
@}

function chr(c)
@{
    # force c to be numeric by adding 0
    return sprintf("%c", c + 0)
@}
@c endfile

#### test code ####
# BEGIN @{
#    for (;;) @{
#        printf("enter a character: ")
#        if (getline var <= 0)
#            break
#        printf("ord(%s) = %d\n", var, ord(var))
#    @}
# @}
@c endfile
@end example

An obvious improvement to these functions is to move the code for the
@code{@w{_ord_init}} function into the body of the @code{BEGIN} rule.  It was
written this way initially for ease of development.
There is a ``test program'' in a @code{BEGIN} rule, to test the
function.  It is commented out for production use.

@node Join Function
@subsection Merging an Array into a String

@cindex libraries of @command{awk} functions, merging arrays into strings
@cindex functions, library, merging arrays into strings
@cindex strings, merging arrays into
@cindex arrays, merging into strings
When doing string processing, it is often useful to be able to join
all the strings in an array into one long string.  The following function,
@code{join()}, accomplishes this task.  It is used later in several of
the application programs
(@pxref{Sample Programs}).

Good function design is important; this function needs to be general, but it
should also have a reasonable default behavior.  It is called with an array
as well as the beginning and ending indices of the elements in the array to be
merged.  This assumes that the array indices are numeric---a reasonable
assumption, as the array was likely created with @code{split()}
(@pxref{String Functions}):

@cindex @code{join()} user-defined function
@example
@c file eg/lib/join.awk
# join.awk --- join an array into a string
@c endfile
@ignore
@c file eg/lib/join.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
@c endfile
@end ignore
@c file eg/lib/join.awk

function join(array, start, end, sep,    result, i)
@{
    if (sep == "")
       sep = " "
    else if (sep == SUBSEP) # magic value
       sep = ""
    result = array[start]
    for (i = start + 1; i <= end; i++)
        result = result sep array[i]
    return result
@}
@c endfile
@end example

An optional additional argument is the separator to use when joining the
strings back together.  If the caller supplies a nonempty value,
@code{join()} uses it; if it is not supplied, it has a null
value.  In this case, @code{join()} uses a single space as a default
separator for the strings.  If the value is equal to @code{SUBSEP},
then @code{join()} joins the strings with no separator between them.
@code{SUBSEP} serves as a ``magic'' value to indicate that there should
be no separation between the component strings.@footnote{It would
be nice if @command{awk} had an assignment operator for concatenation.
The lack of an explicit operator for concatenation makes string operations
more difficult than they really need to be.}

@node Getlocaltime Function
@subsection Managing the Time of Day

@cindex libraries of @command{awk} functions, managing, time
@cindex functions, library, managing time
@cindex timestamps, formatted
@cindex time, managing
The @code{systime()} and @code{strftime()} functions described in
@ref{Time Functions}
provide the minimum functionality necessary for dealing with the time of day
in human-readable form.  Although @code{strftime()} is extensive, the control
formats are not necessarily easy to remember or intuitively obvious when
reading a program.

The following function, @code{getlocaltime()}, populates a user-supplied array
with preformatted time information.  It returns a string with the current
time formatted in the same way as the @command{date} utility:

@cindex @code{getlocaltime()} user-defined function
@example
@c file eg/lib/gettime.awk
# getlocaltime.awk --- get the time of day in a usable format
@c endfile
@ignore
@c file eg/lib/gettime.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain, May 1993
#
@c endfile
@end ignore
@c file eg/lib/gettime.awk

# Returns a string in the format of output of date(1)
# Populates the array argument time with individual values:
#    time["second"]       -- seconds (0 - 59)
#    time["minute"]       -- minutes (0 - 59)
#    time["hour"]         -- hours (0 - 23)
#    time["althour"]      -- hours (0 - 12)
#    time["monthday"]     -- day of month (1 - 31)
#    time["month"]        -- month of year (1 - 12)
#    time["monthname"]    -- name of the month
#    time["shortmonth"]   -- short name of the month
#    time["year"]         -- year modulo 100 (0 - 99)
#    time["fullyear"]     -- full year
#    time["weekday"]      -- day of week (Sunday = 0)
#    time["altweekday"]   -- day of week (Monday = 0)
#    time["dayname"]      -- name of weekday
#    time["shortdayname"] -- short name of weekday
#    time["yearday"]      -- day of year (0 - 365)
#    time["timezone"]     -- abbreviation of timezone name
#    time["ampm"]         -- AM or PM designation
#    time["weeknum"]      -- week number, Sunday first day
#    time["altweeknum"]   -- week number, Monday first day

function getlocaltime(time,    ret, now, i)
@{
    # get time once, avoids unnecessary system calls
    now = systime()

    # return date(1)-style output
    ret = strftime("%a %b %e %H:%M:%S %Z %Y", now)

    # clear out target array
    delete time

    # fill in values, force numeric values to be
    # numeric by adding 0
    time["second"]       = strftime("%S", now) + 0
    time["minute"]       = strftime("%M", now) + 0
    time["hour"]         = strftime("%H", now) + 0
    time["althour"]      = strftime("%I", now) + 0
    time["monthday"]     = strftime("%d", now) + 0
    time["month"]        = strftime("%m", now) + 0
    time["monthname"]    = strftime("%B", now)
    time["shortmonth"]   = strftime("%b", now)
    time["year"]         = strftime("%y", now) + 0
    time["fullyear"]     = strftime("%Y", now) + 0
    time["weekday"]      = strftime("%w", now) + 0
    time["altweekday"]   = strftime("%u", now) + 0
    time["dayname"]      = strftime("%A", now)
    time["shortdayname"] = strftime("%a", now)
    time["yearday"]      = strftime("%j", now) + 0
    time["timezone"]     = strftime("%Z", now)
    time["ampm"]         = strftime("%p", now)
    time["weeknum"]      = strftime("%U", now) + 0
    time["altweeknum"]   = strftime("%W", now) + 0

    return ret
@}
@c endfile
@end example

The string indices are easier to use and read than the various formats
required by @code{strftime()}.  The @code{alarm} program presented in
@ref{Alarm Program}
uses this function.
A more general design for the @code{getlocaltime()} function would have
allowed the user to supply an optional timestamp value to use instead
of the current time.

@node Readfile Function
@subsection Reading a Whole File at Once

Often, it is convenient to have the entire contents of a file available
in memory as a single string. A straightforward but naive way to
do that might be as follows:

@example
function readfile1(file,    tmp, contents)
@{
    if ((getline tmp < file) < 0)
        return

    contents = tmp RT
    while ((getline tmp < file) > 0)
        contents = contents tmp RT

    close(file)
    return contents
@}
@end example

This function reads from @code{file} one record at a time, building
up the full contents of the file in the local variable @code{contents}.
It works, but is not necessarily efficient.

The following function, based on a suggestion by Denis Shirokov,
reads the entire contents of the named file in one shot:

@cindex @code{readfile()} user-defined function
@example
@c file eg/lib/readfile.awk
# readfile.awk --- read an entire file at once
@c endfile
@ignore
@c file eg/lib/readfile.awk
#
# Original idea by Denis Shirokov, cosmogen@@gmail.com, April 2013
#
@c endfile
@end ignore
@c file eg/lib/readfile.awk

function readfile(file,     tmp, save_rs)
@{
    save_rs = RS
    RS = "^$"
    getline tmp < file
    close(file)
    RS = save_rs

    return tmp
@}
@c endfile
@end example

It works by setting @code{RS} to @samp{^$}, a regular expression that
will never match if the file has contents.  @command{gawk} reads data from
the file into @code{tmp}, attempting to match @code{RS}.  The match fails
after each read, but fails quickly, such that @command{gawk} fills
@code{tmp} with the entire contents of the file.
(@xref{Records} for information on @code{RT} and @code{RS}.)

In the case that @code{file} is empty, the return value is the null
string.  Thus, calling code may use something like:

@example
contents = readfile("/some/path")
if (length(contents) == 0)
    # file was empty @dots{}
@end example

This tests the result to see if it is empty or not. An equivalent
test would be @samp{@w{contents == ""}}.

@xref{Extension Sample Readfile} for an extension function that
also reads an entire file into memory.

@node Shell Quoting
@subsection Quoting Strings to Pass to the Shell

@c included by permission
@ignore
Date: Sun, 27 Jul 2014 17:16:16 -0700
Message-ID: <CAKuGj+iCF_obaCLDUX60aSAgbfocFVtguG39GyeoNxTFby5sqQ@mail.gmail.com>
Subject: Useful awk function
From: Mike Brennan <mike@madronabluff.com>
To: Arnold Robbins <arnold@skeeve.com>
@end ignore

Michael Brennan offers the following programming pattern,
which he uses frequently:

@example
#! /bin/sh

awkp='
   @dots{}
   '

@var{input_program} | awk "$awkp" | /bin/sh
@end example

For example, a program of his named @command{flac-edit} has this form:

@example
$ @kbd{flac-edit -song="Whoope! That's Great" file.flac}
@end example

It generates the following output, which is to be piped to
the shell (@file{/bin/sh}):

@example
chmod +w file.flac
metaflac --remove-tag=TITLE file.flac
LANG=en_US.88591 metaflac --set-tag=TITLE='Whoope! That'"'"'s Great' file.flac
chmod -w file.flac
@end example

Note the need for shell quoting.  The function @code{shell_quote()}
does it.  @code{SINGLE} is the one-character string @code{"'"} and
@code{QSINGLE} is the three-character string @code{"\"'\""}:

@example
@c file eg/lib/shellquote.awk
# shell_quote --- quote an argument for passing to the shell
@c endfile
@ignore
@c file eg/lib/shellquote.awk
#
# Michael Brennan
# brennan@@madronabluff.com
# September 2014
@c endfile
@end ignore
@c file eg/lib/shellquote.awk

function shell_quote(s,             # parameter
    SINGLE, QSINGLE, i, X, n, ret)  # locals
@{
    if (s == "")
        return "\"\""

    SINGLE = "\x27"  # single quote
    QSINGLE = "\"\x27\""
    n = split(s, X, SINGLE)

    ret = SINGLE X[1] SINGLE
    for (i = 2; i <= n; i++)
        ret = ret QSINGLE SINGLE X[i] SINGLE

    return ret
@}
@c endfile
@end example

@node Data File Management
@section @value{DDF} Management

@cindex files, managing
@cindex libraries of @command{awk} functions, managing, data files
@cindex functions, library, managing data files
This @value{SECTION} presents functions that are useful for managing
command-line @value{DF}s.

@menu
* Filetrans Function::          A function for handling data file transitions.
* Rewind Function::             A function for rereading the current file.
* File Checking::               Checking that data files are readable.
* Empty Files::                 Checking for zero-length files.
* Ignoring Assigns::            Treating assignments as file names.
@end menu

@node Filetrans Function
@subsection Noting @value{DDF} Boundaries

@cindex files, managing, data file boundaries
@cindex files, initialization and cleanup
The @code{BEGIN} and @code{END} rules are each executed exactly once, at
the beginning and end of your @command{awk} program, respectively
(@pxref{BEGIN/END}).
We (the @command{gawk} authors) once had a user who mistakenly thought that the
@code{BEGIN} rules were executed at the beginning of each @value{DF} and the
@code{END} rules were executed at the end of each @value{DF}.

When informed
that this was not the case, the user requested that we add new special
patterns to @command{gawk}, named @code{BEGIN_FILE} and @code{END_FILE}, that
would have the desired behavior.  He even supplied us the code to do so.

Adding these special patterns to @command{gawk} wasn't necessary;
the job can be done cleanly in @command{awk} itself, as illustrated
by the following library program.
It arranges to call two user-supplied functions, @code{beginfile()} and
@code{endfile()}, at the beginning and end of each @value{DF}.
Besides solving the problem in only nine(!) lines of code, it does so
@emph{portably}; this works with any implementation of @command{awk}:

@example
# transfile.awk
#
# Give the user a hook for filename transitions
#
# The user must supply functions beginfile() and endfile()
# that each take the name of the file being started or
# finished, respectively.
@c #
@c # Arnold Robbins, arnold@@skeeve.com, Public Domain
@c # January 1992

FILENAME != _oldfilename @{
    if (_oldfilename != "")
        endfile(_oldfilename)
    _oldfilename = FILENAME
    beginfile(FILENAME)
@}

END @{ endfile(FILENAME) @}
@end example

This file must be loaded before the user's ``main'' program, so that the
rule it supplies is executed first.

This rule relies on @command{awk}'s @code{FILENAME} variable, which
automatically changes for each new @value{DF}.  The current @value{FN} is
saved in a private variable, @code{_oldfilename}.  If @code{FILENAME} does
not equal @code{_oldfilename}, then a new @value{DF} is being processed and
it is necessary to call @code{endfile()} for the old file.  Because
@code{endfile()} should only be called if a file has been processed, the
program first checks to make sure that @code{_oldfilename} is not the null
string.  The program then assigns the current @value{FN} to
@code{_oldfilename} and calls @code{beginfile()} for the file.
Because, like all @command{awk} variables, @code{_oldfilename} is
initialized to the null string, this rule executes correctly even for the
first @value{DF}.

The program also supplies an @code{END} rule to do the final processing for
the last file.  Because this @code{END} rule comes before any @code{END} rules
supplied in the ``main'' program, @code{endfile()} is called first.  Once
again, the value of multiple @code{BEGIN} and @code{END} rules should be clear.

@cindex @code{beginfile()} user-defined function
@cindex @code{endfile()} user-defined function
If the same @value{DF} occurs twice in a row on the command line, then
@code{endfile()} and @code{beginfile()} are not executed at the end of the
first pass and at the beginning of the second pass.
The following version solves the problem:

@example
@c file eg/lib/ftrans.awk
# ftrans.awk --- handle datafile transitions
#
# user supplies beginfile() and endfile() functions
@c endfile
@ignore
@c file eg/lib/ftrans.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# November 1992
@c endfile
@end ignore
@c file eg/lib/ftrans.awk

FNR == 1 @{
    if (_filename_ != "")
        endfile(_filename_)
    _filename_ = FILENAME
    beginfile(FILENAME)
@}

END @{ endfile(_filename_) @}
@c endfile
@end example

@ref{Wc Program}
shows how this library function can be used and
how it simplifies writing the main program.

@sidebar So Why Does @command{gawk} Have @code{BEGINFILE} and @code{ENDFILE}?

You are probably wondering, if @code{beginfile()} and @code{endfile()}
functions can do the job, why does @command{gawk} have
@code{BEGINFILE} and @code{ENDFILE} patterns?

Good question.  Normally, if @command{awk} cannot open a file, this
causes an immediate fatal error.  In this case, there is no way for a
user-defined function to deal with the problem, as the mechanism for
calling it relies on the file being open and at the first record.  Thus,
the main reason for @code{BEGINFILE} is to give you a ``hook'' to catch
files that cannot be processed.  @code{ENDFILE} exists for symmetry,
and because it provides an easy way to do per-file cleanup processing.
For more information, refer to @ref{BEGINFILE/ENDFILE}.
@end sidebar

@node Rewind Function
@subsection Rereading the Current File

@cindex files, reading
Another request for a new built-in function was for a
function that would make it possible to reread the current file.
The requesting user didn't want to have to use @code{getline}
(@pxref{Getline})
inside a loop.

However, as long as you are not in the @code{END} rule, it is
quite easy to arrange to immediately close the current input file
and then start over with it from the top.
For lack of a better name, we'll call the function @code{rewind()}:

@cindex @code{rewind()} user-defined function
@example
@c file eg/lib/rewind.awk
# rewind.awk --- rewind the current file and start over
@c endfile
@ignore
@c file eg/lib/rewind.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# September 2000
@c endfile
@end ignore
@c file eg/lib/rewind.awk

function rewind(    i)
@{
    # shift remaining arguments up
    for (i = ARGC; i > ARGIND; i--)
        ARGV[i] = ARGV[i-1]

    # make sure gawk knows to keep going
    ARGC++

    # make current file next to get done
    ARGV[ARGIND+1] = FILENAME

    # do it
    nextfile
@}
@c endfile
@end example

The @code{rewind()} function relies on the @code{ARGIND} variable
(@pxref{Auto-set}), which is specific to @command{gawk}.  It also
relies on the @code{nextfile} keyword (@pxref{Nextfile Statement}).
Because of this, you should not call it from an @code{ENDFILE} rule.
(This isn't necessary anyway, because @command{gawk} goes to the next
file as soon as an @code{ENDFILE} rule finishes!)

You need to be careful calling @code{rewind()}.  You can end up
causing infinite recursion if you don't pay attention. Here is an
example use:

@example
$ @kbd{cat data}
@print{} a
@print{} b
@print{} c
@print{} d
@print{} e

$ cat @kbd{test.awk}
@print{} FNR == 3 && ! rewound @{
@print{}    rewound = 1
@print{}    rewind()
@print{} @}
@print{} 
@print{} @{ print FILENAME, FNR, $0 @}

$ @kbd{gawk -f rewind.awk -f test.awk data }
@print{} data 1 a
@print{} data 2 b
@print{} data 1 a
@print{} data 2 b
@print{} data 3 c
@group
@print{} data 4 d
@print{} data 5 e
@end group
@end example

@node File Checking
@subsection Checking for Readable @value{DDF}s

@cindex troubleshooting, readable data files
@cindex readable data files@comma{} checking
@cindex files, skipping
Normally, if you give @command{awk} a @value{DF} that isn't readable,
it stops with a fatal error.  There are times when you might want to
just ignore such files and keep going.@footnote{The @code{BEGINFILE}
special pattern (@pxref{BEGINFILE/ENDFILE}) provides an alternative
mechanism for dealing with files that can't be opened.  However, the
code here provides a portable solution.} You can do this by prepending
the following program to your @command{awk} program:

@cindex @code{readable.awk} program
@example
@c file eg/lib/readable.awk
# readable.awk --- library file to skip over unreadable files
@c endfile
@ignore
@c file eg/lib/readable.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# October 2000
# December 2010
@c endfile
@end ignore
@c file eg/lib/readable.awk

BEGIN @{
    for (i = 1; i < ARGC; i++) @{
        if (ARGV[i] ~ /^[a-zA-Z_][a-zA-Z0-9_]*=.*/ \
            || ARGV[i] == "-" || ARGV[i] == "/dev/stdin")
            continue    # assignment or standard input
        else if ((getline junk < ARGV[i]) < 0) # unreadable
            delete ARGV[i]
        else
            close(ARGV[i])
    @}
@}
@c endfile
@end example

@cindex troubleshooting, @code{getline} function
This works, because the @code{getline} won't be fatal.
Removing the element from @code{ARGV} with @code{delete}
skips the file (because it's no longer in the list).
See also @ref{ARGC and ARGV}.

Because @command{awk} variable names only allow the English letters,
the regular expression check purposely does not use character classes
such as @samp{[:alpha:]} and @samp{[:alnum:]}
(@pxref{Bracket Expressions}).

@node Empty Files
@subsection Checking for Zero-Length Files

All known @command{awk} implementations silently skip over zero-length files.
This is a by-product of @command{awk}'s implicit
read-a-record-and-match-against-the-rules loop: when @command{awk}
tries to read a record from an empty file, it immediately receives an
end-of-file indication, closes the file, and proceeds on to the next
command-line @value{DF}, @emph{without} executing any user-level
@command{awk} program code.

Using @command{gawk}'s @code{ARGIND} variable
(@pxref{Built-in Variables}), it is possible to detect when an empty
@value{DF} has been skipped.  Similar to the library file presented
in @ref{Filetrans Function}, the following library file calls a function named
@code{zerofile()} that the user must provide.  The arguments passed are
the @value{FN} and the position in @code{ARGV} where it was found:

@cindex @code{zerofile.awk} program
@example
@c file eg/lib/zerofile.awk
# zerofile.awk --- library file to process empty input files
@c endfile
@ignore
@c file eg/lib/zerofile.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# June 2003
@c endfile
@end ignore
@c file eg/lib/zerofile.awk

BEGIN @{ Argind = 0 @}

ARGIND > Argind + 1 @{
    for (Argind++; Argind < ARGIND; Argind++)
        zerofile(ARGV[Argind], Argind)
@}

ARGIND != Argind @{ Argind = ARGIND @}

END @{
    if (ARGIND > Argind)
        for (Argind++; Argind <= ARGIND; Argind++)
            zerofile(ARGV[Argind], Argind)
@}
@c endfile
@end example

The user-level variable @code{Argind} allows the @command{awk} program
to track its progress through @code{ARGV}.  Whenever the program detects
that @code{ARGIND} is greater than @samp{Argind + 1}, it means that one or
more empty files were skipped.  The action then calls @code{zerofile()} for
each such file, incrementing @code{Argind} along the way.

The @samp{Argind != ARGIND} rule simply keeps @code{Argind} up to date
in the normal case.

Finally, the @code{END} rule catches the case of any empty files at
the end of the command-line arguments.  Note that the test in the
condition of the @code{for} loop uses the @samp{<=} operator,
not @samp{<}.

@node Ignoring Assigns
@subsection Treating Assignments as @value{FFN}s

@cindex assignments as filenames
@cindex filenames, assignments as
Occasionally, you might not want @command{awk} to process command-line
variable assignments
(@pxref{Assignment Options}).
In particular, if you have a @value{FN} that contains an @samp{=} character,
@command{awk} treats the @value{FN} as an assignment and does not process it.

Some users have suggested an additional command-line option for @command{gawk}
to disable command-line assignments.  However, some simple programming with
a library file does the trick:

@cindex @code{noassign.awk} program
@example
@c file eg/lib/noassign.awk
# noassign.awk --- library file to avoid the need for a
# special option that disables command-line assignments
@c endfile
@ignore
@c file eg/lib/noassign.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# October 1999
@c endfile
@end ignore
@c file eg/lib/noassign.awk

function disable_assigns(argc, argv,    i)
@{
    for (i = 1; i < argc; i++)
        if (argv[i] ~ /^[a-zA-Z_][a-zA-Z0-9_]*=.*/)
            argv[i] = ("./" argv[i])
@}

BEGIN @{
    if (No_command_assign)
        disable_assigns(ARGC, ARGV)
@}
@c endfile
@end example

You then run your program this way:

@example
awk -v No_command_assign=1 -f noassign.awk -f yourprog.awk *
@end example

The function works by looping through the arguments.
It prepends @samp{./} to
any argument that matches the form
of a variable assignment, turning that argument into a @value{FN}.

The use of @code{No_command_assign} allows you to disable command-line
assignments at invocation time, by giving the variable a true value.
When not set, it is initially zero (i.e., false), so the command-line arguments
are left alone.

@node Getopt Function
@section Processing Command-Line Options

@cindex libraries of @command{awk} functions, command-line options
@cindex functions, library, command-line options
@cindex command-line options, processing
@cindex options, command-line, processing
@cindex functions, library, C library
@cindex arguments, processing
Most utilities on POSIX-compatible systems take options on
the command line that can be used to change the way a program behaves.
@command{awk} is an example of such a program
(@pxref{Options}).
Often, options take @dfn{arguments} (i.e., data that the program needs to
correctly obey the command-line option).  For example, @command{awk}'s
@option{-F} option requires a string to use as the field separator.
The first occurrence on the command line of either @option{--} or a
string that does not begin with @samp{-} ends the options.

@cindex @code{getopt()} function (C library)
Modern Unix systems provide a C function named @code{getopt()} for processing
command-line arguments.  The programmer provides a string describing the
one-letter options. If an option requires an argument, it is followed in the
string with a colon.  @code{getopt()} is also passed the
count and values of the command-line arguments and is called in a loop.
@code{getopt()} processes the command-line arguments for option letters.
Each time around the loop, it returns a single character representing the
next option letter that it finds, or @samp{?} if it finds an invalid option.
When it returns @minus{}1, there are no options left on the command line.

When using @code{getopt()}, options that do not take arguments can be
grouped together.  Furthermore, options that take arguments require that the
argument be present.  The argument can immediately follow the option letter,
or it can be a separate command-line argument.

Given a hypothetical program that takes
three command-line options, @option{-a}, @option{-b}, and @option{-c}, where
@option{-b} requires an argument, all of the following are valid ways of
invoking the program:

@example
prog -a -b foo -c data1 data2 data3
prog -ac -bfoo -- data1 data2 data3
prog -acbfoo data1 data2 data3
@end example

Notice that when the argument is grouped with its option, the rest of
the argument is considered to be the option's argument.
In this example, @option{-acbfoo} indicates that all of the
@option{-a}, @option{-b}, and @option{-c} options were supplied,
and that @samp{foo} is the argument to the @option{-b} option.

@code{getopt()} provides four external variables that the programmer can use:

@table @code
@item optind
The index in the argument value array (@code{argv}) where the first
nonoption command-line argument can be found.

@item optarg
The string value of the argument to an option.

@item opterr
Usually @code{getopt()} prints an error message when it finds an invalid
option.  Setting @code{opterr} to zero disables this feature.  (An
application might want to print its own error message.)

@item optopt
The letter representing the command-line option.
@end table

The following C fragment shows how @code{getopt()} might process command-line
arguments for @command{awk}:

@example
int
main(int argc, char *argv[])
@{
    @dots{}
    /* print our own message */
    opterr = 0;
    while ((c = getopt(argc, argv, "v:f:F:W:")) != -1) @{
        switch (c) @{
        case 'f':    /* file */
            @dots{}
            break;
        case 'F':    /* field separator */
            @dots{}
            break;
        case 'v':    /* variable assignment */
            @dots{}
            break;
        case 'W':    /* extension */
            @dots{}
            break;
        case '?':
        default:
            usage();
            break;
        @}
    @}
    @dots{}
@}
@end example

As a side point, @command{gawk} actually uses the GNU @code{getopt_long()}
function to process both normal and GNU-style long options
(@pxref{Options}).

The abstraction provided by @code{getopt()} is very useful and is quite
handy in @command{awk} programs as well.  Following is an @command{awk}
version of @code{getopt()}.  This function highlights one of the
greatest weaknesses in @command{awk}, which is that it is very poor at
manipulating single characters.  Repeated calls to @code{substr()} are
necessary for accessing individual characters
(@pxref{String Functions}).@footnote{This
function was written before @command{gawk} acquired the ability to
split strings into single characters using @code{""} as the separator.
We have left it alone, as using @code{substr()} is more portable.}

The discussion that follows walks through the code a bit at a time:

@cindex @code{getopt()} user-defined function
@example
@c file eg/lib/getopt.awk
# getopt.awk --- Do C library getopt(3) function in awk
@c endfile
@ignore
@c file eg/lib/getopt.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
#
# Initial version: March, 1991
# Revised: May, 1993
@c endfile
@end ignore
@c file eg/lib/getopt.awk

# External variables:
#    Optind -- index in ARGV of first nonoption argument
#    Optarg -- string value of argument to current option
#    Opterr -- if nonzero, print our own diagnostic
#    Optopt -- current option letter

# Returns:
#    -1     at end of options
#    "?"    for unrecognized option
#    <c>    a character representing the current option

# Private Data:
#    _opti  -- index in multiflag option, e.g., -abc
@c endfile
@end example

The function starts out with comments presenting
a list of the global variables it uses,
what the return values are, what they mean, and any global variables that
are ``private'' to this library function.  Such documentation is essential
for any program, and particularly for library functions.

The @code{getopt()} function first checks that it was indeed called with
a string of options (the @code{options} parameter).  If @code{options}
has a zero length, @code{getopt()} immediately returns @minus{}1:

@cindex @code{getopt()} user-defined function
@example
@c file eg/lib/getopt.awk
function getopt(argc, argv, options,    thisopt, i)
@{
    if (length(options) == 0)    # no options given
        return -1

@group
    if (argv[Optind] == "--") @{  # all done
        Optind++
        _opti = 0
        return -1
@end group
    @} else if (argv[Optind] !~ /^-[^:[:space:]]/) @{
        _opti = 0
        return -1
    @}
@c endfile
@end example

The next thing to check for is the end of the options.  A @option{--}
ends the command-line options, as does any command-line argument that
does not begin with a @samp{-}.  @code{Optind} is used to step through
the array of command-line arguments; it retains its value across calls
to @code{getopt()}, because it is a global variable.

The regular expression that is used, @code{@w{/^-[^:[:space:]/}},
checks for a @samp{-} followed by anything
that is not whitespace and not a colon.
If the current command-line argument does not match this pattern,
it is not an option, and it ends option processing. Continuing on:

@example
@c file eg/lib/getopt.awk
    if (_opti == 0)
        _opti = 2
    thisopt = substr(argv[Optind], _opti, 1)
    Optopt = thisopt
    i = index(options, thisopt)
    if (i == 0) @{
        if (Opterr)
            printf("%c -- invalid option\n", thisopt) > "/dev/stderr"
        if (_opti >= length(argv[Optind])) @{
            Optind++
            _opti = 0
        @} else
            _opti++
        return "?"
    @}
@c endfile
@end example

The @code{_opti} variable tracks the position in the current command-line
argument (@code{argv[Optind]}).  If multiple options are
grouped together with one @samp{-} (e.g., @option{-abx}), it is necessary
to return them to the user one at a time.

If @code{_opti} is equal to zero, it is set to two, which is the index in
the string of the next character to look at (we skip the @samp{-}, which
is at position one).  The variable @code{thisopt} holds the character,
obtained with @code{substr()}.  It is saved in @code{Optopt} for the main
program to use.

If @code{thisopt} is not in the @code{options} string, then it is an
invalid option.  If @code{Opterr} is nonzero, @code{getopt()} prints an error
message on the standard error that is similar to the message from the C
version of @code{getopt()}.

Because the option is invalid, it is necessary to skip it and move on to the
next option character.  If @code{_opti} is greater than or equal to the
length of the current command-line argument, it is necessary to move on
to the next argument, so @code{Optind} is incremented and @code{_opti} is reset
to zero. Otherwise, @code{Optind} is left alone and @code{_opti} is merely
incremented.

In any case, because the option is invalid, @code{getopt()} returns @code{"?"}.
The main program can examine @code{Optopt} if it needs to know what the
invalid option letter actually is. Continuing on:

@example
@c file eg/lib/getopt.awk
    if (substr(options, i + 1, 1) == ":") @{
        # get option argument
        if (length(substr(argv[Optind], _opti + 1)) > 0)
            Optarg = substr(argv[Optind], _opti + 1)
        else
            Optarg = argv[++Optind]
        _opti = 0
    @} else
        Optarg = ""
@c endfile
@end example

If the option requires an argument, the option letter is followed by a colon
in the @code{options} string.  If there are remaining characters in the
current command-line argument (@code{argv[Optind]}), then the rest of that
string is assigned to @code{Optarg}.  Otherwise, the next command-line
argument is used (@samp{-xFOO} versus @samp{@w{-x FOO}}). In either case,
@code{_opti} is reset to zero, because there are no more characters left to
examine in the current command-line argument. Continuing:

@example
@c file eg/lib/getopt.awk
    if (_opti == 0 || _opti >= length(argv[Optind])) @{
        Optind++
        _opti = 0
    @} else
        _opti++
    return thisopt
@}
@c endfile
@end example

Finally, if @code{_opti} is either zero or greater than the length of the
current command-line argument, it means this element in @code{argv} is
through being processed, so @code{Optind} is incremented to point to the
next element in @code{argv}.  If neither condition is true, then only
@code{_opti} is incremented, so that the next option letter can be processed
on the next call to @code{getopt()}.

The @code{BEGIN} rule initializes both @code{Opterr} and @code{Optind} to one.
@code{Opterr} is set to one, because the default behavior is for @code{getopt()}
to print a diagnostic message upon seeing an invalid option.  @code{Optind}
is set to one, because there's no reason to look at the program name, which is
in @code{ARGV[0]}:

@example
@c file eg/lib/getopt.awk
BEGIN @{
    Opterr = 1    # default is to diagnose
    Optind = 1    # skip ARGV[0]

    # test program
    if (_getopt_test) @{
        while ((_go_c = getopt(ARGC, ARGV, "ab:cd")) != -1)
            printf("c = <%c>, Optarg = <%s>\n",
                                       _go_c, Optarg)
        printf("non-option arguments:\n")
        for (; Optind < ARGC; Optind++)
            printf("\tARGV[%d] = <%s>\n",
                                    Optind, ARGV[Optind])
    @}
@}
@c endfile
@end example

The rest of the @code{BEGIN} rule is a simple test program.  Here are the
results of two sample runs of the test program:

@example
$ @kbd{awk -f getopt.awk -v _getopt_test=1 -- -a -cbARG bax -x}
@print{} c = <a>, Optarg = <>
@print{} c = <c>, Optarg = <>
@print{} c = <b>, Optarg = <ARG>
@print{} non-option arguments:
@print{}         ARGV[3] = <bax>
@print{}         ARGV[4] = <-x>

$ @kbd{awk -f getopt.awk -v _getopt_test=1 -- -a -x -- xyz abc}
@print{} c = <a>, Optarg = <>
@error{} x -- invalid option
@print{} c = <?>, Optarg = <>
@print{} non-option arguments:
@print{}         ARGV[4] = <xyz>
@print{}         ARGV[5] = <abc>
@end example

In both runs, the first @option{--} terminates the arguments to
@command{awk}, so that it does not try to interpret the @option{-a},
etc., as its own options.

@quotation NOTE
After @code{getopt()} is through,
user-level code must clear out all the elements of @code{ARGV} from 1
to @code{Optind}, so that @command{awk} does not try to process the
command-line options as @value{FN}s.
@end quotation

Using @samp{#!} with the @option{-E} option may help avoid
conflicts between your program's options and @command{gawk}'s options,
as @option{-E} causes @command{gawk} to abandon processing of
further options
(@pxref{Executable Scripts} and
@ifnotdocbook
@pxref{Options}).
@end ifnotdocbook
@ifdocbook
@ref{Options}).
@end ifdocbook

Several of the sample programs presented in
@ref{Sample Programs},
use @code{getopt()} to process their arguments.

@node Passwd Functions
@section Reading the User Database

@cindex libraries of @command{awk} functions, user database, reading
@cindex functions, library, user database@comma{} reading
@cindex user database@comma{} reading
@cindex database, users@comma{} reading
@cindex @code{PROCINFO} array
The @code{PROCINFO} array
(@pxref{Built-in Variables})
provides access to the current user's real and effective user and group ID
numbers, and, if available, the user's supplementary group set.
However, because these are numbers, they do not provide very useful
information to the average user.  There needs to be some way to find the
user information associated with the user and group ID numbers.  This
@value{SECTION} presents a suite of functions for retrieving information from the
user database.  @xref{Group Functions}
for a similar suite that retrieves information from the group database.

@cindex @code{getpwent()} function (C library)
@cindex @code{getpwent()} user-defined function
@cindex users, information about, retrieving
@cindex login information
@cindex account information
@cindex password file
@cindex files, password
The POSIX standard does not define the file where user information is
kept.  Instead, it provides the @code{<pwd.h>} header file
and several C language subroutines for obtaining user information.
The primary function is @code{getpwent()}, for ``get password entry.''
The ``password'' comes from the original user database file,
@file{/etc/passwd}, which stores user information along with the
encrypted passwords (hence the name).

@cindex @command{pwcat} program
Although an @command{awk} program could simply read @file{/etc/passwd}
directly, this file may not contain complete information about the
system's set of users.@footnote{It is often the case that password
information is stored in a network database.} To be sure you are able to
produce a readable and complete version of the user database, it is necessary
to write a small C program that calls @code{getpwent()}.  @code{getpwent()}
is defined as returning a pointer to a @code{struct passwd}.  Each time it
is called, it returns the next entry in the database.  When there are
no more entries, it returns @code{NULL}, the null pointer.  When this
happens, the C program should call @code{endpwent()} to close the database.
Following is @command{pwcat}, a C program that ``cats'' the password database:

@example
@c file eg/lib/pwcat.c
/*
 * pwcat.c
 *
 * Generate a printable version of the password database.
 */
@c endfile
@ignore
@c file eg/lib/pwcat.c
/*
 * Arnold Robbins, arnold@@skeeve.com, May 1993
 * Public Domain
 * December 2010, move to ANSI C definition for main().
 */

#if HAVE_CONFIG_H
#include <config.h>
#endif

@c endfile
@end ignore
@c file eg/lib/pwcat.c
#include <stdio.h>
#include <pwd.h>

@c endfile
@ignore
@c file eg/lib/pwcat.c
#if defined (STDC_HEADERS)
#include <stdlib.h>
#endif

@c endfile
@end ignore
@c file eg/lib/pwcat.c
int
main(int argc, char **argv)
@{
    struct passwd *p;

    while ((p = getpwent()) != NULL)
@c endfile
@ignore
@c file eg/lib/pwcat.c
#ifdef HAVE_STRUCT_PASSWD_PW_PASSWD
@c endfile
@end ignore
@c file eg/lib/pwcat.c
        printf("%s:%s:%ld:%ld:%s:%s:%s\n",
            p->pw_name, p->pw_passwd, (long) p->pw_uid,
            (long) p->pw_gid, p->pw_gecos, p->pw_dir, p->pw_shell);
@c endfile
@ignore
@c file eg/lib/pwcat.c
#else
        printf("%s:*:%ld:%ld:%s:%s\n",
            p->pw_name, (long) p->pw_uid,
            (long) p->pw_gid, p->pw_dir, p->pw_shell);
#endif
@c endfile
@end ignore
@c file eg/lib/pwcat.c

    endpwent();
    return 0;
@}
@c endfile
@end example

If you don't understand C, don't worry about it.
The output from @command{pwcat} is the user database, in the traditional
@file{/etc/passwd} format of colon-separated fields.  The fields are:

@table @asis
@item Login name
The user's login name.

@item Encrypted password
The user's encrypted password.  This may not be available on some systems.

@item User-ID
The user's numeric user ID number.
(On some systems, it's a C @code{long}, and not an @code{int}.  Thus,
we cast it to @code{long} for all cases.)

@item Group-ID
The user's numeric group ID number.
(Similar comments about @code{long} versus @code{int} apply here.)

@item Full name
The user's full name, and perhaps other information associated with the
user.

@item Home directory
The user's login (or ``home'') directory (familiar to shell programmers as
@code{$HOME}).

@item Login shell
The program that is run when the user logs in.  This is usually a
shell, such as Bash.
@end table

A few lines representative of @command{pwcat}'s output are as follows:

@cindex Jacobs, Andrew
@cindex Robbins, Arnold
@cindex Robbins, Miriam
@example
$ @kbd{pwcat}
@print{} root:x:0:1:Operator:/:/bin/sh
@print{} nobody:*:65534:65534::/:
@print{} daemon:*:1:1::/:
@print{} sys:*:2:2::/:/bin/csh
@print{} bin:*:3:3::/bin:
@print{} arnold:xyzzy:2076:10:Arnold Robbins:/home/arnold:/bin/sh
@print{} miriam:yxaay:112:10:Miriam Robbins:/home/miriam:/bin/sh
@print{} andy:abcca2:113:10:Andy Jacobs:/home/andy:/bin/sh
@dots{}
@end example

With that introduction, following is a group of functions for getting user
information.  There are several functions here, corresponding to the C
functions of the same names:

@cindex @code{_pw_init()} user-defined function
@example
@c file eg/lib/passwdawk.in
# passwd.awk --- access password file information
@c endfile
@ignore
@c file eg/lib/passwdawk.in
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised October 2000
# Revised December 2010
@c endfile
@end ignore
@c file eg/lib/passwdawk.in

BEGIN @{
    # tailor this to suit your system
    _pw_awklib = "/usr/local/libexec/awk/"
@}

function _pw_init(    oldfs, oldrs, olddol0, pwcat, using_fw, using_fpat)
@{
    if (_pw_inited)
        return

    oldfs = FS
    oldrs = RS
    olddol0 = $0
    using_fw = (PROCINFO["FS"] == "FIELDWIDTHS")
    using_fpat = (PROCINFO["FS"] == "FPAT")
    FS = ":"
    RS = "\n"

    pwcat = _pw_awklib "pwcat"
    while ((pwcat | getline) > 0) @{
        _pw_byname[$1] = $0
        _pw_byuid[$3] = $0
        _pw_bycount[++_pw_total] = $0
    @}
    close(pwcat)
    _pw_count = 0
    _pw_inited = 1
    FS = oldfs
    if (using_fw)
        FIELDWIDTHS = FIELDWIDTHS
    else if (using_fpat)
        FPAT = FPAT
    RS = oldrs
    $0 = olddol0
@}
@c endfile
@end example

@cindex @code{BEGIN} pattern, @code{pwcat} program
The @code{BEGIN} rule sets a private variable to the directory where
@command{pwcat} is stored.  Because it is used to help out an @command{awk} library
routine, we have chosen to put it in @file{/usr/local/libexec/awk};
however, you might want it to be in a different directory on your system.

The function @code{_pw_init()} fills three copies of the user information
into three associative arrays.  The arrays are indexed by username
(@code{_pw_byname}), by user ID number (@code{_pw_byuid}), and by order of
occurrence (@code{_pw_bycount}).
The variable @code{_pw_inited} is used for efficiency, as @code{_pw_init()}
needs to be called only once.

@cindex @code{PROCINFO} array, testing the field splitting
@cindex @code{getline} command, @code{_pw_init()} function
Because this function uses @code{getline} to read information from
@command{pwcat}, it first saves the values of @code{FS}, @code{RS}, and @code{$0}.
It notes in the variable @code{using_fw} whether field splitting
with @code{FIELDWIDTHS} is in effect or not.
Doing so is necessary, as these functions could be called
from anywhere within a user's program, and the user may have his
or her own way of splitting records and fields.
This makes it possible to restore the correct
field-splitting mechanism later.  The test can only be true for
@command{gawk}.  It is false if using @code{FS} or @code{FPAT},
or on some other @command{awk} implementation.

The code that checks for using @code{FPAT}, using @code{using_fpat}
and @code{PROCINFO["FS"]}, is similar.

The main part of the function uses a loop to read database lines, split
the lines into fields, and then store the lines into each array as necessary.
When the loop is done, @code{@w{_pw_init()}} cleans up by closing the pipeline,
setting @code{@w{_pw_inited}} to one, and restoring @code{FS}
(and @code{FIELDWIDTHS} or @code{FPAT}
if necessary), @code{RS}, and @code{$0}.
The use of @code{@w{_pw_count}} is explained shortly.

@cindex @code{getpwnam()} function (C library)
The @code{getpwnam()} function takes a username as a string argument. If that
user is in the database, it returns the appropriate line. Otherwise, it
relies on the array reference to a nonexistent
element to create the element with the null string as its value:

@cindex @code{getpwnam()} user-defined function
@example
@group
@c file eg/lib/passwdawk.in
function getpwnam(name)
@{
    _pw_init()
    return _pw_byname[name]
@}
@c endfile
@end group
@end example

@cindex @code{getpwuid()} function (C library)
Similarly, the @code{getpwuid()} function takes a user ID number
argument. If that user number is in the database, it returns the
appropriate line. Otherwise, it returns the null string:

@cindex @code{getpwuid()} user-defined function
@example
@c file eg/lib/passwdawk.in
function getpwuid(uid)
@{
    _pw_init()
    return _pw_byuid[uid]
@}
@c endfile
@end example

@cindex @code{getpwent()} function (C library)
The @code{getpwent()} function simply steps through the database, one entry at
a time.  It uses @code{_pw_count} to track its current position in the
@code{_pw_bycount} array:

@cindex @code{getpwent()} user-defined function
@example
@c file eg/lib/passwdawk.in
function getpwent()
@{
    _pw_init()
    if (_pw_count < _pw_total)
        return _pw_bycount[++_pw_count]
    return ""
@}
@c endfile
@end example

@cindex @code{endpwent()} function (C library)
The @code{@w{endpwent()}} function resets @code{@w{_pw_count}} to zero, so that
subsequent calls to @code{getpwent()} start over again:

@cindex @code{endpwent()} user-defined function
@example
@c file eg/lib/passwdawk.in
function endpwent()
@{
    _pw_count = 0
@}
@c endfile
@end example

A conscious design decision in this suite is that each subroutine calls
@code{@w{_pw_init()}} to initialize the database arrays.
The overhead of running
a separate process to generate the user database, and the I/O to scan it,
are only incurred if the user's main program actually calls one of these
functions.  If this library file is loaded along with a user's program, but
none of the routines are ever called, then there is no extra runtime overhead.
(The alternative is move the body of @code{@w{_pw_init()}} into a
@code{BEGIN} rule, which always runs @command{pwcat}.  This simplifies the
code but runs an extra process that may never be needed.)

In turn, calling @code{_pw_init()} is not too expensive, because the
@code{_pw_inited} variable keeps the program from reading the data more than
once.  If you are worried about squeezing every last cycle out of your
@command{awk} program, the check of @code{_pw_inited} could be moved out of
@code{_pw_init()} and duplicated in all the other functions.  In practice,
this is not necessary, as most @command{awk} programs are I/O-bound,
and such a change would clutter up the code.

The @command{id} program in @ref{Id Program}
uses these functions.

@node Group Functions
@section Reading the Group Database

@cindex libraries of @command{awk} functions, group database, reading
@cindex functions, library, group database@comma{} reading
@cindex group database, reading
@cindex database, group, reading
@cindex @code{PROCINFO} array, and group membership
@cindex @code{getgrent()} function (C library)
@cindex @code{getgrent()} user-defined function
@cindex groups@comma{} information about
@cindex account information
@cindex group file
@cindex files, group
Much of the discussion presented in
@ref{Passwd Functions}
applies to the group database as well.  Although there has traditionally
been a well-known file (@file{/etc/group}) in a well-known format, the POSIX
standard only provides a set of C library routines
(@code{<grp.h>} and @code{getgrent()})
for accessing the information.
Even though this file may exist, it may not have
complete information.  Therefore, as with the user database, it is necessary
to have a small C program that generates the group database as its output.
@command{grcat}, a C program that ``cats'' the group database,
is as follows:

@cindex @command{grcat} program
@example
@c file eg/lib/grcat.c
/*
 * grcat.c
 *
 * Generate a printable version of the group database.
 */
@c endfile
@ignore
@c file eg/lib/grcat.c
/*
 * Arnold Robbins, arnold@@skeeve.com, May 1993
 * Public Domain
 * December 2010, move to ANSI C definition for main().
 */

#if HAVE_CONFIG_H
#include <config.h>
#endif

#if defined (STDC_HEADERS)
#include <stdlib.h>
#endif

#ifndef HAVE_GETGRENT
int main() { return 0; }
#else
@c endfile
@end ignore
@c file eg/lib/grcat.c
#include <stdio.h>
#include <grp.h>

int
main(int argc, char **argv)
@{
    struct group *g;
    int i;

    while ((g = getgrent()) != NULL) @{
@c endfile
@ignore
@c file eg/lib/grcat.c
#ifdef HAVE_STRUCT_GROUP_GR_PASSWD
@c endfile
@end ignore
@c file eg/lib/grcat.c
        printf("%s:%s:%ld:", g->gr_name, g->gr_passwd,
                                     (long) g->gr_gid);
@c endfile
@ignore
@c file eg/lib/grcat.c
#else
        printf("%s:*:%ld:", g->gr_name, (long) g->gr_gid);
#endif
@c endfile
@end ignore
@c file eg/lib/grcat.c
        for (i = 0; g->gr_mem[i] != NULL; i++) @{
            printf("%s", g->gr_mem[i]);
@group
            if (g->gr_mem[i+1] != NULL)
                putchar(',');
        @}
@end group
        putchar('\n');
    @}
    endgrent();
    return 0;
@}
@c endfile
@ignore
@c file eg/lib/grcat.c
#endif /* HAVE_GETGRENT */
@c endfile
@end ignore
@end example

Each line in the group database represents one group.  The fields are
separated with colons and represent the following information:

@table @asis
@item Group Name
The group's name.

@item Group Password
The group's encrypted password. In practice, this field is never used;
it is usually empty or set to @samp{*}.

@item Group ID Number
The group's numeric group ID number;
the association of name to number must be unique within the file.
(On some systems it's a C @code{long}, and not an @code{int}.  Thus,
we cast it to @code{long} for all cases.)

@item Group Member List
A comma-separated list of usernames.  These users are members of the group.
Modern Unix systems allow users to be members of several groups
simultaneously.  If your system does, then there are elements
@code{"group1"} through @code{"group@var{N}"} in @code{PROCINFO}
for those group ID numbers.
(Note that @code{PROCINFO} is a @command{gawk} extension;
@pxref{Built-in Variables}.)
@end table

Here is what running @command{grcat} might produce:

@example
$ @kbd{grcat}
@print{} wheel:*:0:arnold
@print{} nogroup:*:65534:
@print{} daemon:*:1:
@print{} kmem:*:2:
@print{} staff:*:10:arnold,miriam,andy
@print{} other:*:20:
@dots{}
@end example

Here are the functions for obtaining information from the group database.
There are several, modeled after the C library functions of the same names:

@cindex @code{getline} command, @code{_gr_init()} user-defined function
@cindex @code{_gr_init()} user-defined function
@example
@c file eg/lib/groupawk.in
# group.awk --- functions for dealing with the group file
@c endfile
@ignore
@c file eg/lib/groupawk.in
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised October 2000
# Revised December 2010
@c endfile
@end ignore
@c line break on _gr_init for smallbook
@c file eg/lib/groupawk.in

BEGIN @{
    # Change to suit your system
    _gr_awklib = "/usr/local/libexec/awk/"
@}

function _gr_init(    oldfs, oldrs, olddol0, grcat,
                             using_fw, using_fpat, n, a, i)
@{
    if (_gr_inited)
        return

    oldfs = FS
    oldrs = RS
    olddol0 = $0
    using_fw = (PROCINFO["FS"] == "FIELDWIDTHS")
    using_fpat = (PROCINFO["FS"] == "FPAT")
    FS = ":"
    RS = "\n"

    grcat = _gr_awklib "grcat"
    while ((grcat | getline) > 0) @{
        if ($1 in _gr_byname)
            _gr_byname[$1] = _gr_byname[$1] "," $4
        else
            _gr_byname[$1] = $0
        if ($3 in _gr_bygid)
            _gr_bygid[$3] = _gr_bygid[$3] "," $4
        else
            _gr_bygid[$3] = $0

        n = split($4, a, "[ \t]*,[ \t]*")
        for (i = 1; i <= n; i++)
            if (a[i] in _gr_groupsbyuser)
                _gr_groupsbyuser[a[i]] = _gr_groupsbyuser[a[i]] " " $1
            else
                _gr_groupsbyuser[a[i]] = $1

        _gr_bycount[++_gr_count] = $0
    @}
    close(grcat)
    _gr_count = 0
    _gr_inited++
    FS = oldfs
    if (using_fw)
        FIELDWIDTHS = FIELDWIDTHS
    else if (using_fpat)
        FPAT = FPAT
    RS = oldrs
    $0 = olddol0
@}
@c endfile
@end example

The @code{BEGIN} rule sets a private variable to the directory where
@command{grcat} is stored.  Because it is used to help out an @command{awk} library
routine, we have chosen to put it in @file{/usr/local/libexec/awk}.  You might
want it to be in a different directory on your system.

These routines follow the same general outline as the user database routines
(@pxref{Passwd Functions}).
The @code{@w{_gr_inited}} variable is used to
ensure that the database is scanned no more than once.
The @code{@w{_gr_init()}} function first saves @code{FS},
@code{RS}, and
@code{$0}, and then sets @code{FS} and @code{RS} to the correct values for
scanning the group information.
It also takes care to note whether @code{FIELDWIDTHS} or @code{FPAT}
is being used, and to restore the appropriate field-splitting mechanism.

The group information is stored in several associative arrays.
The arrays are indexed by group name (@code{@w{_gr_byname}}), by group ID number
(@code{@w{_gr_bygid}}), and by position in the database (@code{@w{_gr_bycount}}).
There is an additional array indexed by username (@code{@w{_gr_groupsbyuser}}),
which is a space-separated list of groups to which each user belongs.

Unlike in the user database, it is possible to have multiple records in the
database for the same group.  This is common when a group has a large number
of members.  A pair of such entries might look like the following:

@example
tvpeople:*:101:johnny,jay,arsenio
tvpeople:*:101:david,conan,tom,joan
@end example

For this reason, @code{_gr_init()} looks to see if a group name or
group ID number is already seen.  If so, the usernames are
simply concatenated onto the previous list of users.@footnote{There is a
subtle problem with the code just presented.  Suppose that
the first time there were no names. This code adds the names with
a leading comma. It also doesn't check that there is a @code{$4}.}

Finally, @code{_gr_init()} closes the pipeline to @command{grcat}, restores
@code{FS} (and @code{FIELDWIDTHS} or @code{FPAT}, if necessary), @code{RS}, and @code{$0},
initializes @code{_gr_count} to zero
(it is used later), and makes @code{_gr_inited} nonzero.

@cindex @code{getgrnam()} function (C library)
The @code{getgrnam()} function takes a group name as its argument, and if that
group exists, it is returned.
Otherwise, it
relies on the array reference to a nonexistent
element to create the element with the null string as its value:

@cindex @code{getgrnam()} user-defined function
@example
@c file eg/lib/groupawk.in
function getgrnam(group)
@{
    _gr_init()
    return _gr_byname[group]
@}
@c endfile
@end example

@cindex @code{getgrgid()} function (C library)
The @code{getgrgid()} function is similar; it takes a numeric group ID and
looks up the information associated with that group ID:

@cindex @code{getgrgid()} user-defined function
@example
@c file eg/lib/groupawk.in
function getgrgid(gid)
@{
    _gr_init()
    return _gr_bygid[gid]
@}
@c endfile
@end example

@cindex @code{getgruser()} function (C library)
The @code{getgruser()} function does not have a C counterpart. It takes a
username and returns the list of groups that have the user as a member:

@cindex @code{getgruser()} function, user-defined
@example
@c file eg/lib/groupawk.in
function getgruser(user)
@{
    _gr_init()
    return _gr_groupsbyuser[user]
@}
@c endfile
@end example

@cindex @code{getgrent()} function (C library)
The @code{getgrent()} function steps through the database one entry at a time.
It uses @code{_gr_count} to track its position in the list:

@cindex @code{getgrent()} user-defined function
@example
@c file eg/lib/groupawk.in
function getgrent()
@{
    _gr_init()
    if (++_gr_count in _gr_bycount)
        return _gr_bycount[_gr_count]
@group
    return ""
@}
@end group
@c endfile
@end example

@cindex @code{endgrent()} function (C library)
The @code{endgrent()} function resets @code{_gr_count} to zero so that @code{getgrent()} can
start over again:

@cindex @code{endgrent()} user-defined function
@example
@c file eg/lib/groupawk.in
function endgrent()
@{
    _gr_count = 0
@}
@c endfile
@end example

As with the user database routines, each function calls @code{_gr_init()} to
initialize the arrays.  Doing so only incurs the extra overhead of running
@command{grcat} if these functions are used (as opposed to moving the body of
@code{_gr_init()} into a @code{BEGIN} rule).

Most of the work is in scanning the database and building the various
associative arrays.  The functions that the user calls are themselves very
simple, relying on @command{awk}'s associative arrays to do work.

The @command{id} program in @ref{Id Program}
uses these functions.

@node Walking Arrays
@section Traversing Arrays of Arrays

@ref{Arrays of Arrays} described how @command{gawk}
provides arrays of arrays.  In particular, any element of
an array may be either a scalar or another array. The
@code{isarray()} function (@pxref{Type Functions})
lets you distinguish an array
from a scalar.
The following function, @code{walk_array()}, recursively traverses
an array, printing the element indices and values.
You call it with the array and a string representing the name
of the array:

@cindex @code{walk_array()} user-defined function
@example
@c file eg/lib/walkarray.awk
function walk_array(arr, name,      i)
@{
    for (i in arr) @{
        if (isarray(arr[i]))
            walk_array(arr[i], (name "[" i "]"))
        else
            printf("%s[%s] = %s\n", name, i, arr[i])
    @}
@}
@c endfile
@end example

@noindent
It works by looping over each element of the array. If any given
element is itself an array, the function calls itself recursively,
passing the subarray and a new string representing the current index.
Otherwise, the function simply prints the element's name, index, and value.
Here is a main program to demonstrate:

@example
BEGIN @{
    a[1] = 1
    a[2][1] = 21
    a[2][2] = 22
    a[3] = 3
    a[4][1][1] = 411
    a[4][2] = 42

    walk_array(a, "a")
@}
@end example

When run, the program produces the following output:

@example
$ @kbd{gawk -f walk_array.awk}
@print{} a[1] = 1
@print{} a[2][1] = 21
@print{} a[2][2] = 22
@print{} a[3] = 3
@print{} a[4][1][1] = 411
@print{} a[4][2] = 42
@end example

The function just presented simply prints the
name and value of each scalar array element. However, it is easy to
generalize it, by passing in the name of a function to call
when walking an array. The modified function looks like this:

@example
@c file eg/lib/processarray.awk
function process_array(arr, name, process, do_arrays,   i, new_name)
@{
    for (i in arr) @{
        new_name = (name "[" i "]")
        if (isarray(arr[i])) @{
            if (do_arrays)
                @@process(new_name, arr[i])
            process_array(arr[i], new_name, process, do_arrays)
        @} else
            @@process(new_name, arr[i])
    @}
@}
@c endfile
@end example

The arguments are as follows:

@table @code
@item arr
The array.

@item name
The name of the array (a string).

@item process
The name of the function to call.

@item do_arrays
If this is true, the function can handle elements that are subarrays.
@end table

If subarrays are to be processed, that is done before walking them further.

When run with the following scaffolding, the function produces the same
results as does the earlier version of @code{walk_array()}:

@example
BEGIN @{
    a[1] = 1
    a[2][1] = 21
    a[2][2] = 22
    a[3] = 3
    a[4][1][1] = 411
    a[4][2] = 42

    process_array(a, "a", "do_print", 0)
@}

function do_print(name, element)
@{
    printf "%s = %s\n", name, element
@}
@end example

@node Library Functions Summary
@section Summary

@itemize @value{BULLET}
@item
Reading programs is an excellent way to learn Good Programming.
The functions and programs provided in this @value{CHAPTER} and the next
are intended to serve that purpose.

@item
When writing general-purpose library functions, put some thought into how
to name any global variables so that they won't conflict with variables
from a user's program.

@item
The functions presented here fit into the following categories:

@c nested list
@table @asis
@item General problems
Number-to-string conversion, testing assertions, rounding, random number
generation, converting characters to numbers, joining strings, getting
easily usable time-of-day information, and reading a whole file in
one shot

@item Managing @value{DF}s
Noting @value{DF} boundaries, rereading the current file, checking for
readable files, checking for zero-length files, and treating assignments
as @value{FN}s

@item Processing command-line options
An @command{awk} version of the standard C @code{getopt()} function

@item Reading the user and group databases
Two sets of routines that parallel the C library versions

@item Traversing arrays of arrays
Two functions that traverse an array of arrays to any depth
@end table
@c end nested list

@end itemize

@c EXCLUDE START
@node Library Exercises
@section Exercises

@enumerate
@item
In @ref{Empty Files}, we presented the @file{zerofile.awk} program,
which made use of @command{gawk}'s @code{ARGIND} variable.  Can this
problem be solved without relying on @code{ARGIND}?  If so, how?

@ignore
# zerofile2.awk --- same thing, portably

BEGIN @{
    ARGIND = Argind = 0
    for (i = 1; i < ARGC; i++)
        Fnames[ARGV[i]]++

@}
FNR == 1 @{
    while (ARGV[ARGIND] != FILENAME)
        ARGIND++
    Seen[FILENAME]++
    if (Seen[FILENAME] == Fnames[FILENAME])
        do
            ARGIND++
        while (ARGV[ARGIND] != FILENAME)
@}
ARGIND > Argind + 1 @{
    for (Argind++; Argind < ARGIND; Argind++)
        zerofile(ARGV[Argind], Argind)
@}
ARGIND != Argind @{
    Argind = ARGIND
@}
END @{
    if (ARGIND < ARGC - 1)
        ARGIND = ARGC - 1
    if (ARGIND > Argind)
        for (Argind++; Argind <= ARGIND; Argind++)
            zerofile(ARGV[Argind], Argind)
@}
@end ignore

@item
As a related challenge, revise that code to handle the case where
an intervening value in @code{ARGV} is a variable assignment.

@ignore
@c June 13 2015: Antonio points out that this is answered in the text. Ooops.
@item
@ref{Walking Arrays} presented a function that walked a multidimensional
array to print it out.  However, walking an array and processing
each element is a general-purpose operation.  Generalize the
@code{walk_array()} function by adding an additional parameter named
@code{process}.

Then, inside the loop, instead of printing the array element's index and
value, use the indirect function call syntax (@pxref{Indirect Calls})
on @code{process}, passing it the index and the value.

When calling @code{walk_array()}, you would pass the name of a
user-defined function that expects to receive an index and a value,
and then processes the element.

Test your new version by printing the array; you should end up with
output identical to that of the original version.
@end ignore

@end enumerate
@c EXCLUDE END


@node Sample Programs
@chapter Practical @command{awk} Programs
@cindex @command{awk} programs, examples of

@c FULLXREF ON
@ref{Library Functions},
presents the idea that reading programs in a language contributes to
learning that language.  This @value{CHAPTER} continues that theme,
presenting a potpourri of @command{awk} programs for your reading
enjoyment.
@c FULLXREF OFF
@ifnotinfo
There are three @value{SECTION}s.
The first describes how to run the programs presented
in this @value{CHAPTER}.

The second presents @command{awk}
versions of several common POSIX utilities.
These are programs that you are hopefully already familiar with,
and therefore whose problems are understood.
By reimplementing these programs in @command{awk},
you can focus on the @command{awk}-related aspects of solving
the programming problems.

The third is a grab bag of interesting programs.
These solve a number of different data-manipulation and management
problems.  Many of the programs are short, which emphasizes @command{awk}'s
ability to do a lot in just a few lines of code.
@end ifnotinfo

Many of these programs use library functions presented in
@ref{Library Functions}.

@menu
* Running Examples::            How to run these examples.
* Clones::                      Clones of common utilities.
* Miscellaneous Programs::      Some interesting @command{awk} programs.
* Programs Summary::            Summary of programs.
* Programs Exercises::          Exercises.
@end menu

@node Running Examples
@section Running the Example Programs

To run a given program, you would typically do something like this:

@example
awk -f @var{program} -- @var{options} @var{files}
@end example

@noindent
Here, @var{program} is the name of the @command{awk} program (such as
@file{cut.awk}), @var{options} are any command-line options for the
program that start with a @samp{-}, and @var{files} are the actual @value{DF}s.

If your system supports the @samp{#!} executable interpreter mechanism
(@pxref{Executable Scripts}),
you can instead run your program directly:

@example
cut.awk -c1-8 myfiles > results
@end example

If your @command{awk} is not @command{gawk}, you may instead need to use this:

@example
cut.awk -- -c1-8 myfiles > results
@end example

@node Clones
@section Reinventing Wheels for Fun and Profit
@cindex POSIX, programs@comma{} implementing in @command{awk}

This @value{SECTION} presents a number of POSIX utilities implemented in
@command{awk}.  Reinventing these programs in @command{awk} is often enjoyable,
because the algorithms can be very clearly expressed, and the code is usually
very concise and simple.  This is true because @command{awk} does so much for you.

It should be noted that these programs are not necessarily intended to
replace the installed versions on your system.
Nor may all of these programs be fully compliant with the most recent
POSIX standard.  This is not a problem; their
purpose is to illustrate @command{awk} language programming for ``real-world''
tasks.

The programs are presented in alphabetical order.

@menu
* Cut Program::                 The @command{cut} utility.
* Egrep Program::               The @command{egrep} utility.
* Id Program::                  The @command{id} utility.
* Split Program::               The @command{split} utility.
* Tee Program::                 The @command{tee} utility.
* Uniq Program::                The @command{uniq} utility.
* Wc Program::                  The @command{wc} utility.
@end menu

@node Cut Program
@subsection Cutting Out Fields and Columns

@cindex @command{cut} utility
@cindex @command{cut} utility
@cindex fields, cutting
@cindex columns, cutting
The @command{cut} utility selects, or ``cuts,'' characters or fields
from its standard input and sends them to its standard output.
Fields are separated by TABs by default,
but you may supply a command-line option to change the field
@dfn{delimiter} (i.e., the field-separator character). @command{cut}'s
definition of fields is less general than @command{awk}'s.

A common use of @command{cut} might be to pull out just the login names of
logged-on users from the output of @command{who}.  For example, the following
pipeline generates a sorted, unique list of the logged-on users:

@example
who | cut -c1-8 | sort | uniq
@end example

The options for @command{cut} are:

@table @code
@item -c @var{list}
Use @var{list} as the list of characters to cut out.  Items within the list
may be separated by commas, and ranges of characters can be separated with
dashes.  The list @samp{1-8,15,22-35} specifies characters 1 through
8, 15, and 22 through 35.

@item -f @var{list}
Use @var{list} as the list of fields to cut out.

@item -d @var{delim}
Use @var{delim} as the field-separator character instead of the TAB
character.

@item -s
Suppress printing of lines that do not contain the field delimiter.
@end table

The @command{awk} implementation of @command{cut} uses the @code{getopt()} library
function (@pxref{Getopt Function})
and the @code{join()} library function
(@pxref{Join Function}).

The program begins with a comment describing the options, the library
functions needed, and a @code{usage()} function that prints out a usage
message and exits.  @code{usage()} is called if invalid arguments are
supplied:

@cindex @code{cut.awk} program
@example
@c file eg/prog/cut.awk
# cut.awk --- implement cut in awk
@c endfile
@ignore
@c file eg/prog/cut.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
@c endfile
@end ignore
@c file eg/prog/cut.awk

# Options:
#    -f list     Cut fields
#    -d c        Field delimiter character
#    -c list     Cut characters
#
#    -s          Suppress lines without the delimiter
#
# Requires getopt() and join() library functions

@group
function usage()
@{
    print("usage: cut [-f list] [-d c] [-s] [files...]") > "/dev/stderr"
    print("usage: cut [-c list] [files...]") > "/dev/stderr"
    exit 1
@}
@end group
@c endfile
@end example

@cindex @code{BEGIN} pattern, running @command{awk} programs and
@cindex @code{FS} variable, running @command{awk} programs and
Next comes a @code{BEGIN} rule that parses the command-line options.
It sets @code{FS} to a single TAB character, because that is @command{cut}'s
default field separator. The rule then sets the output field separator to be the
same as the input field separator.  A loop using @code{getopt()} steps
through the command-line options.  Exactly one of the variables
@code{by_fields} or @code{by_chars} is set to true, to indicate that
processing should be done by fields or by characters, respectively.
When cutting by characters, the output field separator is set to the null
string:

@example
@c file eg/prog/cut.awk
BEGIN @{
    FS = "\t"    # default
    OFS = FS
    while ((c = getopt(ARGC, ARGV, "sf:c:d:")) != -1) @{
        if (c == "f") @{
            by_fields = 1
            fieldlist = Optarg
        @} else if (c == "c") @{
            by_chars = 1
            fieldlist = Optarg
            OFS = ""
        @} else if (c == "d") @{
            if (length(Optarg) > 1) @{
                printf("cut: using first character of %s" \
                       " for delimiter\n", Optarg) > "/dev/stderr"
                Optarg = substr(Optarg, 1, 1)
            @}
            fs = FS = Optarg
            OFS = FS
            if (FS == " ")    # defeat awk semantics
                FS = "[ ]"
        @} else if (c == "s")
            suppress = 1
        else
            usage()
    @}

    # Clear out options
    for (i = 1; i < Optind; i++)
        ARGV[i] = ""
@c endfile
@end example

@cindex field separators, spaces as
The code must take
special care when the field delimiter is a space.  Using
a single space (@code{@w{" "}}) for the value of @code{FS} is
incorrect---@command{awk} would separate fields with runs of spaces,
TABs, and/or newlines, and we want them to be separated with individual
spaces.
To this end, we save the original space character in the variable
@code{fs} for later use; after setting @code{FS} to @code{"[ ]"} we can't
use it directly to see if the field delimiter character is in the string.

Also remember that after @code{getopt()} is through
(as described in @ref{Getopt Function}),
we have to
clear out all the elements of @code{ARGV} from 1 to @code{Optind},
so that @command{awk} does not try to process the command-line options
as @value{FN}s.

After dealing with the command-line options, the program verifies that the
options make sense.  Only one or the other of @option{-c} and @option{-f}
should be used, and both require a field list.  Then the program calls
either @code{set_fieldlist()} or @code{set_charlist()} to pull apart the
list of fields or characters:

@example
@c file eg/prog/cut.awk
    if (by_fields && by_chars)
        usage()

    if (by_fields == 0 && by_chars == 0)
        by_fields = 1    # default

@group
    if (fieldlist == "") @{
        print "cut: needs list for -c or -f" > "/dev/stderr"
        exit 1
    @}
@end group

    if (by_fields)
        set_fieldlist()
    else
        set_charlist()
@}
@c endfile
@end example

@code{set_fieldlist()} splits the field list apart at the commas
into an array.  Then, for each element of the array, it looks to
see if the element is actually a range, and if so, splits it apart.
The function checks the range
to make sure that the first number is smaller than the second.
Each number in the list is added to the @code{flist} array, which
simply lists the fields that will be printed.  Normal field splitting
is used.  The program lets @command{awk} handle the job of doing the
field splitting:

@example
@c file eg/prog/cut.awk
function set_fieldlist(        n, m, i, j, k, f, g)
@{
    n = split(fieldlist, f, ",")
    j = 1    # index in flist
    for (i = 1; i <= n; i++) @{
        if (index(f[i], "-") != 0) @{ # a range
            m = split(f[i], g, "-")
@group
            if (m != 2 || g[1] >= g[2]) @{
                printf("cut: bad field list: %s\n",
                                  f[i]) > "/dev/stderr"
                exit 1
            @}
@end group
            for (k = g[1]; k <= g[2]; k++)
                flist[j++] = k
        @} else
            flist[j++] = f[i]
    @}
    nfields = j - 1
@}
@c endfile
@end example

The @code{set_charlist()} function is more complicated than
@code{set_fieldlist()}.
The idea here is to use @command{gawk}'s @code{FIELDWIDTHS} variable
(@pxref{Constant Size}),
which describes constant-width input.  When using a character list, that is
exactly what we have.

Setting up @code{FIELDWIDTHS} is more complicated than simply listing the
fields that need to be printed.  We have to keep track of the fields to
print and also the intervening characters that have to be skipped.
For example, suppose you wanted characters 1 through 8, 15, and
22 through 35.  You would use @samp{-c 1-8,15,22-35}.  The necessary value
for @code{FIELDWIDTHS} is @code{@w{"8 6 1 6 14"}}.  This yields five
fields, and the fields to print
are @code{$1}, @code{$3}, and @code{$5}.
The intermediate fields are @dfn{filler},
which is stuff in between the desired data.
@code{flist} lists the fields to print, and @code{t} tracks the
complete field list, including filler fields:

@example
@c file eg/prog/cut.awk
function set_charlist(    field, i, j, f, g, n, m, t,
                          filler, last, len)
@{
    field = 1   # count total fields
    n = split(fieldlist, f, ",")
    j = 1       # index in flist
    for (i = 1; i <= n; i++) @{
        if (index(f[i], "-") != 0) @{ # range
            m = split(f[i], g, "-")
            if (m != 2 || g[1] >= g[2]) @{
                printf("cut: bad character list: %s\n",
                               f[i]) > "/dev/stderr"
                exit 1
            @}
            len = g[2] - g[1] + 1
            if (g[1] > 1)  # compute length of filler
                filler = g[1] - last - 1
            else
                filler = 0
@group
            if (filler)
                t[field++] = filler
@end group
            t[field++] = len  # length of field
            last = g[2]
            flist[j++] = field - 1
        @} else @{
            if (f[i] > 1)
                filler = f[i] - last - 1
            else
                filler = 0
            if (filler)
                t[field++] = filler
            t[field++] = 1
            last = f[i]
            flist[j++] = field - 1
        @}
    @}
    FIELDWIDTHS = join(t, 1, field - 1)
    nfields = j - 1
@}
@c endfile
@end example

Next is the rule that processes the data.  If the @option{-s} option
is given, then @code{suppress} is true.  The first @code{if} statement
makes sure that the input record does have the field separator.  If
@command{cut} is processing fields, @code{suppress} is true, and the field
separator character is not in the record, then the record is skipped.

If the record is valid, then @command{gawk} has split the data
into fields, either using the character in @code{FS} or using fixed-length
fields and @code{FIELDWIDTHS}.  The loop goes through the list of fields
that should be printed.  The corresponding field is printed if it contains data.
If the next field also has data, then the separator character is
written out between the fields:

@example
@c file eg/prog/cut.awk
@{
    if (by_fields && suppress && index($0, fs) == 0)
        next

    for (i = 1; i <= nfields; i++) @{
        if ($flist[i] != "") @{
            printf "%s", $flist[i]
            if (i < nfields && $flist[i+1] != "")
                printf "%s", OFS
        @}
    @}
    print ""
@}
@c endfile
@end example

This version of @command{cut} relies on @command{gawk}'s @code{FIELDWIDTHS}
variable to do the character-based cutting.  It is possible in
other @command{awk} implementations to use @code{substr()}
(@pxref{String Functions}), but
it is also extremely painful.
The @code{FIELDWIDTHS} variable supplies an elegant solution to the problem
of picking the input line apart by characters.


@node Egrep Program
@subsection Searching for Regular Expressions in Files

@cindex regular expressions, searching for
@cindex searching, files for regular expressions
@cindex files, searching for regular expressions
@cindex @command{egrep} utility
The @command{egrep} utility searches files for patterns.  It uses regular
expressions that are almost identical to those available in @command{awk}
(@pxref{Regexp}).
You invoke it as follows:

@display
@command{egrep} [@var{options}] @code{'@var{pattern}'} @var{files} @dots{}
@end display

The @var{pattern} is a regular expression.  In typical usage, the regular
expression is quoted to prevent the shell from expanding any of the
special characters as @value{FN} wildcards.  Normally, @command{egrep}
prints the lines that matched.  If multiple @value{FN}s are provided on
the command line, each output line is preceded by the name of the file
and a colon.

The options to @command{egrep} are as follows:

@table @code
@item -c
Print out a count of the lines that matched the pattern, instead of the
lines themselves.

@item -s
Be silent.  No output is produced and the exit value indicates whether
the pattern was matched.

@item -v
Invert the sense of the test. @command{egrep} prints the lines that do
@emph{not} match the pattern and exits successfully if the pattern is not
matched.

@item -i
Ignore case distinctions in both the pattern and the input data.

@item -l
Only print (list) the names of the files that matched, not the lines that matched.

@item -e @var{pattern}
Use @var{pattern} as the regexp to match.  The purpose of the @option{-e}
option is to allow patterns that start with a @samp{-}.
@end table

This version uses the @code{getopt()} library function
(@pxref{Getopt Function})
and the file transition library program
(@pxref{Filetrans Function}).

The program begins with a descriptive comment and then a @code{BEGIN} rule
that processes the command-line arguments with @code{getopt()}.  The @option{-i}
(ignore case) option is particularly easy with @command{gawk}; we just use the
@code{IGNORECASE} predefined variable
(@pxref{Built-in Variables}):

@cindex @code{egrep.awk} program
@example
@c file eg/prog/egrep.awk
# egrep.awk --- simulate egrep in awk
#
@c endfile
@ignore
@c file eg/prog/egrep.awk
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993

@c endfile
@end ignore
@c file eg/prog/egrep.awk
# Options:
#    -c    count of lines
#    -s    silent - use exit value
#    -v    invert test, success if no match
#    -i    ignore case
#    -l    print filenames only
#    -e    argument is pattern
#
# Requires getopt and file transition library functions

BEGIN @{
    while ((c = getopt(ARGC, ARGV, "ce:svil")) != -1) @{
        if (c == "c")
            count_only++
        else if (c == "s")
            no_print++
        else if (c == "v")
            invert++
        else if (c == "i")
            IGNORECASE = 1
        else if (c == "l")
            filenames_only++
        else if (c == "e")
            pattern = Optarg
        else
            usage()
    @}
@c endfile
@end example

Next comes the code that handles the @command{egrep}-specific behavior. If no
pattern is supplied with @option{-e}, the first nonoption on the
command line is used.  The @command{awk} command-line arguments up to @code{ARGV[Optind]}
are cleared, so that @command{awk} won't try to process them as files.  If no
files are specified, the standard input is used, and if multiple files are
specified, we make sure to note this so that the @value{FN}s can precede the
matched lines in the output:

@example
@c file eg/prog/egrep.awk
    if (pattern == "")
        pattern = ARGV[Optind++]

    for (i = 1; i < Optind; i++)
        ARGV[i] = ""
    if (Optind >= ARGC) @{
        ARGV[1] = "-"
        ARGC = 2
    @} else if (ARGC - Optind > 1)
        do_filenames++

#    if (IGNORECASE)
#        pattern = tolower(pattern)
@}
@c endfile
@end example

The last two lines are commented out, as they are not needed in
@command{gawk}.  They should be uncommented if you have to use another version
of @command{awk}.

The next set of lines should be uncommented if you are not using
@command{gawk}.  This rule translates all the characters in the input line
into lowercase if the @option{-i} option is specified.@footnote{It
also introduces a subtle bug;
if a match happens, we output the translated line, not the original.}
The rule is
commented out as it is not necessary with @command{gawk}:

@example
@c file eg/prog/egrep.awk
#@{
#    if (IGNORECASE)
#        $0 = tolower($0)
#@}
@c endfile
@end example

The @code{beginfile()} function is called by the rule in @file{ftrans.awk}
when each new file is processed.  In this case, it is very simple; all it
does is initialize a variable @code{fcount} to zero. @code{fcount} tracks
how many lines in the current file matched the pattern.
Naming the parameter @code{junk} shows we know that @code{beginfile()}
is called with a parameter, but that we're not interested in its value:

@example
@c file eg/prog/egrep.awk
function beginfile(junk)
@{
    fcount = 0
@}
@c endfile
@end example

The @code{endfile()} function is called after each file has been processed.
It affects the output only when the user wants a count of the number of lines that
matched.  @code{no_print} is true only if the exit status is desired.
@code{count_only} is true if line counts are desired.  @command{egrep}
therefore only prints line counts if printing and counting are enabled.
The output format must be adjusted depending upon the number of files to
process.  Finally, @code{fcount} is added to @code{total}, so that we
know the total number of lines that matched the pattern:

@example
@c file eg/prog/egrep.awk
function endfile(file)
@{
    if (! no_print && count_only) @{
        if (do_filenames)
            print file ":" fcount
        else
            print fcount
    @}

@group
    total += fcount
@}
@end group
@c endfile
@end example

The @code{BEGINFILE} and @code{ENDFILE} special patterns
(@pxref{BEGINFILE/ENDFILE}) could be used, but then the program would be
@command{gawk}-specific. Additionally, this example was written before
@command{gawk} acquired @code{BEGINFILE} and @code{ENDFILE}.

The following rule does most of the work of matching lines. The variable
@code{matches} is true if the line matched the pattern. If the user
wants lines that did not match, the sense of @code{matches} is inverted
using the @samp{!} operator. @code{fcount} is incremented with the value of
@code{matches}, which is either one or zero, depending upon a
successful or unsuccessful match.  If the line does not match, the
@code{next} statement just moves on to the next record.

A number of additional tests are made, but they are only done if we
are not counting lines.  First, if the user only wants the exit status
(@code{no_print} is true), then it is enough to know that @emph{one}
line in this file matched, and we can skip on to the next file with
@code{nextfile}.  Similarly, if we are only printing @value{FN}s, we can
print the @value{FN}, and then skip to the next file with @code{nextfile}.
Finally, each line is printed, with a leading @value{FN} and colon
if necessary:

@cindex @code{!} (exclamation point), @code{!} operator
@cindex exclamation point (@code{!}), @code{!} operator
@example
@c file eg/prog/egrep.awk
@{
    matches = ($0 ~ pattern)
    if (invert)
        matches = ! matches

    fcount += matches    # 1 or 0

    if (! matches)
        next

    if (! count_only) @{
        if (no_print)
            nextfile

        if (filenames_only) @{
            print FILENAME
            nextfile
        @}

        if (do_filenames)
            print FILENAME ":" $0
        else
            print
    @}
@}
@c endfile
@end example

The @code{END} rule takes care of producing the correct exit status. If
there are no matches, the exit status is one; otherwise, it is zero:

@example
@c file eg/prog/egrep.awk
END @{
    exit (total == 0)
@}
@c endfile
@end example

The @code{usage()} function prints a usage message in case of invalid options,
and then exits:

@example
@c file eg/prog/egrep.awk
function usage()
@{
    print("Usage: egrep [-csvil] [-e pat] [files ...]") > "/dev/stderr"
    print("\n\tegrep [-csvil] pat [files ...]") > "/dev/stderr"
    exit 1
@}
@c endfile
@end example


@node Id Program
@subsection Printing Out User Information

@cindex printing, user information
@cindex users, information about, printing
@cindex @command{id} utility
The @command{id} utility lists a user's real and effective user ID numbers,
real and effective group ID numbers, and the user's group set, if any.
@command{id} only prints the effective user ID and group ID if they are
different from the real ones.  If possible, @command{id} also supplies the
corresponding user and group names.  The output might look like this:

@example
$ @kbd{id}
@print{} uid=1000(arnold) gid=1000(arnold) groups=1000(arnold),4(adm),7(lp),27(sudo)
@end example

@cindex @code{PROCINFO} array, and user and group ID numbers
This information is part of what is provided by @command{gawk}'s
@code{PROCINFO} array (@pxref{Built-in Variables}).
However, the @command{id} utility provides a more palatable output than just
individual numbers.

Here is a simple version of @command{id} written in @command{awk}.
It uses the user database library functions
(@pxref{Passwd Functions})
and the group database library functions
(@pxref{Group Functions})
from @ref{Library Functions}.

The program is fairly straightforward.  All the work is done in the
@code{BEGIN} rule.  The user and group ID numbers are obtained from
@code{PROCINFO}.
The code is repetitive.  The entry in the user database for the real user ID
number is split into parts at the @samp{:}. The name is the first field.
Similar code is used for the effective user ID number and the group
numbers:

@cindex @code{id.awk} program
@example
@c file eg/prog/id.awk
# id.awk --- implement id in awk
#
# Requires user and group library functions
@c endfile
@ignore
@c file eg/prog/id.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised February 1996
# Revised May 2014
# Revised September 2014

@c endfile
@end ignore
@c file eg/prog/id.awk
# output is:
# uid=12(foo) euid=34(bar) gid=3(baz) \
#             egid=5(blat) groups=9(nine),2(two),1(one)

@group
BEGIN @{
    uid = PROCINFO["uid"]
    euid = PROCINFO["euid"]
    gid = PROCINFO["gid"]
    egid = PROCINFO["egid"]
@end group

    printf("uid=%d", uid)
    pw = getpwuid(uid)
    pr_first_field(pw)

@group
    if (euid != uid) @{
        printf(" euid=%d", euid)
        pw = getpwuid(euid)
@end group
@group
        pr_first_field(pw)
    @}
@end group

    printf(" gid=%d", gid)
    pw = getgrgid(gid)
    pr_first_field(pw)

    if (egid != gid) @{
        printf(" egid=%d", egid)
        pw = getgrgid(egid)
        pr_first_field(pw)
    @}

    for (i = 1; ("group" i) in PROCINFO; i++) @{
        if (i == 1)
            printf(" groups=")
        group = PROCINFO["group" i]
        printf("%d", group)
        pw = getgrgid(group)
        pr_first_field(pw)
        if (("group" (i+1)) in PROCINFO)
            printf(",")
    @}

    print ""
@}

function pr_first_field(str,  a)
@{
    if (str != "") @{
        split(str, a, ":")
        printf("(%s)", a[1])
    @}
@}
@c endfile
@end example

The test in the @code{for} loop is worth noting.
Any supplementary groups in the @code{PROCINFO} array have the
indices @code{"group1"} through @code{"group@var{N}"} for some
@var{N} (i.e., the total number of supplementary groups).
However, we don't know in advance how many of these groups
there are.

This loop works by starting at one, concatenating the value with
@code{"group"}, and then using @code{in} to see if that value is
in the array (@pxref{Reference to Elements}).  Eventually, @code{i} is incremented past
the last group in the array and the loop exits.

The loop is also correct if there are @emph{no} supplementary
groups; then the condition is false the first time it's
tested, and the loop body never executes.

The @code{pr_first_field()} function simply isolates out some
code that is used repeatedly, making the whole program
shorter and cleaner. In particular, moving the check for
the empty string into this function saves several lines of code.


@node Split Program
@subsection Splitting a Large File into Pieces

@c FIXME: One day, update to current POSIX version of split

@cindex files, splitting
@cindex @code{split} utility
The @command{split} program splits large text files into smaller pieces.
Usage is as follows:@footnote{This is the traditional usage. The
POSIX usage is different, but not relevant for what the program
aims to demonstrate.}

@display
@command{split} [@code{-@var{count}}] [@var{file}] [@var{prefix}]
@end display

By default,
the output files are named @file{xaa}, @file{xab}, and so on. Each file has
1,000 lines in it, with the likely exception of the last file. To change the
number of lines in each file, supply a number on the command line
preceded with a minus sign (e.g., @samp{-500} for files with 500 lines in them
instead of 1,000).  To change the names of the output files to something like
@file{myfileaa}, @file{myfileab}, and so on, supply an additional
argument that specifies the @value{FN} prefix.

Here is a version of @command{split} in @command{awk}. It uses the
@code{ord()} and @code{chr()} functions presented in
@ref{Ordinal Functions}.

The program first sets its defaults, and then tests to make sure there are
not too many arguments.  It then looks at each argument in turn.  The
first argument could be a minus sign followed by a number. If it is, this happens
to look like a negative number, so it is made positive, and that is the
count of lines.  The @value{DF} name is skipped over and the final argument
is used as the prefix for the output @value{FN}s:

@cindex @code{split.awk} program
@example
@c file eg/prog/split.awk
# split.awk --- do split in awk
#
# Requires ord() and chr() library functions
@c endfile
@ignore
@c file eg/prog/split.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised slightly, May 2014

@c endfile
@end ignore
@c file eg/prog/split.awk
# usage: split [-count] [file] [outname]

BEGIN @{
    outfile = "x"    # default
    count = 1000
    if (ARGC > 4)
        usage()

    i = 1
    if (i in ARGV && ARGV[i] ~ /^-[[:digit:]]+$/) @{
        count = -ARGV[i]
        ARGV[i] = ""
        i++
    @}
    # test argv in case reading from stdin instead of file
    if (i in ARGV)
        i++    # skip datafile name
@group
    if (i in ARGV) @{
        outfile = ARGV[i]
        ARGV[i] = ""
    @}
@end group
@group
    s1 = s2 = "a"
    out = (outfile s1 s2)
@}
@end group
@c endfile
@end example

The next rule does most of the work. @code{tcount} (temporary count) tracks
how many lines have been printed to the output file so far. If it is greater
than @code{count}, it is time to close the current file and start a new one.
@code{s1} and @code{s2} track the current suffixes for the @value{FN}. If
they are both @samp{z}, the file is just too big.  Otherwise, @code{s1}
moves to the next letter in the alphabet and @code{s2} starts over again at
@samp{a}:

@c else on separate line here for page breaking
@example
@c file eg/prog/split.awk
@{
    if (++tcount > count) @{
        close(out)
        if (s2 == "z") @{
            if (s1 == "z") @{
                printf("split: %s is too large to split\n",
                       FILENAME) > "/dev/stderr"
                exit 1
            @}
            s1 = chr(ord(s1) + 1)
            s2 = "a"
        @}
@group
        else
            s2 = chr(ord(s2) + 1)
@end group
        out = (outfile s1 s2)
        tcount = 1
    @}
    print > out
@}
@c endfile
@end example

@noindent
The @code{usage()} function simply prints an error message and exits:

@example
@c file eg/prog/split.awk
function usage()
@{
    print("usage: split [-num] [file] [outname]") > "/dev/stderr"
    exit 1
@}
@c endfile
@end example

This program is a bit sloppy; it relies on @command{awk} to automatically close the last file
instead of doing it in an @code{END} rule.
It also assumes that letters are contiguous in the character set,
which isn't true for EBCDIC systems.

@ifset FOR_PRINT
You might want to consider how to eliminate the use of
@code{ord()} and @code{chr()}; this can be done in such a
way as to solve the EBCDIC issue as well.
@end ifset


@node Tee Program
@subsection Duplicating Output into Multiple Files

@cindex files, multiple@comma{} duplicating output into
@cindex output, duplicating into files
@cindex @code{tee} utility
The @code{tee} program is known as a ``pipe fitting.''  @code{tee} copies
its standard input to its standard output and also duplicates it to the
files named on the command line.  Its usage is as follows:

@display
@command{tee} [@option{-a}] @var{file} @dots{}
@end display

The @option{-a} option tells @code{tee} to append to the named files, instead of
truncating them and starting over.

The @code{BEGIN} rule first makes a copy of all the command-line arguments
into an array named @code{copy}.
@code{ARGV[0]} is not needed, so it is not copied.
@code{tee} cannot use @code{ARGV} directly, because @command{awk} attempts to
process each @value{FN} in @code{ARGV} as input data.

@cindex flag variables
If the first argument is @option{-a}, then the flag variable
@code{append} is set to true, and both @code{ARGV[1]} and
@code{copy[1]} are deleted. If @code{ARGC} is less than two, then no
@value{FN}s were supplied and @code{tee} prints a usage message and exits.
Finally, @command{awk} is forced to read the standard input by setting
@code{ARGV[1]} to @code{"-"} and @code{ARGC} to two:

@cindex @code{tee.awk} program
@example
@c file eg/prog/tee.awk
# tee.awk --- tee in awk
#
# Copy standard input to all named output files.
# Append content if -a option is supplied.
#
@c endfile
@ignore
@c file eg/prog/tee.awk
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised December 1995

@c endfile
@end ignore
@c file eg/prog/tee.awk
BEGIN @{
    for (i = 1; i < ARGC; i++)
        copy[i] = ARGV[i]

    if (ARGV[1] == "-a") @{
        append = 1
        delete ARGV[1]
        delete copy[1]
        ARGC--
    @}
    if (ARGC < 2) @{
        print "usage: tee [-a] file ..." > "/dev/stderr"
        exit 1
    @}
    ARGV[1] = "-"
    ARGC = 2
@}
@c endfile
@end example

The following single rule does all the work.  Because there is no pattern, it is
executed for each line of input.  The body of the rule simply prints the
line into each file on the command line, and then to the standard output:

@example
@c file eg/prog/tee.awk
@{
    # moving the if outside the loop makes it run faster
    if (append)
        for (i in copy)
            print >> copy[i]
    else
        for (i in copy)
            print > copy[i]
    print
@}
@c endfile
@end example

@noindent
It is also possible to write the loop this way:

@example
@group
for (i in copy)
    if (append)
        print >> copy[i]
@end group
@group
    else
        print > copy[i]
@end group
@end example

@noindent
This is more concise, but it is also less efficient.  The @samp{if} is
tested for each record and for each output file.  By duplicating the loop
body, the @samp{if} is only tested once for each input record.  If there are
@var{N} input records and @var{M} output files, the first method only
executes @var{N} @samp{if} statements, while the second executes
@var{N}@code{*}@var{M} @samp{if} statements.

Finally, the @code{END} rule cleans up by closing all the output files:

@example
@c file eg/prog/tee.awk
END @{
    for (i in copy)
        close(copy[i])
@}
@c endfile
@end example

@node Uniq Program
@subsection Printing Nonduplicated Lines of Text

@c FIXME: One day, update to current POSIX version of uniq

@cindex printing, unduplicated lines of text
@cindex text@comma{} printing, unduplicated lines of
@cindex @command{uniq} utility
The @command{uniq} utility reads sorted lines of data on its standard
input, and by default removes duplicate lines.  In other words, it only
prints unique lines---hence the name.  @command{uniq} has a number of
options. The usage is as follows:

@display
@command{uniq} [@option{-udc} [@code{-@var{n}}]] [@code{+@var{n}}] [@var{inputfile} [@var{outputfile}]]
@end display

The options for @command{uniq} are:

@table @code
@item -d
Print only repeated (duplicated) lines.

@item -u
Print only nonrepeated (unique) lines.

@item -c
Count lines. This option overrides @option{-d} and @option{-u}.  Both repeated
and nonrepeated lines are counted.

@item -@var{n}
Skip @var{n} fields before comparing lines.  The definition of fields
is similar to @command{awk}'s default: nonwhitespace characters separated
by runs of spaces and/or TABs.

@item +@var{n}
Skip @var{n} characters before comparing lines.  Any fields specified with
@samp{-@var{n}} are skipped first.

@item @var{inputfile}
Data is read from the input file named on the command line, instead of from
the standard input.

@item @var{outputfile}
The generated output is sent to the named output file, instead of to the
standard output.
@end table

Normally @command{uniq} behaves as if both the @option{-d} and
@option{-u} options are provided.

@command{uniq} uses the
@code{getopt()} library function
(@pxref{Getopt Function})
and the @code{join()} library function
(@pxref{Join Function}).

The program begins with a @code{usage()} function and then a brief outline of
the options and their meanings in comments.
The @code{BEGIN} rule deals with the command-line arguments and options. It
uses a trick to get @code{getopt()} to handle options of the form @samp{-25},
treating such an option as the option letter @samp{2} with an argument of
@samp{5}. If indeed two or more digits are supplied (@code{Optarg} looks
like a number), @code{Optarg} is
concatenated with the option digit and then the result is added to zero to make
it into a number.  If there is only one digit in the option, then
@code{Optarg} is not needed. In this case, @code{Optind} must be decremented so that
@code{getopt()} processes it next time.  This code is admittedly a bit
tricky.

If no options are supplied, then the default is taken, to print both
repeated and nonrepeated lines.  The output file, if provided, is assigned
to @code{outputfile}.  Early on, @code{outputfile} is initialized to the
standard output, @file{/dev/stdout}:

@cindex @code{uniq.awk} program
@example
@c file eg/prog/uniq.awk
@group
# uniq.awk --- do uniq in awk
#
# Requires getopt() and join() library functions
@end group
@c endfile
@ignore
@c file eg/prog/uniq.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
@c endfile
@end ignore
@c file eg/prog/uniq.awk

function usage()
@{
    print("Usage: uniq [-udc [-n]] [+n] [ in [ out ]]") > "/dev/stderr"
    exit 1
@}

# -c    count lines. overrides -d and -u
# -d    only repeated lines
# -u    only nonrepeated lines
# -n    skip n fields
# +n    skip n characters, skip fields first

BEGIN @{
    count = 1
    outputfile = "/dev/stdout"
    opts = "udc0:1:2:3:4:5:6:7:8:9:"
    while ((c = getopt(ARGC, ARGV, opts)) != -1) @{
        if (c == "u")
            non_repeated_only++
        else if (c == "d")
            repeated_only++
        else if (c == "c")
            do_count++
        else if (index("0123456789", c) != 0) @{
            # getopt() requires args to options
            # this messes us up for things like -5
            if (Optarg ~ /^[[:digit:]]+$/)
                fcount = (c Optarg) + 0
            else @{
                fcount = c + 0
                Optind--
            @}
        @} else
            usage()
    @}

@group
    if (ARGV[Optind] ~ /^\+[[:digit:]]+$/) @{
        charcount = substr(ARGV[Optind], 2) + 0
        Optind++
    @}
@end group

    for (i = 1; i < Optind; i++)
        ARGV[i] = ""

    if (repeated_only == 0 && non_repeated_only == 0)
        repeated_only = non_repeated_only = 1

    if (ARGC - Optind == 2) @{
        outputfile = ARGV[ARGC - 1]
        ARGV[ARGC - 1] = ""
    @}
@}
@c endfile
@end example

The following function, @code{are_equal()}, compares the current line,
@code{$0}, to the previous line, @code{last}.  It handles skipping fields
and characters.  If no field count and no character count are specified,
@code{are_equal()} returns one or zero depending upon the result of a
simple string comparison of @code{last} and @code{$0}.

Otherwise, things get more complicated.  If fields have to be skipped,
each line is broken into an array using @code{split()} (@pxref{String
Functions}); the desired fields are then joined back into a line
using @code{join()}.  The joined lines are stored in @code{clast} and
@code{cline}.  If no fields are skipped, @code{clast} and @code{cline}
are set to @code{last} and @code{$0}, respectively.  Finally, if
characters are skipped, @code{substr()} is used to strip off the leading
@code{charcount} characters in @code{clast} and @code{cline}.  The two
strings are then compared and @code{are_equal()} returns the result:

@example
@c file eg/prog/uniq.awk
@group
function are_equal(    n, m, clast, cline, alast, aline)
@{
    if (fcount == 0 && charcount == 0)
        return (last == $0)
@end group

    if (fcount > 0) @{
        n = split(last, alast)
        m = split($0, aline)
        clast = join(alast, fcount+1, n)
        cline = join(aline, fcount+1, m)
    @} else @{
        clast = last
        cline = $0
    @}
    if (charcount) @{
        clast = substr(clast, charcount + 1)
        cline = substr(cline, charcount + 1)
    @}
@group

    return (clast == cline)
@}
@end group
@c endfile
@end example

The following two rules are the body of the program.  The first one is
executed only for the very first line of data.  It sets @code{last} equal to
@code{$0}, so that subsequent lines of text have something to be compared to.

The second rule does the work. The variable @code{equal} is one or zero,
depending upon the results of @code{are_equal()}'s comparison. If @command{uniq}
is counting repeated lines, and the lines are equal, then it increments the @code{count} variable.
Otherwise, it prints the line and resets @code{count},
because the two lines are not equal.

If @command{uniq} is not counting, and if the lines are equal, @code{count} is incremented.
Nothing is printed, as the point is to remove duplicates.
Otherwise, if @command{uniq} is counting repeated lines and more than
one line is seen, or if @command{uniq} is counting nonrepeated lines
and only one line is seen, then the line is printed, and @code{count}
is reset.

Finally, similar logic is used in the @code{END} rule to print the final
line of input data:

@example
@c file eg/prog/uniq.awk
NR == 1 @{
    last = $0
    next
@}

@{
    equal = are_equal()

    if (do_count) @{    # overrides -d and -u
        if (equal)
            count++
        else @{
            printf("%4d %s\n", count, last) > outputfile
            last = $0
            count = 1    # reset
        @}
        next
    @}

    if (equal)
        count++
    else @{
        if ((repeated_only && count > 1) ||
            (non_repeated_only && count == 1))
                print last > outputfile
        last = $0
        count = 1
    @}
@}

END @{
    if (do_count)
        printf("%4d %s\n", count, last) > outputfile
@group
    else if ((repeated_only && count > 1) ||
            (non_repeated_only && count == 1))
        print last > outputfile
    close(outputfile)
@}
@end group
@c endfile
@end example

@c FIXME: Include this?
@ignore
This program does not follow our recommended convention of naming
global variables with a leading capital letter.  Doing that would
make the program a little easier to follow.
@end ignore

@ifset FOR_PRINT
The logic for choosing which lines to print represents a @dfn{state
machine}, which is ``a device that can be in one of a set number of stable
conditions depending on its previous condition and on the present values
of its inputs.''@footnote{This is the definition returned from entering
@code{define: state machine} into Google.}
Brian Kernighan suggests that
``an alternative approach to state machines is to just read
the input into an array, then use indexing.  It's almost always
easier code, and for most inputs where you would use this, just
as fast.''  Consider how to rewrite the logic to follow this
suggestion.
@end ifset



@node Wc Program
@subsection Counting Things

@c FIXME: One day, update to current POSIX version of wc

@cindex counting
@cindex input files, counting elements in
@cindex words, counting
@cindex characters, counting
@cindex lines, counting
@cindex @command{wc} utility
The @command{wc} (word count) utility counts lines, words, and characters in
one or more input files. Its usage is as follows:

@display
@command{wc} [@option{-lwc}] [@var{files} @dots{}]
@end display

If no files are specified on the command line, @command{wc} reads its standard
input. If there are multiple files, it also prints total counts for all
the files.  The options and their meanings are as follows:

@table @code
@item -l
Count only lines.

@item -w
Count only words.
A ``word'' is a contiguous sequence of nonwhitespace characters, separated
by spaces and/or TABs.  Luckily, this is the normal way @command{awk} separates
fields in its input data.

@item -c
Count only characters.
@end table

Implementing @command{wc} in @command{awk} is particularly elegant,
because @command{awk} does a lot of the work for us; it splits lines into
words (i.e., fields) and counts them, it counts lines (i.e., records),
and it can easily tell us how long a line is.

This program uses the @code{getopt()} library function
(@pxref{Getopt Function})
and the file-transition functions
(@pxref{Filetrans Function}).

This version has one notable difference from traditional versions of
@command{wc}: it always prints the counts in the order lines, words,
and characters.  Traditional versions note the order of the @option{-l},
@option{-w}, and @option{-c} options on the command line, and print the
counts in that order.

The @code{BEGIN} rule does the argument processing.  The variable
@code{print_total} is true if more than one file is named on the
command line:

@cindex @code{wc.awk} program
@example
@c file eg/prog/wc.awk
# wc.awk --- count lines, words, characters
@c endfile
@ignore
@c file eg/prog/wc.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
@c endfile
@end ignore
@c file eg/prog/wc.awk

# Options:
#    -l    only count lines
#    -w    only count words
#    -c    only count characters
#
# Default is to count lines, words, characters
#
# Requires getopt() and file transition library functions

BEGIN @{
    # let getopt() print a message about
    # invalid options. we ignore them
    while ((c = getopt(ARGC, ARGV, "lwc")) != -1) @{
        if (c == "l")
            do_lines = 1
        else if (c == "w")
            do_words = 1
        else if (c == "c")
            do_chars = 1
    @}
    for (i = 1; i < Optind; i++)
        ARGV[i] = ""

    # if no options, do all
    if (! do_lines && ! do_words && ! do_chars)
        do_lines = do_words = do_chars = 1

    print_total = (ARGC - i > 1)
@}
@c endfile
@end example

The @code{beginfile()} function is simple; it just resets the counts of lines,
words, and characters to zero, and saves the current @value{FN} in
@code{fname}:

@example
@c file eg/prog/wc.awk
function beginfile(file)
@{
    lines = words = chars = 0
    fname = FILENAME
@}
@c endfile
@end example

The @code{endfile()} function adds the current file's numbers to the
running totals of lines, words, and characters.  It then prints out those
numbers for the file that was just read. It relies on @code{beginfile()}
to reset the numbers for the following @value{DF}:

@example
@c file eg/prog/wc.awk
function endfile(file)
@{
    tlines += lines
    twords += words
    tchars += chars
    if (do_lines)
        printf "\t%d", lines
@group
    if (do_words)
        printf "\t%d", words
@end group
    if (do_chars)
        printf "\t%d", chars
    printf "\t%s\n", fname
@}
@c endfile
@end example

There is one rule that is executed for each line. It adds the length of
the record, plus one, to @code{chars}.@footnote{Because @command{gawk}
understands multibyte locales, this code counts characters, not bytes.}
Adding one plus the record length
is needed because the newline character separating records (the value
of @code{RS}) is not part of the record itself, and thus not included
in its length.  Next, @code{lines} is incremented for each line read,
and @code{words} is incremented by the value of @code{NF}, which is the
number of ``words'' on this line:

@example
@c file eg/prog/wc.awk
# do per line
@{
    chars += length($0) + 1    # get newline
    lines++
    words += NF
@}
@c endfile
@end example

Finally, the @code{END} rule simply prints the totals for all the files:

@example
@c file eg/prog/wc.awk
END @{
    if (print_total) @{
        if (do_lines)
            printf "\t%d", tlines
        if (do_words)
            printf "\t%d", twords
        if (do_chars)
            printf "\t%d", tchars
        print "\ttotal"
    @}
@}
@c endfile
@end example

@node Miscellaneous Programs
@section A Grab Bag of @command{awk} Programs

This @value{SECTION} is a large ``grab bag'' of miscellaneous programs.
We hope you find them both interesting and enjoyable.

@menu
* Dupword Program::             Finding duplicated words in a document.
* Alarm Program::               An alarm clock.
* Translate Program::           A program similar to the @command{tr} utility.
* Labels Program::              Printing mailing labels.
* Word Sorting::                A program to produce a word usage count.
* History Sorting::             Eliminating duplicate entries from a history
                                file.
* Extract Program::             Pulling out programs from Texinfo source
                                files.
* Simple Sed::                  A Simple Stream Editor.
* Igawk Program::               A wrapper for @command{awk} that includes
                                files.
* Anagram Program::             Finding anagrams from a dictionary.
* Signature Program::           People do amazing things with too much time on
                                their hands.
@end menu

@node Dupword Program
@subsection Finding Duplicated Words in a Document

@cindex words, duplicate@comma{} searching for
@cindex searching, for words
@cindex documents@comma{} searching
A common error when writing large amounts of prose is to accidentally
duplicate words.  Typically you will see this in text as something like ``the
the program does the following@dots{}''  When the text is online, often
the duplicated words occur at the end of one line and the
@iftex
the
@end iftex
beginning of
another, making them very difficult to spot.
@c as here!

This program, @file{dupword.awk}, scans through a file one line at a time
and looks for adjacent occurrences of the same word.  It also saves the last
word on a line (in the variable @code{prev}) for comparison with the first
word on the next line.

@cindex Texinfo
The first two statements make sure that the line is all lowercase,
so that, for example, ``The'' and ``the'' compare equal to each other.
The next statement replaces nonalphanumeric and nonwhitespace characters
with spaces, so that punctuation does not affect the comparison either.
The characters are replaced with spaces so that formatting controls
don't create nonsense words (e.g., the Texinfo @samp{@@code@{NF@}}
becomes @samp{codeNF} if punctuation is simply deleted).  The record is
then resplit into fields, yielding just the actual words on the line,
and ensuring that there are no empty fields.

If there are no fields left after removing all the punctuation, the
current record is skipped.  Otherwise, the program loops through each
word, comparing it to the previous one:

@cindex @code{dupword.awk} program
@example
@c file eg/prog/dupword.awk
# dupword.awk --- find duplicate words in text
@c endfile
@ignore
@c file eg/prog/dupword.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# December 1991
# Revised October 2000

@c endfile
@end ignore
@c file eg/prog/dupword.awk
@{
    $0 = tolower($0)
    gsub(/[^[:alnum:][:blank:]]/, " ");
    $0 = $0         # re-split
    if (NF == 0)
        next
    if ($1 == prev)
        printf("%s:%d: duplicate %s\n",
            FILENAME, FNR, $1)
    for (i = 2; i <= NF; i++)
        if ($i == $(i-1))
            printf("%s:%d: duplicate %s\n",
                FILENAME, FNR, $i)
    prev = $NF
@}
@c endfile
@end example

@node Alarm Program
@subsection An Alarm Clock Program
@cindex insomnia, cure for
@cindex Robbins, Arnold
@quotation
@i{Nothing cures insomnia like a ringing alarm clock.}
@author Arnold Robbins
@end quotation
@cindex Quanstrom, Erik
@ignore
Date: Sat, 15 Feb 2014 16:47:09 -0500
Subject: Re: 9atom install question
Message-ID: <l2jcvx6j6mey60xnrkb0hhob.1392500829294@email.android.com>
From: Erik Quanstrom <quanstro@quanstro.net>
To: Aharon Robbins <arnold@skeeve.com>

yes.

- erik

Aharon Robbins <arnold@skeeve.com> wrote:

>> sleep is for web developers.
>
>Can I quote you, in the gawk manual?
>
>Thanks,
>
>Arnold
@end ignore
@quotation
@i{Sleep is for web developers.}
@author Erik Quanstrom
@end quotation

@cindex time, alarm clock example program
@cindex alarm clock example program
The following program is a simple ``alarm clock'' program.
You give it a time of day and an optional message.  At the specified time,
it prints the message on the standard output. In addition, you can give it
the number of times to repeat the message as well as a delay between
repetitions.

This program uses the @code{getlocaltime()} function from
@ref{Getlocaltime Function}.

All the work is done in the @code{BEGIN} rule.  The first part is argument
checking and setting of defaults: the delay, the count, and the message to
print.  If the user supplied a message without the ASCII BEL
character (known as the ``alert'' character, @code{"\a"}), then it is added to
the message.  (On many systems, printing the ASCII BEL generates an
audible alert. Thus, when the alarm goes off, the system calls attention
to itself in case the user is not looking at the computer.)
Just for a change, this program uses a @code{switch} statement
(@pxref{Switch Statement}), but the processing could be done with a series of
@code{if}-@code{else} statements instead.
Here is the program:

@cindex @code{alarm.awk} program
@example
@c file eg/prog/alarm.awk
# alarm.awk --- set an alarm
#
# Requires getlocaltime() library function
@c endfile
@ignore
@c file eg/prog/alarm.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised December 2010

@c endfile
@end ignore
@c file eg/prog/alarm.awk
# usage: alarm time [ "message" [ count [ delay ] ] ]

BEGIN @{
    # Initial argument sanity checking
    usage1 = "usage: alarm time ['message' [count [delay]]]"
    usage2 = sprintf("\t(%s) time ::= hh:mm", ARGV[1])

    if (ARGC < 2) @{
        print usage1 > "/dev/stderr"
        print usage2 > "/dev/stderr"
        exit 1
    @}
    switch (ARGC) @{
    case 5:
        delay = ARGV[4] + 0
        # fall through
    case 4:
        count = ARGV[3] + 0
        # fall through
    case 3:
        message = ARGV[2]
        break
    default:
        if (ARGV[1] !~ /[[:digit:]]?[[:digit:]]:[[:digit:]]@{2@}/) @{
            print usage1 > "/dev/stderr"
            print usage2 > "/dev/stderr"
            exit 1
        @}
        break
    @}

    # set defaults for once we reach the desired time
    if (delay == 0)
        delay = 180    # 3 minutes
@group
    if (count == 0)
        count = 5
@end group
    if (message == "")
        message = sprintf("\aIt is now %s!\a", ARGV[1])
    else if (index(message, "\a") == 0)
        message = "\a" message "\a"
@c endfile
@end example

The next @value{SECTION} of code turns the alarm time into hours and minutes,
converts it (if necessary) to a 24-hour clock, and then turns that
time into a count of the seconds since midnight.  Next it turns the current
time into a count of seconds since midnight.  The difference between the two
is how long to wait before setting off the alarm:

@example
@c file eg/prog/alarm.awk
    # split up alarm time
    split(ARGV[1], atime, ":")
    hour = atime[1] + 0    # force numeric
    minute = atime[2] + 0  # force numeric

    # get current broken down time
    getlocaltime(now)

    # if time given is 12-hour hours and it's after that
    # hour, e.g., `alarm 5:30' at 9 a.m. means 5:30 p.m.,
    # then add 12 to real hour
    if (hour < 12 && now["hour"] > hour)
        hour += 12

    # set target time in seconds since midnight
    target = (hour * 60 * 60) + (minute * 60)

    # get current time in seconds since midnight
    current = (now["hour"] * 60 * 60) + \
               (now["minute"] * 60) + now["second"]

    # how long to sleep for
    naptime = target - current
    if (naptime <= 0) @{
        print "alarm: time is in the past!" > "/dev/stderr"
        exit 1
    @}
@c endfile
@end example

@cindex @command{sleep} utility
Finally, the program uses the @code{system()} function
(@pxref{I/O Functions})
to call the @command{sleep} utility.  The @command{sleep} utility simply pauses
for the given number of seconds.  If the exit status is not zero,
the program assumes that @command{sleep} was interrupted and exits. If
@command{sleep} exited with an OK status (zero), then the program prints the
message in a loop, again using @command{sleep} to delay for however many
seconds are necessary:

@example
@c file eg/prog/alarm.awk
    # zzzzzz..... go away if interrupted
    if (system(sprintf("sleep %d", naptime)) != 0)
        exit 1

    # time to notify!
    command = sprintf("sleep %d", delay)
    for (i = 1; i <= count; i++) @{
        print message
        # if sleep command interrupted, go away
        if (system(command) != 0)
            break
    @}

    exit 0
@}
@c endfile
@end example

@node Translate Program
@subsection Transliterating Characters

@cindex characters, transliterating
@cindex @command{tr} utility
The system @command{tr} utility transliterates characters.  For example, it is
often used to map uppercase letters into lowercase for further processing:

@example
@var{generate data} | tr 'A-Z' 'a-z' | @var{process data} @dots{}
@end example

@command{tr} requires two lists of characters.@footnote{On some older
systems, including Solaris, the system version of @command{tr} may require
that the lists be written as range expressions enclosed in square brackets
(@samp{[a-z]}) and quoted, to prevent the shell from attempting a
@value{FN} expansion.  This is not a feature.}  When processing the input, the
first character in the first list is replaced with the first character
in the second list, the second character in the first list is replaced
with the second character in the second list, and so on.  If there are
more characters in the ``from'' list than in the ``to'' list, the last
character of the ``to'' list is used for the remaining characters in the
``from'' list.

Once upon a time,
@c early or mid-1989!
a user proposed adding a transliteration function
to @command{gawk}.
@c Wishing to avoid gratuitous new features,
@c at least theoretically
The following program was written to
prove that character transliteration could be done with a user-level
function.  This program is not as complete as the system @command{tr} utility,
but it does most of the job.

The @command{translate} program was written long before @command{gawk}
acquired the ability to split each character in a string into separate
array elements.  Thus, it makes repeated use of the @code{substr()},
@code{index()}, and @code{gsub()} built-in functions (@pxref{String
Functions}).  There are two functions.  The first, @code{stranslate()},
takes three arguments:

@table @code
@item from
A list of characters from which to translate

@item to
A list of characters to which to translate

@item target
The string on which to do the translation
@end table

Associative arrays make the translation part fairly easy. @code{t_ar} holds
the ``to'' characters, indexed by the ``from'' characters.  Then a simple
loop goes through @code{from}, one character at a time.  For each character
in @code{from}, if the character appears in @code{target},
it is replaced with the corresponding @code{to} character.

The @code{translate()} function calls @code{stranslate()}, using @code{$0}
as the target.  The main program sets two global variables, @code{FROM} and
@code{TO}, from the command line, and then changes @code{ARGV} so that
@command{awk} reads from the standard input.

Finally, the processing rule simply calls @code{translate()} for each record:

@cindex @code{translate.awk} program
@example
@c file eg/prog/translate.awk
# translate.awk --- do tr-like stuff
@c endfile
@ignore
@c file eg/prog/translate.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# August 1989
# February 2009 - bug fix

@c endfile
@end ignore
@c file eg/prog/translate.awk
# Bugs: does not handle things like tr A-Z a-z; it has
# to be spelled out. However, if `to' is shorter than `from',
# the last character in `to' is used for the rest of `from'.

function stranslate(from, to, target,     lf, lt, ltarget, t_ar, i, c,
                                                               result)
@{
    lf = length(from)
    lt = length(to)
    ltarget = length(target)
    for (i = 1; i <= lt; i++)
        t_ar[substr(from, i, 1)] = substr(to, i, 1)
    if (lt < lf)
        for (; i <= lf; i++)
            t_ar[substr(from, i, 1)] = substr(to, lt, 1)
    for (i = 1; i <= ltarget; i++) @{
        c = substr(target, i, 1)
        if (c in t_ar)
            c = t_ar[c]
        result = result c
    @}
    return result
@}

function translate(from, to)
@{
    return $0 = stranslate(from, to, $0)
@}

# main program
BEGIN @{
@group
    if (ARGC < 3) @{
        print "usage: translate from to" > "/dev/stderr"
        exit
    @}
@end group
    FROM = ARGV[1]
    TO = ARGV[2]
    ARGC = 2
    ARGV[1] = "-"
@}

@{
    translate(FROM, TO)
    print
@}
@c endfile
@end example

It is possible to do character transliteration in a user-level
function, but it is not necessarily efficient, and we (the @command{gawk}
developers) started to consider adding a built-in function.  However,
shortly after writing this program, we learned that Brian Kernighan
had added the @code{toupper()} and @code{tolower()} functions to his
@command{awk} (@pxref{String Functions}).  These functions handle the
vast majority of the cases where character transliteration is necessary,
and so we chose to simply add those functions to @command{gawk} as well
and then leave well enough alone.

An obvious improvement to this program would be to set up the
@code{t_ar} array only once, in a @code{BEGIN} rule. However, this
assumes that the ``from'' and ``to'' lists
will never change throughout the lifetime of the program.

Another obvious improvement is to enable the use of ranges,
such as @samp{a-z}, as allowed by the @command{tr} utility.
Look at the code for @file{cut.awk} (@pxref{Cut Program})
for inspiration.


@node Labels Program
@subsection Printing Mailing Labels

@cindex printing, mailing labels
@cindex mailing labels@comma{} printing
Here is a ``real-world''@footnote{``Real world'' is defined as
``a program actually used to get something done.''}
program.  This
script reads lists of names and
addresses and generates mailing labels.  Each page of labels has 20 labels
on it, two across and 10 down.  The addresses are guaranteed to be no more
than five lines of data.  Each address is separated from the next by a blank
line.

The basic idea is to read 20 labels' worth of data.  Each line of each label
is stored in the @code{line} array.  The single rule takes care of filling
the @code{line} array and printing the page when 20 labels have been read.

The @code{BEGIN} rule simply sets @code{RS} to the empty string, so that
@command{awk} splits records at blank lines
(@pxref{Records}).
It sets @code{MAXLINES} to 100, because 100 is the maximum number
of lines on the page
@iftex
(@math{20 @cdot 5 = 100}).
@end iftex
@ifnottex
@ifnotdocbook
(20 * 5 = 100).
@end ifnotdocbook
@end ifnottex
@docbook
(20 &sdot; 5 = 100).
@end docbook

Most of the work is done in the @code{printpage()} function.
The label lines are stored sequentially in the @code{line} array.  But they
have to print horizontally: @code{line[1]} next to @code{line[6]},
@code{line[2]} next to @code{line[7]}, and so on.  Two loops
accomplish this.  The outer loop, controlled by @code{i}, steps through
every 10 lines of data; this is each row of labels.  The inner loop,
controlled by @code{j}, goes through the lines within the row.
As @code{j} goes from 0 to 4, @samp{i+j} is the @code{j}th line in
the row, and @samp{i+j+5} is the entry next to it.  The output ends up
looking something like this:

@example
line 1          line 6
line 2          line 7
line 3          line 8
line 4          line 9
line 5          line 10
@dots{}
@end example

@noindent
The @code{printf} format string @samp{%-41s} left-aligns
the data and prints it within a fixed-width field.

As a final note, an extra blank line is printed at lines 21 and 61, to keep
the output lined up on the labels.  This is dependent on the particular
brand of labels in use when the program was written.  You will also note
that there are two blank lines at the top and two blank lines at the bottom.

The @code{END} rule arranges to flush the final page of labels; there may
not have been an even multiple of 20 labels in the data:

@cindex @code{labels.awk} program
@example
@c file eg/prog/labels.awk
# labels.awk --- print mailing labels
@c endfile
@ignore
@c file eg/prog/labels.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# June 1992
# December 2010, minor edits
@c endfile
@end ignore
@c file eg/prog/labels.awk

# Each label is 5 lines of data that may have blank lines.
# The label sheets have 2 blank lines at the top and 2 at
# the bottom.

BEGIN    @{ RS = "" ; MAXLINES = 100 @}

function printpage(    i, j)
@{
    if (Nlines <= 0)
        return

    printf "\n\n"        # header

    for (i = 1; i <= Nlines; i += 10) @{
        if (i == 21 || i == 61)
            print ""
        for (j = 0; j < 5; j++) @{
            if (i + j > MAXLINES)
                break
            printf "   %-41s %s\n", line[i+j], line[i+j+5]
        @}
        print ""
    @}

    printf "\n\n"        # footer

    delete line
@}

# main rule
@{
    if (Count >= 20) @{
        printpage()
        Count = 0
        Nlines = 0
    @}
    n = split($0, a, "\n")
    for (i = 1; i <= n; i++)
        line[++Nlines] = a[i]
    for (; i <= 5; i++)
        line[++Nlines] = ""
    Count++
@}

END @{
    printpage()
@}
@c endfile
@end example

@node Word Sorting
@subsection Generating Word-Usage Counts

@cindex words, usage counts@comma{} generating

When working with large amounts of text, it can be interesting to know
how often different words appear.  For example, an author may overuse
certain words, in which case he or she might wish to find synonyms to substitute
for words that appear too often. This @value{SUBSECTION} develops a
program for counting words and presenting the frequency information
in a useful format.

At first glance, a program like this would seem to do the job:

@example
# wordfreq-first-try.awk --- print list of word frequencies

@{
    for (i = 1; i <= NF; i++)
        freq[$i]++
@}

@group
END @{
    for (word in freq)
        printf "%s\t%d\n", word, freq[word]
@}
@end group
@end example

The program relies on @command{awk}'s default field-splitting
mechanism to break each line up into ``words'' and uses an
associative array named @code{freq}, indexed by each word, to count
the number of times the word occurs. In the @code{END} rule,
it prints the counts.

This program has several problems that prevent it from being
useful on real text files:

@itemize @value{BULLET}
@item
The @command{awk} language considers upper- and lowercase characters to be
distinct.  Therefore, ``bartender'' and ``Bartender'' are not treated
as the same word.  This is undesirable, because words are capitalized
if they begin sentences in normal text, and a frequency analyzer should
not be sensitive to capitalization.

@item
Words are detected using the @command{awk} convention that fields are
separated just by whitespace.  Other characters in the input (except
newlines) don't have any special meaning to @command{awk}.  This means that
punctuation characters count as part of words.

@item
The output does not come out in any useful order.  You're more likely to be
interested in which words occur most frequently or in having an alphabetized
table of how frequently each word occurs.
@end itemize

@cindex @command{sort} utility
The first problem can be solved by using @code{tolower()} to remove case
distinctions.  The second problem can be solved by using @code{gsub()}
to remove punctuation characters.  Finally, we solve the third problem
by using the system @command{sort} utility to process the output of the
@command{awk} script.  Here is the new version of the program:

@cindex @code{wordfreq.awk} program
@example
@c file eg/prog/wordfreq.awk
# wordfreq.awk --- print list of word frequencies

@{
    $0 = tolower($0)    # remove case distinctions
    # remove punctuation
    gsub(/[^[:alnum:]_[:blank:]]/, "", $0)
    for (i = 1; i <= NF; i++)
        freq[$i]++
@}

@c endfile
END @{
    for (word in freq)
        printf "%s\t%d\n", word, freq[word]
@}
@end example

The regexp @code{/[^[:alnum:]_[:blank:]]/} might have been written
@code{/[[:punct:]]/}, but then underscores would also be removed,
and we want to keep them.

Assuming we have saved this program in a file named @file{wordfreq.awk},
and that the data is in @file{file1}, the following pipeline:

@example
awk -f wordfreq.awk file1 | sort -k 2nr
@end example

@noindent
produces a table of the words appearing in @file{file1} in order of
decreasing frequency.

The @command{awk} program suitably massages the
data and produces a word frequency table, which is not ordered.
The @command{awk} script's output is then sorted by the @command{sort}
utility and printed on the screen.

The options given to @command{sort}
specify a sort that uses the second field of each input line (skipping
one field), that the sort keys should be treated as numeric quantities
(otherwise @samp{15} would come before @samp{5}), and that the sorting
should be done in descending (reverse) order.

The @command{sort} could even be done from within the program, by changing
the @code{END} action to:

@example
@c file eg/prog/wordfreq.awk
END @{
    sort = "sort -k 2nr"
    for (word in freq)
        printf "%s\t%d\n", word, freq[word] | sort
    close(sort)
@}
@c endfile
@end example

This way of sorting must be used on systems that do not
have true pipes at the command-line (or batch-file) level.
See the general operating system documentation for more information on how
to use the @command{sort} program.

@node History Sorting
@subsection Removing Duplicates from Unsorted Text

@cindex lines, duplicate@comma{} removing
The @command{uniq} program
(@pxref{Uniq Program})
removes duplicate lines from @emph{sorted} data.

Suppose, however, you need to remove duplicate lines from a @value{DF} but
that you want to preserve the order the lines are in.  A good example of
this might be a shell history file.  The history file keeps a copy of all
the commands you have entered, and it is not unusual to repeat a command
several times in a row.  Occasionally you might want to compact the history
by removing duplicate entries.  Yet it is desirable to maintain the order
of the original commands.

This simple program does the job.  It uses two arrays.  The @code{data}
array is indexed by the text of each line.
For each line, @code{data[$0]} is incremented.
If a particular line has not
been seen before, then @code{data[$0]} is zero.
In this case, the text of the line is stored in @code{lines[count]}.
Each element of @code{lines} is a unique command, and the indices of
@code{lines} indicate the order in which those lines are encountered.
The @code{END} rule simply prints out the lines, in order:

@cindex Rakitzis, Byron
@cindex @code{histsort.awk} program
@example
@c file eg/prog/histsort.awk
# histsort.awk --- compact a shell history file
# Thanks to Byron Rakitzis for the general idea
@c endfile
@ignore
@c file eg/prog/histsort.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
@c endfile
@end ignore
@c file eg/prog/histsort.awk

@group
@{
    if (data[$0]++ == 0)
        lines[++count] = $0
@}
@end group

@group
END @{
    for (i = 1; i <= count; i++)
        print lines[i]
@}
@end group
@c endfile
@end example

This program also provides a foundation for generating other useful
information.  For example, using the following @code{print} statement in the
@code{END} rule indicates how often a particular command is used:

@example
print data[lines[i]], lines[i]
@end example

@noindent
This works because @code{data[$0]} is incremented each time a line is
seen.

@node Extract Program
@subsection Extracting Programs from Texinfo Source Files

@cindex Texinfo, extracting programs from source files
@cindex files, Texinfo@comma{} extracting programs from
@ifnotinfo
Both this chapter and the previous chapter
(@ref{Library Functions})
present a large number of @command{awk} programs.
@end ifnotinfo
@ifinfo
The nodes
@ref{Library Functions},
and @ref{Sample Programs},
are the top level nodes for a large number of @command{awk} programs.
@end ifinfo
If you want to experiment with these programs, it is tedious to type
them in by hand.  Here we present a program that can extract parts of a
Texinfo input file into separate files.

@cindex Texinfo
This @value{DOCUMENT} is written in @uref{https://www.gnu.org/software/texinfo/, Texinfo},
the GNU Project's document formatting language.
A single Texinfo source file can be used to produce both
printed documentation, with @TeX{}, and online documentation.
@ifnotinfo
(Texinfo is fully documented in the book
@cite{Texinfo---The GNU Documentation Format},
available from the Free Software Foundation,
and also available @uref{https://www.gnu.org/software/texinfo/manual/texinfo/, online}.)
@end ifnotinfo
@ifinfo
(The Texinfo language is described fully, starting with
@inforef{Top, , Texinfo, texinfo,Texinfo---The GNU Documentation Format}.)
@end ifinfo

For our purposes, it is enough to know three things about Texinfo input
files:

@itemize @value{BULLET}
@item
The ``at'' symbol (@samp{@@}) is special in Texinfo, much as
the backslash (@samp{\}) is in C
or @command{awk}.  Literal @samp{@@} symbols are represented in Texinfo source
files as @samp{@@@@}.

@item
Comments start with either @samp{@@c} or @samp{@@comment}.
The file-extraction program works by using special comments that start
at the beginning of a line.

@item
Lines containing @samp{@@group} and @samp{@@end group} commands bracket
example text that should not be split across a page boundary.
(Unfortunately, @TeX{} isn't always smart enough to do things exactly right,
so we have to give it some help.)
@end itemize

The following program, @file{extract.awk}, reads through a Texinfo source
file and does two things, based on the special comments.
Upon seeing @samp{@w{@@c system @dots{}}},
it runs a command, by extracting the command text from the
control line and passing it on to the @code{system()} function
(@pxref{I/O Functions}).
Upon seeing @samp{@@c file @var{filename}}, each subsequent line is sent to
the file @var{filename}, until @samp{@@c endfile} is encountered.
The rules in @file{extract.awk} match either @samp{@@c} or
@samp{@@comment} by letting the @samp{omment} part be optional.
Lines containing @samp{@@group} and @samp{@@end group} are simply removed.
@file{extract.awk} uses the @code{join()} library function
(@pxref{Join Function}).

The example programs in the online Texinfo source for @cite{@value{TITLE}}
(@file{gawktexi.in}) have all been bracketed inside @samp{file} and
@samp{endfile} lines.  The @command{gawk} distribution uses a copy of
@file{extract.awk} to extract the sample programs and install many
of them in a standard directory where @command{gawk} can find them.
The Texinfo file looks something like this:

@example
@dots{}
This program has a @@code@{BEGIN@} rule
that prints a nice message:

@@example
@@c file examples/messages.awk
BEGIN @@@{ print "Don't panic!" @@@}
@@c endfile
@@end example

It also prints some final advice:

@@example
@@c file examples/messages.awk
END @@@{ print "Always avoid bored archaeologists!" @@@}
@@c endfile
@@end example
@dots{}
@end example

@file{extract.awk} begins by setting @code{IGNORECASE} to one, so that
mixed upper- and lowercase letters in the directives won't matter.

The first rule handles calling @code{system()}, checking that a command is
given (@code{NF} is at least three) and also checking that the command
exits with a zero exit status, signifying OK:

@cindex @code{extract.awk} program
@example
@c file eg/prog/extract.awk
# extract.awk --- extract files and run programs from Texinfo files
@c endfile
@ignore
@c file eg/prog/extract.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# May 1993
# Revised September 2000
@c endfile
@end ignore
@c file eg/prog/extract.awk

BEGIN    @{ IGNORECASE = 1 @}

/^@@c(omment)?[ \t]+system/ @{
    if (NF < 3) @{
        e = ("extract: " FILENAME ":" FNR)
        e = (e  ": badly formed `system' line")
        print e > "/dev/stderr"
        next
    @}
    $1 = ""
    $2 = ""
    stat = system($0)
    if (stat != 0) @{
        e = ("extract: " FILENAME ":" FNR)
        e = (e ": warning: system returned " stat)
        print e > "/dev/stderr"
    @}
@}
@c endfile
@end example

@noindent
The variable @code{e} is used so that the rule
fits nicely on the @value{PAGE}.

The second rule handles moving data into files.  It verifies that a
@value{FN} is given in the directive.  If the file named is not the
current file, then the current file is closed.  Keeping the current file
open until a new file is encountered allows the use of the @samp{>}
redirection for printing the contents, keeping open-file management
simple.

The @code{for} loop does the work.  It reads lines using @code{getline}
(@pxref{Getline}).
For an unexpected end-of-file, it calls the @code{@w{unexpected_eof()}}
function.  If the line is an ``endfile'' line, then it breaks out of
the loop.
If the line is an @samp{@@group} or @samp{@@end group} line, then it
ignores it and goes on to the next line.
Similarly, comments within examples are also ignored.

Most of the work is in the following few lines.  If the line has no @samp{@@}
symbols, the program can print it directly.
Otherwise, each leading @samp{@@} must be stripped off.
To remove the @samp{@@} symbols, the line is split into separate elements of
the array @code{a}, using the @code{split()} function
(@pxref{String Functions}).
The @samp{@@} symbol is used as the separator character.
Each element of @code{a} that is empty indicates two successive @samp{@@}
symbols in the original line.  For each two empty elements (@samp{@@@@} in
the original file), we have to add a single @samp{@@} symbol back in.

When the processing of the array is finished, @code{join()} is called with the
value of @code{SUBSEP} (@pxref{Multidimensional}),
to rejoin the pieces back into a single
line.  That line is then printed to the output file:

@example
@c file eg/prog/extract.awk
/^@@c(omment)?[ \t]+file/ @{
    if (NF != 3) @{
        e = ("extract: " FILENAME ":" FNR ": badly formed `file' line")
        print e > "/dev/stderr"
        next
    @}
    if ($3 != curfile) @{
        if (curfile != "")
            close(curfile)
        curfile = $3
    @}

    for (;;) @{
        if ((getline line) <= 0)
            unexpected_eof()
        if (line ~ /^@@c(omment)?[ \t]+endfile/)
            break
        else if (line ~ /^@@(end[ \t]+)?group/)
            continue
        else if (line ~ /^@@c(omment+)?[ \t]+/)
            continue
        if (index(line, "@@") == 0) @{
            print line > curfile
            continue
        @}
        n = split(line, a, "@@")
        # if a[1] == "", means leading @@,
        # don't add one back in.
        for (i = 2; i <= n; i++) @{
            if (a[i] == "") @{ # was an @@@@
                a[i] = "@@"
                if (a[i+1] == "")
                    i++
            @}
        @}
@group
        print join(a, 1, n, SUBSEP) > curfile
    @}
@}
@end group
@c endfile
@end example

An important thing to note is the use of the @samp{>} redirection.
Output done with @samp{>} only opens the file once; it stays open and
subsequent output is appended to the file
(@pxref{Redirection}).
This makes it easy to mix program text and explanatory prose for the same
sample source file (as has been done here!) without any hassle.  The file is
only closed when a new @value{DF} name is encountered or at the end of the
input file.

Finally, the function @code{@w{unexpected_eof()}} prints an appropriate
error message and then exits.
The @code{END} rule handles the final cleanup, closing the open file:

@example
@c file eg/prog/extract.awk
@group
function unexpected_eof()
@{
    printf("extract: %s:%d: unexpected EOF or error\n",
                     FILENAME, FNR) > "/dev/stderr"
    exit 1
@}
@end group

END @{
    if (curfile)
        close(curfile)
@}
@c endfile
@end example

@node Simple Sed
@subsection A Simple Stream Editor

@cindex @command{sed} utility
@cindex stream editors
The @command{sed} utility is a @dfn{stream editor}, a program that reads a
stream of data, makes changes to it, and passes it on.
It is often used to make global changes to a large file or to a stream
of data generated by a pipeline of commands.
Although @command{sed} is a complicated program in its own right, its most common
use is to perform global substitutions in the middle of a pipeline:

@example
@var{command1} < orig.data | sed 's/old/new/g' | @var{command2} > result
@end example

Here, @samp{s/old/new/g} tells @command{sed} to look for the regexp
@samp{old} on each input line and globally replace it with the text
@samp{new} (i.e., all the occurrences on a line).  This is similar to
@command{awk}'s @code{gsub()} function
(@pxref{String Functions}).

The following program, @file{awksed.awk}, accepts at least two command-line
arguments: the pattern to look for and the text to replace it with. Any
additional arguments are treated as @value{DF} names to process. If none
are provided, the standard input is used:

@cindex Brennan, Michael
@cindex @command{awksed.awk} program
@c @cindex simple stream editor
@c @cindex stream editor, simple
@example
@c file eg/prog/awksed.awk
# awksed.awk --- do s/foo/bar/g using just print
#    Thanks to Michael Brennan for the idea
@c endfile
@ignore
@c file eg/prog/awksed.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# August 1995
@c endfile
@end ignore
@c file eg/prog/awksed.awk

function usage()
@{
    print "usage: awksed pat repl [files...]" > "/dev/stderr"
    exit 1
@}

@group
BEGIN @{
    # validate arguments
    if (ARGC < 3)
        usage()
@end group

    RS = ARGV[1]
    ORS = ARGV[2]

    # don't use arguments as files
    ARGV[1] = ARGV[2] = ""
@}

@group
# look ma, no hands!
@{
    if (RT == "")
        printf "%s", $0
    else
        print
@}
@end group
@c endfile
@end example

The program relies on @command{gawk}'s ability to have @code{RS} be a regexp,
as well as on the setting of @code{RT} to the actual text that terminates the
record (@pxref{Records}).

The idea is to have @code{RS} be the pattern to look for. @command{gawk}
automatically sets @code{$0} to the text between matches of the pattern.
This is text that we want to keep, unmodified.  Then, by setting @code{ORS}
to the replacement text, a simple @code{print} statement outputs the
text we want to keep, followed by the replacement text.

There is one wrinkle to this scheme, which is what to do if the last record
doesn't end with text that matches @code{RS}.  Using a @code{print}
statement unconditionally prints the replacement text, which is not correct.
However, if the file did not end in text that matches @code{RS}, @code{RT}
is set to the null string.  In this case, we can print @code{$0} using
@code{printf}
(@pxref{Printf}).

The @code{BEGIN} rule handles the setup, checking for the right number
of arguments and calling @code{usage()} if there is a problem. Then it sets
@code{RS} and @code{ORS} from the command-line arguments and sets
@code{ARGV[1]} and @code{ARGV[2]} to the null string, so that they are
not treated as @value{FN}s
(@pxref{ARGC and ARGV}).

The @code{usage()} function prints an error message and exits.
Finally, the single rule handles the printing scheme outlined earlier,
using @code{print} or @code{printf} as appropriate, depending upon the
value of @code{RT}.

@node Igawk Program
@subsection An Easy Way to Use Library Functions

@cindex libraries of @command{awk} functions, example program for using
@cindex functions, library, example program for using
In @ref{Include Files}, we saw how @command{gawk} provides a built-in
file-inclusion capability.  However, this is a @command{gawk} extension.
This @value{SECTION} provides the motivation for making file inclusion
available for standard @command{awk}, and shows how to do it using a
combination of shell and @command{awk} programming.

Using library functions in @command{awk} can be very beneficial. It
encourages code reuse and the writing of general functions. Programs are
smaller and therefore clearer.
However, using library functions is only easy when writing @command{awk}
programs; it is painful when running them, requiring multiple @option{-f}
options.  If @command{gawk} is unavailable, then so too is the @env{AWKPATH}
environment variable and the ability to put @command{awk} functions into a
library directory (@pxref{Options}).
It would be nice to be able to write programs in the following manner:

@example
# library functions
@@include getopt.awk
@@include join.awk
@dots{}

# main program
BEGIN @{
    while ((c = getopt(ARGC, ARGV, "a:b:cde")) != -1)
        @dots{}
    @dots{}
@}
@end example

The following program, @file{igawk.sh}, provides this service.
It simulates @command{gawk}'s searching of the @env{AWKPATH} variable
and also allows @dfn{nested} includes (i.e., a file that is included
with @code{@@include} can contain further @code{@@include} statements).
@command{igawk} makes an effort to only include files once, so that nested
includes don't accidentally include a library function twice.

@command{igawk} should behave just like @command{gawk} externally.  This
means it should accept all of @command{gawk}'s command-line arguments,
including the ability to have multiple source files specified via
@option{-f} and the ability to mix command-line and library source files.

The program is written using the POSIX Shell (@command{sh}) command
language.@footnote{Fully explaining the @command{sh} language is beyond
the scope of this book. We provide some minimal explanations, but see
a good shell programming book if you wish to understand things in more
depth.} It works as follows:

@enumerate
@item
Loop through the arguments, saving anything that doesn't represent
@command{awk} source code for later, when the expanded program is run.

@item
For any arguments that do represent @command{awk} text, put the arguments into
a shell variable that will be expanded.  There are two cases:

@enumerate a
@item
Literal text, provided with @option{-e} or @option{--source}.  This
text is just appended directly.

@item
Source @value{FN}s, provided with @option{-f}.  We use a neat trick and
append @samp{@@include @var{filename}} to the shell variable's contents.
Because the file-inclusion program works the way @command{gawk} does, this
gets the text of the file included in the program at the correct point.
@end enumerate

@item
Run an @command{awk} program (naturally) over the shell variable's contents to expand
@code{@@include} statements.  The expanded program is placed in a second
shell variable.

@item
Run the expanded program with @command{gawk} and any other original command-line
arguments that the user supplied (such as the @value{DF} names).
@end enumerate

This program uses shell variables extensively: for storing command-line arguments and
the text of the @command{awk} program that will expand the user's program, for the
user's original program, and for the expanded program.  Doing so removes some
potential problems that might arise were we to use temporary files instead,
at the cost of making the script somewhat more complicated.

The initial part of the program turns on shell tracing if the first
argument is @samp{debug}.

The next part loops through all the command-line arguments.
There are several cases of interest:

@c @asis for docbook
@table @asis
@item @option{--}
This ends the arguments to @command{igawk}.  Anything else should be passed on
to the user's @command{awk} program without being evaluated.

@item @option{-W}
This indicates that the next option is specific to @command{gawk}.  To make
argument processing easier, the @option{-W} is appended to the front of the
remaining arguments and the loop continues.  (This is an @command{sh}
programming trick.  Don't worry about it if you are not familiar with
@command{sh}.)

@item @option{-v}, @option{-F}
These are saved and passed on to @command{gawk}.

@item @option{-f}, @option{--file}, @option{--file=}, @option{-Wfile=}
The @value{FN} is appended to the shell variable @code{program} with an
@code{@@include} statement.
The @command{expr} utility is used to remove the leading option part of the
argument (e.g., @samp{--file=}).
(Typical @command{sh} usage would be to use the @command{echo} and @command{sed}
utilities to do this work.  Unfortunately, some versions of @command{echo} evaluate
escape sequences in their arguments, possibly mangling the program text.
Using @command{expr} avoids this problem.)

@item @option{--source}, @option{--source=}, @option{-Wsource=}
The source text is appended to @code{program}.

@item @option{--version}, @option{-Wversion}
@command{igawk} prints its version number, runs @samp{gawk --version}
to get the @command{gawk} version information, and then exits.
@end table

If none of the @option{-f}, @option{--file}, @option{-Wfile}, @option{--source},
or @option{-Wsource} arguments are supplied, then the first nonoption argument
should be the @command{awk} program.  If there are no command-line
arguments left, @command{igawk} prints an error message and exits.
Otherwise, the first argument is appended to @code{program}.
In any case, after the arguments have been processed,
the shell variable
@code{program} contains the complete text of the original @command{awk}
program.

The program is as follows:

@cindex @code{igawk.sh} program
@example
@c file eg/prog/igawk.sh
#! /bin/sh
# igawk --- like gawk but do @@include processing
@c endfile
@ignore
@c file eg/prog/igawk.sh
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# July 1993
# December 2010, minor edits
@c endfile
@end ignore
@c file eg/prog/igawk.sh

if [ "$1" = debug ]
then
    set -x
    shift
fi

# A literal newline, so that program text is formatted correctly
n='
'

# Initialize variables to empty
program=
opts=

while [ $# -ne 0 ] # loop over arguments
do
    case $1 in
    --)     shift
            break ;;

    -W)     shift
            # The $@{x?'message here'@} construct prints a
            # diagnostic if $x is the null string
            set -- -W"$@{@@?'missing operand'@}"
            continue ;;

    -[vF])  opts="$opts $1 '$@{2?'missing operand'@}'"
            shift ;;

    -[vF]*) opts="$opts '$1'" ;;

    -f)     program="$program$n@@include $@{2?'missing operand'@}"
            shift ;;

    -f*)    f=$(expr "$1" : '-f\(.*\)')
            program="$program$n@@include $f" ;;

    -[W-]file=*)
            f=$(expr "$1" : '-.file=\(.*\)')
            program="$program$n@@include $f" ;;

    -[W-]file)
            program="$program$n@@include $@{2?'missing operand'@}"
            shift ;;

    -[W-]source=*)
            t=$(expr "$1" : '-.source=\(.*\)')
            program="$program$n$t" ;;

    -[W-]source)
            program="$program$n$@{2?'missing operand'@}"
            shift ;;

    -[W-]version)
            echo igawk: version 3.0 1>&2
            gawk --version
            exit 0 ;;

    -[W-]*) opts="$opts '$1'" ;;

    *)      break ;;
    esac
    shift
done

if [ -z "$program" ]
then
     program=$@{1?'missing program'@}
     shift
fi

# At this point, `program' has the program.
@c endfile
@end example

The @command{awk} program to process @code{@@include} directives
is stored in the shell variable @code{expand_prog}.  Doing this keeps
the shell script readable.  The @command{awk} program
reads through the user's program, one line at a time, using @code{getline}
(@pxref{Getline}).  The input
@value{FN}s and @code{@@include} statements are managed using a stack.
As each @code{@@include} is encountered, the current @value{FN} is
``pushed'' onto the stack and the file named in the @code{@@include}
directive becomes the current @value{FN}.  As each file is finished,
the stack is ``popped,'' and the previous input file becomes the current
input file again.  The process is started by making the original file
the first one on the stack.

The @code{pathto()} function does the work of finding the full path to
a file.  It simulates @command{gawk}'s behavior when searching the
@env{AWKPATH} environment variable
(@pxref{AWKPATH Variable}).
If a @value{FN} has a @samp{/} in it, no path search is done.
Similarly, if the @value{FN} is @code{"-"}, then that string is
used as-is.  Otherwise,
the @value{FN} is concatenated with the name of each directory in
the path, and an attempt is made to open the generated @value{FN}.
The only way to test if a file can be read in @command{awk} is to go
ahead and try to read it with @code{getline}; this is what @code{pathto()}
does.@footnote{On some very old versions of @command{awk}, the test
@samp{getline junk < t} can loop forever if the file exists but is empty.}
If the file can be read, it is closed and the @value{FN}
is returned:

@ignore
An alternative way to test for the file's existence would be to call
@samp{system("test -r " t)}, which uses the @command{test} utility to
see if the file exists and is readable.  The disadvantage to this method
is that it requires creating an extra process and can thus be slightly
slower.
@end ignore

@example
@c file eg/prog/igawk.sh
expand_prog='

function pathto(file,    i, t, junk)
@{
    if (index(file, "/") != 0)
        return file

    if (file == "-")
        return file

    for (i = 1; i <= ndirs; i++) @{
        t = (pathlist[i] "/" file)
@group
        if ((getline junk < t) > 0) @{
            # found it
            close(t)
            return t
        @}
@end group
    @}
    return ""
@}
@c endfile
@end example

The main program is contained inside one @code{BEGIN} rule.  The first thing it
does is set up the @code{pathlist} array that @code{pathto()} uses.  After
splitting the path on @samp{:}, null elements are replaced with @code{"."},
which represents the current directory:

@example
@c file eg/prog/igawk.sh
BEGIN @{
    path = ENVIRON["AWKPATH"]
    ndirs = split(path, pathlist, ":")
    for (i = 1; i <= ndirs; i++) @{
        if (pathlist[i] == "")
            pathlist[i] = "."
    @}
@c endfile
@end example

The stack is initialized with @code{ARGV[1]}, which will be @code{"/dev/stdin"}.
The main loop comes next.  Input lines are read in succession. Lines that
do not start with @code{@@include} are printed verbatim.
If the line does start with @code{@@include}, the @value{FN} is in @code{$2}.
@code{pathto()} is called to generate the full path.  If it cannot, then the program
prints an error message and continues.

The next thing to check is if the file is included already.  The
@code{processed} array is indexed by the full @value{FN} of each included
file and it tracks this information for us.  If the file is
seen again, a warning message is printed. Otherwise, the new @value{FN} is
pushed onto the stack and processing continues.

Finally, when @code{getline} encounters the end of the input file, the file
is closed and the stack is popped.  When @code{stackptr} is less than zero,
the program is done:

@example
@c file eg/prog/igawk.sh
    stackptr = 0
    input[stackptr] = ARGV[1] # ARGV[1] is first file

    for (; stackptr >= 0; stackptr--) @{
        while ((getline < input[stackptr]) > 0) @{
            if (tolower($1) != "@@include") @{
                print
                continue
            @}
            fpath = pathto($2)
            if (fpath == "") @{
                printf("igawk: %s:%d: cannot find %s\n",
                    input[stackptr], FNR, $2) > "/dev/stderr"
                continue
            @}
            if (! (fpath in processed)) @{
                processed[fpath] = input[stackptr]
                input[++stackptr] = fpath  # push onto stack
            @} else
                print $2, "included in", input[stackptr],
                    "already included in",
                    processed[fpath] > "/dev/stderr"
        @}
        close(input[stackptr])
    @}
@}'  # close quote ends `expand_prog' variable

processed_program=$(gawk -- "$expand_prog" /dev/stdin << EOF
$program
EOF
)
@c endfile
@end example

The shell construct @samp{@var{command} << @var{marker}} is called
a @dfn{here document}.  Everything in the shell script up to the
@var{marker} is fed to @var{command} as input.  The shell processes
the contents of the here document for variable and command substitution
(and possibly other things as well, depending upon the shell).

The shell construct @samp{$(@dots{})} is called @dfn{command substitution}.
The output of the command inside the parentheses is substituted
into the command line.
Because the result is used in a variable assignment,
it is saved as a single string, even if the results contain whitespace.

The expanded program is saved in the variable @code{processed_program}.
It's done in these steps:

@enumerate
@item
Run @command{gawk} with the @code{@@include}-processing program (the
value of the @code{expand_prog} shell variable) reading standard input.

@item
Standard input is the contents of the user's program,
from the shell variable @code{program}.
Feed its contents to @command{gawk} via a here document.

@item
Save the results of this processing in the shell variable
@code{processed_program} by using command substitution.
@end enumerate

The last step is to call @command{gawk} with the expanded program,
along with the original
options and command-line arguments that the user supplied:

@example
@c file eg/prog/igawk.sh
eval gawk $opts -- '"$processed_program"' '"$@@"'
@c endfile
@end example

The @command{eval} command is a shell construct that reruns the shell's parsing
process.  This keeps things properly quoted.

This version of @command{igawk} represents the fifth version of this program.
There are four key simplifications that make the program work better:

@itemize @value{BULLET}
@item
Using @code{@@include} even for the files named with @option{-f} makes building
the initial collected @command{awk} program much simpler; all the
@code{@@include} processing can be done once.

@item
Not trying to save the line read with @code{getline}
in the @code{pathto()} function when testing for the
file's accessibility for use with the main program simplifies things
considerably.

@item
Using a @code{getline} loop in the @code{BEGIN} rule does it all in one
place.  It is not necessary to call out to a separate loop for processing
nested @code{@@include} statements.

@item
Instead of saving the expanded program in a temporary file, putting it in a shell variable
avoids some potential security problems.
This has the disadvantage that the script relies upon more features
of the @command{sh} language, making it harder to follow for those who
aren't familiar with @command{sh}.
@end itemize

Also, this program illustrates that it is often worthwhile to combine
@command{sh} and @command{awk} programming together.  You can usually
accomplish quite a lot, without having to resort to low-level programming
in C or C++, and it is frequently easier to do certain kinds of string
and argument manipulation using the shell than it is in @command{awk}.

Finally, @command{igawk} shows that it is not always necessary to add new
features to a program; they can often be layered on top.@footnote{@command{gawk}
does @code{@@include} processing itself in order to support the use
of @command{awk} programs as Web CGI scripts.}


@node Anagram Program
@subsection Finding Anagrams from a Dictionary

@cindex anagrams, finding
An interesting programming challenge is to
search for @dfn{anagrams} in a
word list (such as
@file{/usr/share/dict/words} on many GNU/Linux systems).
One word is an anagram of another if both words contain
the same letters
(e.g., ``babbling'' and ``blabbing'').

Column 2, Problem C, of Jon Bentley's @cite{Programming Pearls}, Second
Edition, presents an elegant algorithm.  The idea is to give words that
are anagrams a common signature, sort all the words together by their
signatures, and then print them.  Dr.@: Bentley observes that taking the
letters in each word and sorting them produces those common signatures.

The following program uses arrays of arrays to bring together
words with the same signature and array sorting to print the words
in sorted order:

@cindex @code{anagram.awk} program
@example
@c file eg/prog/anagram.awk
# anagram.awk --- An implementation of the anagram-finding algorithm
#                 from Jon Bentley's "Programming Pearls," 2nd edition.
#                 Addison Wesley, 2000, ISBN 0-201-65788-0.
#                 Column 2, Problem C, section 2.8, pp 18-20.
@c endfile
@ignore
@c file eg/prog/anagram.awk
#
# This program requires gawk 4.0 or newer.
# Required gawk-specific features:
#   - True multidimensional arrays
#   - split() with "" as separator splits out individual characters
#   - asort() and asorti() functions
#
# See https://savannah.gnu.org/projects/gawk.
#
# Arnold Robbins
# arnold@@skeeve.com
# Public Domain
# January, 2011
@c endfile
@end ignore
@c file eg/prog/anagram.awk

/'s$/   @{ next @}        # Skip possessives
@c endfile
@end example

The program starts with a header, and then a rule to skip
possessives in the dictionary file. The next rule builds
up the data structure. The first dimension of the array
is indexed by the signature; the second dimension is the word
itself:

@example
@c file eg/prog/anagram.awk
@{
    key = word2key($1)  # Build signature
    data[key][$1] = $1  # Store word with signature
@}
@c endfile
@end example

The @code{word2key()} function creates the signature.
It splits the word apart into individual letters,
sorts the letters, and then joins them back together:

@example
@c file eg/prog/anagram.awk
# word2key --- split word apart into letters, sort, and join back together

function word2key(word,     a, i, n, result)
@{
    n = split(word, a, "")
    asort(a)

    for (i = 1; i <= n; i++)
        result = result a[i]

    return result
@}
@c endfile
@end example

Finally, the @code{END} rule traverses the array
and prints out the anagram lists.  It sends the output
to the system @command{sort} command because otherwise
the anagrams would appear in arbitrary order:

@example
@c file eg/prog/anagram.awk
END @{
    sort = "sort"
    for (key in data) @{
        # Sort words with same key
        nwords = asorti(data[key], words)
        if (nwords == 1)
            continue

        # And print. Minor glitch: trailing space at end of each line
        for (j = 1; j <= nwords; j++)
            printf("%s ", words[j]) | sort
        print "" | sort
    @}
    close(sort)
@}
@c endfile
@end example

Here is some partial output when the program is run:

@example
$ @kbd{gawk -f anagram.awk /usr/share/dict/words | grep '^b'}
@dots{}
babbled blabbed
babbler blabber brabble
babblers blabbers brabbles
babbling blabbing
babbly blabby
babel bable
babels beslab
babery yabber
@dots{}
@end example


@node Signature Program
@subsection And Now for Something Completely Different

@cindex signature program
@cindex Brini, Davide
The following program was written by Davide Brini
@c (@email{dave_br@@gmx.com})
and is published on @uref{http://backreference.org/2011/02/03/obfuscated-awk/,
his website}.
It serves as his signature in the Usenet group @code{comp.lang.awk}.
He supplies the following copyright terms:

@quotation
Copyright @copyright{} 2008 Davide Brini

Copying and distribution of the code published in this page, with or without
modification, are permitted in any medium without royalty provided the copyright
notice and this notice are preserved.
@end quotation

Here is the program:

@example
@group
awk 'BEGIN@{O="~"~"~";o="=="=="==";o+=+o;x=O""O;while(X++<=x+o+o)c=c"%c";
printf c,(x-O)*(x-O),x*(x-o)-o,x*(x-O)+x-O-o,+x*(x-O)-x+o,X*(o*o+O)+x-O,
X*(X-x)-o*o,(x+X)*o*o+o,x*(X-x)-O-O,x-O+(O+o+X+x)*(o+O),X*X-X*(x-O)-x+O,
O+X*(o*(o+O)+O),+x+O+X*o,x*(x-o),(o+X+x)*o*o-(x-O-O),O+(X-x)*(X+O),x-O@}'
@end group
@end example

@cindex Johansen, Chris
We leave it to you to determine what the program does.  (If you are
truly desperate to understand it, see Chris Johansen's explanation,
which is embedded in the Texinfo source file for this @value{DOCUMENT}.)

@ignore
To: "Arnold Robbins" <arnold@skeeve.com>
Date: Sat, 20 Aug 2011 13:50:46 -0400
Subject: The GNU Awk User's Guide, Section 13.3.11
From: "Chris Johansen" <johansen@main.nc.us>
Message-ID: <op.v0iw6wlv7finx3@asusodin.thrudvang.lan>

Arnold, you don't know me, but we have a tenuous connection.  My wife is
Barbara A. Field, FAIA, GIT '65 (B. Arch.).

I have had a couple of paper copies of "Effective Awk Programming" for
years, and now I'm going through a Kindle version of "The GNU Awk User's
Guide" again.  When I got to section 13.3.11, I reformatted and lightly
commented Davide Brin's signature script to understand its workings.

It occurs to me that this might have pedagogical value as an example
(although imperfect) of the value of whitespace and comments, and a
starting point for that discussion.  It certainly helped _me_ understand
what's going on.  You are welcome to it, as-is or modified (subject to
Davide's constraints, of course, which I think I have met).

If I were to include it in a future edition, I would put it at some
distance from section 13.3.11, say, as a note or an appendix, so as not to
be a "spoiler" to the puzzle.

Best regards,
--
Chris Johansen {johansen at main dot nc dot us}
  . . . collapsing the probability wave function, sending ripples of
certainty through the space-time continuum.


#! /usr/bin/gawk -f

# From "13.3.11 And Now For Something Completely Different"
#   https://www.gnu.org/software/gawk/manual/html_node/Signature-Program.html#Signature-Program

# Copyright @copyright{} 2008 Davide Brini

# Copying and distribution of the code published in this page, with
# or without modification, are permitted in any medium without
# royalty provided the copyright notice and this notice are preserved.

BEGIN {
  O = "~" ~ "~";    #  1
  o = "==" == "=="; #  1
  o += +o;          #  2
  x = O "" O;       # 11


  while ( X++ <= x + o + o ) c = c "%c";

  # O is  1
  # o is  2
  # x is 11
  # X is 17
  # c is "%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c"

  printf c,
    ( x - O )*( x - O),                  # 100 d
    x*( x - o ) - o,                     #  97 a
    x*( x - O ) + x - O - o,             # 118 v
    +x*( x - O ) - x + o,                # 101 e
    X*( o*o + O ) + x - O,               #  95 _
    X*( X - x ) - o*o,                   #  98 b
    ( x + X )*o*o + o,                   # 114 r
    x*( X - x ) - O - O,                 #  64 @
    x - O + ( O + o + X + x )*( o + O ), # 103 g
    X*X - X*( x - O ) - x + O,           # 109 m
    O + X*( o*( o + O ) + O ),           # 120 x
    +x + O + X*o,                        #  46 .
    x*( x - o),                          #  99 c
    ( o + X + x )*o*o - ( x - O - O ),   # 111 0
    O + ( X - x )*( X + O ),             # 109 m
    x - O                                #  10 \n
}
@end ignore

@node Programs Summary
@section Summary

@itemize @value{BULLET}
@item
The programs provided in this @value{CHAPTER}
continue on the theme that reading programs is an excellent way to learn
Good Programming.

@item
Using @samp{#!} to make @command{awk} programs directly runnable makes
them easier to use.  Otherwise, invoke the program using @samp{awk
-f @dots{}}.

@item
Reimplementing standard POSIX programs in @command{awk} is a pleasant
exercise; @command{awk}'s expressive power lets you write such programs
in relatively few lines of code, yet they are functionally complete
and usable.

@item
One of standard @command{awk}'s weaknesses is working with individual
characters.  The ability to use @code{split()} with the empty string as
the separator can considerably simplify such tasks.

@item
The examples here demonstrate the usefulness of the library
functions from @ref{Library Functions}
for a number of real (if small) programs.

@item
Besides reinventing POSIX wheels, other programs solved a selection of
interesting problems, such as finding duplicate words in text, printing
mailing labels, and finding anagrams.

@end itemize

@c EXCLUDE START
@node Programs Exercises
@section Exercises

@enumerate
@item
Rewrite @file{cut.awk} (@pxref{Cut Program})
using @code{split()} with @code{""} as the separator.

@item
In @ref{Egrep Program}, we mentioned that @samp{egrep -i} could be
simulated in versions of @command{awk} without @code{IGNORECASE} by
using @code{tolower()} on the line and the pattern. In a footnote there,
we also mentioned that this solution has a bug: the translated line is
output, and not the original one.  Fix this problem.
@c Exercise: Fix this, w/array and new line as key to original line

@item
The POSIX version of @command{id} takes options that control which
information is printed.  Modify the @command{awk} version
(@pxref{Id Program}) to accept the same arguments and perform in the
same way.

@item
The @code{split.awk} program (@pxref{Split Program}) assumes
that letters are contiguous in the character set,
which isn't true for EBCDIC systems.
Fix this problem.
(Hint: Consider a different way to work through the alphabet,
without relying on @code{ord()} and @code{chr()}.)

@item
In @file{uniq.awk} (@pxref{Uniq Program}, the
logic for choosing which lines to print represents a @dfn{state
machine}, which is ``a device that can be in one of a set number of stable
conditions depending on its previous condition and on the present values
of its inputs.''@footnote{This is the definition returned from entering
@code{define: state machine} into Google.}
Brian Kernighan suggests that
``an alternative approach to state machines is to just read
the input into an array, then use indexing.  It's almost always
easier code, and for most inputs where you would use this, just
as fast.''  Rewrite the logic to follow this
suggestion.


@item
Why can't the @file{wc.awk} program (@pxref{Wc Program}) just
use the value of @code{FNR} in @code{endfile()}?
Hint: Examine the code in @ref{Filetrans Function}.

@ignore
@command{wc} can't just use the value of @code{FNR} in
@code{endfile()}. If you examine the code in @ref{Filetrans Function},
you will see that @code{FNR} has already been reset by the time
@code{endfile()} is called.
@end ignore

@item
Manipulation of individual characters in the @command{translate} program
(@pxref{Translate Program}) is painful using standard @command{awk}
functions.  Given that @command{gawk} can split strings into individual
characters using @code{""} as the separator, how might you use this
feature to simplify the program?

@item
The @file{extract.awk} program (@pxref{Extract Program}) was written
before @command{gawk} had the @code{gensub()} function.  Use it
to simplify the code.

@item
Compare the performance of the @file{awksed.awk} program
(@pxref{Simple Sed}) with the more straightforward:

@example
BEGIN @{
    pat = ARGV[1]
    repl = ARGV[2]
    ARGV[1] = ARGV[2] = ""
@}

@{ gsub(pat, repl); print @}
@end example

@item
What are the advantages and disadvantages of @file{awksed.awk} versus
the real @command{sed} utility?

@ignore
  Advantage: egrep regexps
             speed (?)
  Disadvantage: no & in replacement text

Others?
@end ignore

@item
In @ref{Igawk Program}, we mentioned that not trying to save the line
read with @code{getline} in the @code{pathto()} function when testing
for the file's accessibility for use with the main program simplifies
things considerably.  What problem does this engender though?
@c answer, reading from "-" or /dev/stdin

@cindex search paths
@cindex search paths, for source files
@cindex source files@comma{} search path for
@cindex files, source@comma{} search path for
@cindex directories, searching
@item
As an additional example of the idea that it is not always necessary to
add new features to a program, consider the idea of having two files in
a directory in the search path:

@table @file
@item default.awk
This file contains a set of default library functions, such
as @code{getopt()} and @code{assert()}.

@item site.awk
This file contains library functions that are specific to a site or
installation; i.e., locally developed functions.
Having a separate file allows @file{default.awk} to change with
new @command{gawk} releases, without requiring the system administrator to
update it each time by adding the local functions.
@end table

One user
@c Karl Berry, karl@ileaf.com, 10/95
suggested that @command{gawk} be modified to automatically read these files
upon startup.  Instead, it would be very simple to modify @command{igawk}
to do this. Since @command{igawk} can process nested @code{@@include}
directives, @file{default.awk} could simply contain @code{@@include}
statements for the desired library functions.
Make this change.

@item
Modify @file{anagram.awk} (@pxref{Anagram Program}), to avoid
the use of the external @command{sort} utility.

@end enumerate
@c EXCLUDE END

@ifnotinfo
@part @value{PART3}Moving Beyond Standard @command{awk} with @command{gawk}
@end ifnotinfo

@ifdocbook
Part III focuses on features specific to @command{gawk}.
It contains the following chapters:

@itemize @value{BULLET}
@item
@ref{Advanced Features}

@item
@ref{Internationalization}

@item
@ref{Debugger}

@item
@ref{Arbitrary Precision Arithmetic}

@item
@ref{Dynamic Extensions}
@end itemize
@end ifdocbook

@node Advanced Features
@chapter Advanced Features of @command{gawk}
@cindex @command{gawk}, features, advanced
@cindex advanced features, @command{gawk}
@ignore
Contributed by: Peter Langston <pud!psl@bellcore.bellcore.com>

    Found in Steve English's "signature" line:

"Write documentation as if whoever reads it is a violent psychopath
who knows where you live."
@end ignore
@cindex Langston, Peter
@cindex English, Steve
@quotation
@i{Write documentation as if whoever reads it is
a violent psychopath who knows where you live.}
@author Steve English, as quoted by Peter Langston
@end quotation

This @value{CHAPTER} discusses advanced features in @command{gawk}.
It's a bit of a ``grab bag'' of items that are otherwise unrelated
to each other.
First, we look at a command-line option that allows @command{gawk} to recognize
nondecimal numbers in input data, not just in @command{awk}
programs.
Then, @command{gawk}'s special features for sorting arrays are presented.
Next, two-way I/O, discussed briefly in earlier parts of this
@value{DOCUMENT}, is described in full detail, along with the basics
of TCP/IP networking.  Finally, we see how @command{gawk}
can @dfn{profile} an @command{awk} program, making it possible to tune
it for performance.

@c FULLXREF ON
Additional advanced features are discussed in separate @value{CHAPTER}s of their
own:

@itemize @value{BULLET}
@item
@ref{Internationalization}, discusses how to internationalize
your @command{awk} programs, so that they can speak multiple
national languages.

@item
@ref{Debugger}, describes @command{gawk}'s built-in command-line
debugger for debugging @command{awk} programs.

@item
@ref{Arbitrary Precision Arithmetic}, describes how you can use
@command{gawk} to perform arbitrary-precision arithmetic.

@item
@ref{Dynamic Extensions},
discusses the ability to dynamically add new built-in functions to
@command{gawk}.
@end itemize
@c FULLXREF OFF

@menu
* Nondecimal Data::             Allowing nondecimal input data.
* Array Sorting::               Facilities for controlling array traversal and
                                sorting arrays.
* Two-way I/O::                 Two-way communications with another process.
* TCP/IP Networking::           Using @command{gawk} for network programming.
* Profiling::                   Profiling your @command{awk} programs.
* Advanced Features Summary::   Summary of advanced features.
@end menu

@node Nondecimal Data
@section Allowing Nondecimal Input Data
@cindex @option{--non-decimal-data} option
@cindex advanced features, nondecimal input data
@cindex input, data@comma{} nondecimal
@cindex constants, nondecimal

If you run @command{gawk} with the @option{--non-decimal-data} option,
you can have nondecimal values in your input data:

@example
$ @kbd{echo 0123 123 0x123 |}
> @kbd{gawk --non-decimal-data '@{ printf "%d, %d, %d\n", $1, $2, $3 @}'}
@print{} 83, 123, 291
@end example

For this feature to work, write your program so that
@command{gawk} treats your data as numeric:

@example
$ @kbd{echo 0123 123 0x123 | gawk '@{ print $1, $2, $3 @}'}
@print{} 0123 123 0x123
@end example

@noindent
The @code{print} statement treats its expressions as strings.
Although the fields can act as numbers when necessary,
they are still strings, so @code{print} does not try to treat them
numerically.  You need to add zero to a field to force it to
be treated as a number.  For example:

@example
$ @kbd{echo 0123 123 0x123 | gawk --non-decimal-data '}
> @kbd{@{ print $1, $2, $3}
>   @kbd{print $1 + 0, $2 + 0, $3 + 0 @}'}
@print{} 0123 123 0x123
@print{} 83 123 291
@end example

Because it is common to have decimal data with leading zeros, and because
using this facility could lead to surprising results, the default is to leave it
disabled.  If you want it, you must explicitly request it.

@cindex programming conventions, @code{--non-decimal-data} option
@cindex @option{--non-decimal-data} option, @code{strtonum()} function and
@cindex @code{strtonum()} function (@command{gawk}), @code{--non-decimal-data} option and
@quotation CAUTION
@emph{Use of this option is not recommended.}
It can break old programs very badly.
Instead, use the @code{strtonum()} function to convert your data
(@pxref{String Functions}).
This makes your programs easier to write and easier to read, and
leads to less surprising results.

This option may disappear in a future version of @command{gawk}.
@end quotation

@node Array Sorting
@section Controlling Array Traversal and Array Sorting

@command{gawk} lets you control the order in which a
@samp{for (@var{indx} in @var{array})}
loop traverses an array.

In addition, two built-in functions, @code{asort()} and @code{asorti()},
let you sort arrays based on the array values and indices, respectively.
These two functions also provide control over the sorting criteria used
to order the elements during sorting.

@menu
* Controlling Array Traversal:: How to use PROCINFO["sorted_in"].
* Array Sorting Functions::     How to use @code{asort()} and @code{asorti()}.
@end menu

@node Controlling Array Traversal
@subsection Controlling Array Traversal

By default, the order in which a @samp{for (@var{indx} in @var{array})} loop
scans an array is not defined; it is generally based upon
the internal implementation of arrays inside @command{awk}.

Often, though, it is desirable to be able to loop over the elements
in a particular order that you, the programmer, choose.  @command{gawk}
lets you do this.

@ref{Controlling Scanning} describes how you can assign special,
predefined values to @code{PROCINFO["sorted_in"]} in order to
control the order in which @command{gawk} traverses an array
during a @code{for} loop.

In addition, the value of @code{PROCINFO["sorted_in"]} can be a
function name.@footnote{This is why the predefined sorting orders
start with an @samp{@@} character, which cannot be part of an identifier.}
This lets you traverse an array based on any custom criterion.
The array elements are ordered according to the return value of this
function.  The comparison function should be defined with at least
four arguments:

@example
function comp_func(i1, v1, i2, v2)
@{
    @var{compare elements 1 and 2 in some fashion}
    @var{return < 0; 0; or > 0}
@}
@end example

Here, @code{i1} and @code{i2} are the indices, and @code{v1} and @code{v2}
are the corresponding values of the two elements being compared.
Either @code{v1} or @code{v2}, or both, can be arrays if the array being
traversed contains subarrays as values.
(@xref{Arrays of Arrays} for more information about subarrays.)
The three possible return values are interpreted as follows:

@table @code
@item comp_func(i1, v1, i2, v2) < 0
Index @code{i1} comes before index @code{i2} during loop traversal.

@item comp_func(i1, v1, i2, v2) == 0
Indices @code{i1} and @code{i2}
come together, but the relative order with respect to each other is undefined.

@item comp_func(i1, v1, i2, v2) > 0
Index @code{i1} comes after index @code{i2} during loop traversal.
@end table

Our first comparison function can be used to scan an array in
numerical order of the indices:

@example
@group
function cmp_num_idx(i1, v1, i2, v2)
@{
     # numerical index comparison, ascending order
     return (i1 - i2)
@}
@end group
@end example

Our second function traverses an array based on the string order of
the element values rather than by indices:

@example
function cmp_str_val(i1, v1, i2, v2)
@{
    # string value comparison, ascending order
    v1 = v1 ""
    v2 = v2 ""
    if (v1 < v2)
        return -1
    return (v1 != v2)
@}
@end example

The third
comparison function makes all numbers, and numeric strings without
any leading or trailing spaces, come out first during loop traversal:

@example
function cmp_num_str_val(i1, v1, i2, v2,   n1, n2)
@{
     # numbers before string value comparison, ascending order
     n1 = v1 + 0
     n2 = v2 + 0
     if (n1 == v1)
         return (n2 == v2) ? (n1 - n2) : -1
     else if (n2 == v2)
         return 1
     return (v1 < v2) ? -1 : (v1 != v2)
@}
@end example

Here is a main program to demonstrate how @command{gawk}
behaves using each of the previous functions:

@example
BEGIN @{
    data["one"] = 10
    data["two"] = 20
    data[10] = "one"
    data[100] = 100
    data[20] = "two"

    f[1] = "cmp_num_idx"
    f[2] = "cmp_str_val"
    f[3] = "cmp_num_str_val"
    for (i = 1; i <= 3; i++) @{
        printf("Sort function: %s\n", f[i])
        PROCINFO["sorted_in"] = f[i]
        for (j in data)
            printf("\tdata[%s] = %s\n", j, data[j])
        print ""
    @}
@}
@end example

Here are the results when the program is run:

@example
$ @kbd{gawk -f compdemo.awk}
@print{} Sort function: cmp_num_idx      @ii{Sort by numeric index}
@print{}     data[two] = 20
@print{}     data[one] = 10              @ii{Both strings are numerically zero}
@print{}     data[10] = one
@print{}     data[20] = two
@print{}     data[100] = 100
@print{}
@print{} Sort function: cmp_str_val      @ii{Sort by element values as strings}
@print{}     data[one] = 10
@print{}     data[100] = 100             @ii{String 100 is less than string 20}
@print{}     data[two] = 20
@print{}     data[10] = one
@print{}     data[20] = two
@print{}
@print{} Sort function: cmp_num_str_val  @ii{Sort all numeric values before all strings}
@print{}     data[one] = 10
@print{}     data[two] = 20
@print{}     data[100] = 100
@print{}     data[10] = one
@print{}     data[20] = two
@end example

Consider sorting the entries of a GNU/Linux system password file
according to login name.  The following program sorts records
by a specific field position and can be used for this purpose:

@example
# passwd-sort.awk --- simple program to sort by field position
# field position is specified by the global variable POS

function cmp_field(i1, v1, i2, v2)
@{
    # comparison by value, as string, and ascending order
    return v1[POS] < v2[POS] ? -1 : (v1[POS] != v2[POS])
@}

@{
    for (i = 1; i <= NF; i++)
        a[NR][i] = $i
@}

@group
END @{
    PROCINFO["sorted_in"] = "cmp_field"
@end group
    if (POS < 1 || POS > NF)
        POS = 1

    for (i in a) @{
        for (j = 1; j <= NF; j++)
            printf("%s%c", a[i][j], j < NF ? ":" : "")
        print ""
    @}
@}
@end example

The first field in each entry of the password file is the user's login name,
and the fields are separated by colons.
Each record defines a subarray,
with each field as an element in the subarray.
Running the program produces the
following output:

@example
$ @kbd{gawk -v POS=1 -F: -f sort.awk /etc/passwd}
@print{} adm:x:3:4:adm:/var/adm:/sbin/nologin
@print{} apache:x:48:48:Apache:/var/www:/sbin/nologin
@print{} avahi:x:70:70:Avahi daemon:/:/sbin/nologin
@dots{}
@end example

The comparison should normally always return the same value when given a
specific pair of array elements as its arguments.  If inconsistent
results are returned, then the order is undefined.  This behavior can be
exploited to introduce random order into otherwise seemingly
ordered data:

@example
function cmp_randomize(i1, v1, i2, v2)
@{
    # random order (caution: this may never terminate!)
    return (2 - 4 * rand())
@}
@end example

As already mentioned, the order of the indices is arbitrary if two
elements compare equal.  This is usually not a problem, but letting
the tied elements come out in arbitrary order can be an issue, especially
when comparing item values.  The partial ordering of the equal elements
may change the next time the array is traversed, if other elements are added to or
removed from the array.  One way to resolve ties when comparing elements
with otherwise equal values is to include the indices in the comparison
rules.  Note that doing this may make the loop traversal less efficient,
so consider it only if necessary.  The following comparison functions
force a deterministic order, and are based on the fact that the
(string) indices of two elements are never equal:

@example
function cmp_numeric(i1, v1, i2, v2)
@{
    # numerical value (and index) comparison, descending order
    return (v1 != v2) ? (v2 - v1) : (i2 - i1)
@}

@group
function cmp_string(i1, v1, i2, v2)
@{
    # string value (and index) comparison, descending order
    v1 = v1 i1
    v2 = v2 i2
    return (v1 > v2) ? -1 : (v1 != v2)
@}
@end group
@end example

@c Avoid using the term ``stable'' when describing the unpredictable behavior
@c if two items compare equal.  Usually, the goal of a "stable algorithm"
@c is to maintain the original order of the items, which is a meaningless
@c concept for a list constructed from a hash.

A custom comparison function can often simplify ordered loop
traversal, and the sky is really the limit when it comes to
designing such a function.

When string comparisons are made during a sort, either for element
values where one or both aren't numbers, or for element indices
handled as strings, the value of @code{IGNORECASE}
(@pxref{Built-in Variables}) controls whether
the comparisons treat corresponding upper- and lowercase letters as
equivalent or distinct.

Another point to keep in mind is that in the case of subarrays,
the element values can themselves be arrays; a production comparison
function should use the @code{isarray()} function
(@pxref{Type Functions})
to check for this, and choose a defined sorting order for subarrays.

All sorting based on @code{PROCINFO["sorted_in"]}
is disabled in POSIX mode,
because the @code{PROCINFO} array is not special in that case.

As a side note, sorting the array indices before traversing
the array has been reported to add a 15% to 20% overhead to the
execution time of @command{awk} programs. For this reason,
sorted array traversal is not the default.

@c The @command{gawk}
@c maintainers believe that only the people who wish to use a
@c feature should have to pay for it.

@node Array Sorting Functions
@subsection Sorting Array Values and Indices with @command{gawk}

@cindex arrays, sorting
@cindexgawkfunc{asort}
@cindex @code{asort()} function (@command{gawk}), arrays@comma{} sorting
@cindexgawkfunc{asorti}
@cindex @code{asorti()} function (@command{gawk}), arrays@comma{} sorting
@cindex sort function, arrays, sorting
In most @command{awk} implementations, sorting an array requires writing
a @code{sort()} function.  This can be educational for exploring
different sorting algorithms, but usually that's not the point of the program.
@command{gawk} provides the built-in @code{asort()} and @code{asorti()}
functions (@pxref{String Functions}) for sorting arrays.  For example:

@example
@var{populate the array} data
n = asort(data)
for (i = 1; i <= n; i++)
    @var{do something with} data[i]
@end example

After the call to @code{asort()}, the array @code{data} is indexed from 1
to some number @var{n}, the total number of elements in @code{data}.
(This count is @code{asort()}'s return value.)
@code{data[1]} @value{LEQ} @code{data[2]} @value{LEQ} @code{data[3]}, and so on.
The default comparison is based on the type of the elements
(@pxref{Typing and Comparison}).
All numeric values come before all string values,
which in turn come before all subarrays.

@cindex side effects, @code{asort()} function
An important side effect of calling @code{asort()} is that
@emph{the array's original indices are irrevocably lost}.
As this isn't always desirable, @code{asort()} accepts a
second argument:

@example
@var{populate the array} source
n = asort(source, dest)
for (i = 1; i <= n; i++)
    @var{do something with} dest[i]
@end example

In this case, @command{gawk} copies the @code{source} array into the
@code{dest} array and then sorts @code{dest}, destroying its indices.
However, the @code{source} array is not affected.

Often, what's needed is to sort on the values of the @emph{indices}
instead of the values of the elements.  To do that, use the
@code{asorti()} function.  The interface and behavior are identical to
that of @code{asort()}, except that the index values are used for sorting
and become the values of the result array:

@example
@{ source[$0] = some_func($0) @}

END @{
    n = asorti(source, dest)
    for (i = 1; i <= n; i++) @{
        @ii{Work with sorted indices directly:}
        @var{do something with} dest[i]
        @dots{}
        @ii{Access original array via sorted indices:}
        @var{do something with} source[dest[i]]
    @}
@}
@end example

So far, so good. Now it starts to get interesting.  Both @code{asort()}
and @code{asorti()} accept a third string argument to control comparison
of array elements.  When we introduced @code{asort()} and @code{asorti()}
in @ref{String Functions}, we ignored this third argument; however,
now is the time to describe how this argument affects these two functions.

Basically, the third argument specifies how the array is to be sorted.
There are two possibilities.  As with @code{PROCINFO["sorted_in"]},
this argument may be one of the predefined names that @command{gawk}
provides (@pxref{Controlling Scanning}), or it may be the name of a
user-defined function (@pxref{Controlling Array Traversal}).

In the latter case, @emph{the function can compare elements in any way
it chooses}, taking into account just the indices, just the values,
or both.  This is extremely powerful.

Once the array is sorted, @code{asort()} takes the @emph{values} in
their final order and uses them to fill in the result array, whereas
@code{asorti()} takes the @emph{indices} in their final order and uses
them to fill in the result array.

@cindex reference counting, sorting arrays
@quotation NOTE
Copying array indices and elements isn't expensive in terms of memory.
Internally, @command{gawk} maintains @dfn{reference counts} to data.
For example, when @code{asort()} copies the first array to the second one,
there is only one copy of the original array elements' data, even though
both arrays use the values.
@end quotation

@c Document It And Call It A Feature. Sigh.
@cindex @command{gawk}, @code{IGNORECASE} variable in
@cindex arrays, sorting, and @code{IGNORECASE} variable
@cindex @code{IGNORECASE} variable, and array sorting functions
Because @code{IGNORECASE} affects string comparisons, the value
of @code{IGNORECASE} also affects sorting for both @code{asort()} and @code{asorti()}.
Note also that the locale's sorting order does @emph{not}
come into play; comparisons are based on character values only.@footnote{This
is true because locale-based comparison occurs only when in
POSIX-compatibility mode, and because @code{asort()} and @code{asorti()} are
@command{gawk} extensions, they are not available in that case.}

The following example demonstrates the use of a comparison function with
@code{asort()}.  The comparison function, @code{case_fold_compare()}, maps
both values to lowercase in order to compare them ignoring case.

@example
@group
# case_fold_compare --- compare as strings, ignoring case

function case_fold_compare(i1, v1, i2, v2,    l, r)
@{
    l = tolower(v1)
@end group
    r = tolower(v2)

    if (l < r)
        return -1
    else if (l == r)
        return 0
    else
        return 1
@}
@end example

And here is the test program for it:

@example
# Test program

BEGIN @{
    Letters = "abcdefghijklmnopqrstuvwxyz" \
              "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
    split(Letters, data, "")

    asort(data, result, "case_fold_compare")

    j = length(result)
    for (i = 1; i <= j; i++) @{
        printf("%s", result[i])
        if (i % (j/2) == 0)
            printf("\n")
        else
            printf(" ")
    @}
@}
@end example

When run, we get the following:

@example
$ @kbd{gawk -f case_fold_compare.awk}
@print{} A a B b c C D d e E F f g G H h i I J j k K l L M m
@print{} n N O o p P Q q r R S s t T u U V v w W X x y Y z Z
@end example

@node Two-way I/O
@section Two-Way Communications with Another Process

@c 8/2014. Neither Mike nor BWK saw this as relevant. Commenting it out.
@ignore
@cindex Brennan, Michael
@cindex programmers, attractiveness of
@smallexample
@c Path: cssun.mathcs.emory.edu!gatech!newsxfer3.itd.umich.edu!news-peer.sprintlink.net!news-sea-19.sprintlink.net!news-in-west.sprintlink.net!news.sprintlink.net!Sprint!204.94.52.5!news.whidbey.com!brennan
From: brennan@@whidbey.com (Mike Brennan)
Newsgroups: comp.lang.awk
Subject: Re: Learn the SECRET to Attract Women Easily
Date: 4 Aug 1997 17:34:46 GMT
@c Organization: WhidbeyNet
@c Lines: 12
Message-ID: <5s53rm$eca@@news.whidbey.com>
@c References: <5s20dn$2e1@chronicle.concentric.net>
@c Reply-To: brennan@whidbey.com
@c NNTP-Posting-Host: asn202.whidbey.com
@c X-Newsreader: slrn (0.9.4.1 UNIX)
@c Xref: cssun.mathcs.emory.edu comp.lang.awk:5403

On 3 Aug 1997 13:17:43 GMT, Want More Dates???
<tracy78@@kilgrona.com> wrote:
>Learn the SECRET to Attract Women Easily
>
>The SCENT(tm)  Pheromone Sex Attractant For Men to Attract Women

The scent of awk programmers is a lot more attractive to women than
the scent of perl programmers.
--
Mike Brennan
@c brennan@@whidbey.com
@end smallexample
@end ignore

@cindex advanced features, processes@comma{} communicating with
@cindex processes, two-way communications with
It is often useful to be able to
send data to a separate program for
processing and then read the result.  This can always be
done with temporary files:

@example
# Write the data for processing
tempfile = ("mydata." PROCINFO["pid"])
while (@var{not done with data})
    print @var{data} | ("subprogram > " tempfile)
close("subprogram > " tempfile)

# Read the results, remove tempfile when done
while ((getline newdata < tempfile) > 0)
    @var{process} newdata @var{appropriately}
close(tempfile)
system("rm " tempfile)
@end example

@noindent
This works, but not elegantly.  Among other things, it requires that
the program be run in a directory that cannot be shared among users;
for example, @file{/tmp} will not do, as another user might happen
to be using a temporary file with the same name.@footnote{Michael
Brennan suggests the use of @command{rand()} to generate unique
@value{FN}s. This is a valid point; nevertheless, temporary files
remain more difficult to use than two-way pipes.} @c 8/2014

@cindex coprocesses
@cindex input/output, two-way
@cindex @code{|} (vertical bar), @code{|&} operator (I/O)
@cindex vertical bar (@code{|}), @code{|&} operator (I/O)
@cindex @command{csh} utility, @code{|&} operator, comparison with
However, with @command{gawk}, it is possible to
open a @emph{two-way} pipe to another process.  The second process is
termed a @dfn{coprocess}, as it runs in parallel with @command{gawk}.
The two-way connection is created using the @samp{|&} operator
(borrowed from the Korn shell, @command{ksh}):@footnote{This is very
different from the same operator in the C shell and in Bash.}

@example
do @{
    print @var{data} |& "subprogram"
    "subprogram" |& getline results
@} while (@var{data left to process})
close("subprogram")
@end example

The first time an I/O operation is executed using the @samp{|&}
operator, @command{gawk} creates a two-way pipeline to a child process
that runs the other program.  Output created with @code{print}
or @code{printf} is written to the program's standard input, and
output from the program's standard output can be read by the @command{gawk}
program using @code{getline}.
As is the case with processes started by @samp{|}, the subprogram
can be any program, or pipeline of programs, that can be started by
the shell.

There are some cautionary items to be aware of:

@itemize @value{BULLET}
@item
As the code inside @command{gawk} currently stands, the coprocess's
standard error goes to the same place that the parent @command{gawk}'s
standard error goes. It is not possible to read the child's
standard error separately.

@cindex deadlocks
@cindex buffering, input/output
@cindex @code{getline} command, deadlock and
@item
I/O buffering may be a problem.  @command{gawk} automatically
flushes all output down the pipe to the coprocess.
However, if the coprocess does not flush its output,
@command{gawk} may hang when doing a @code{getline} in order to read
the coprocess's results.  This could lead to a situation
known as @dfn{deadlock}, where each process is waiting for the
other one to do something.
@end itemize

@cindex @code{close()} function, two-way pipes and
It is possible to close just one end of the two-way pipe to
a coprocess, by supplying a second argument to the @code{close()}
function of either @code{"to"} or @code{"from"}
(@pxref{Close Files And Pipes}).
These strings tell @command{gawk} to close the end of the pipe
that sends data to the coprocess or the end that reads from it,
respectively.

@cindex @command{sort} utility, coprocesses and
This is particularly necessary in order to use
the system @command{sort} utility as part of a coprocess;
@command{sort} must read @emph{all} of its input
data before it can produce any output.
The @command{sort} program does not receive an end-of-file indication
until @command{gawk} closes the write end of the pipe.

When you have finished writing data to the @command{sort}
utility, you can close the @code{"to"} end of the pipe, and
then start reading sorted data via @code{getline}.
For example:

@example
BEGIN @{
    command = "LC_ALL=C sort"
    n = split("abcdefghijklmnopqrstuvwxyz", a, "")

    for (i = n; i > 0; i--)
        print a[i] |& command
    close(command, "to")

    while ((command |& getline line) > 0)
        print "got", line
    close(command)
@}
@end example

This program writes the letters of the alphabet in reverse order, one
per line, down the two-way pipe to @command{sort}.  It then closes the
write end of the pipe, so that @command{sort} receives an end-of-file
indication.  This causes @command{sort} to sort the data and write the
sorted data back to the @command{gawk} program.  Once all of the data
has been read, @command{gawk} terminates the coprocess and exits.

As a side note, the assignment @samp{LC_ALL=C} in the @command{sort}
command ensures traditional Unix (ASCII) sorting from @command{sort}.
This is not strictly necessary here, but it's good to know how to do this.

Be careful when closing the @code{"from"} end of a two-way pipe; in this
case @command{gawk} waits for the child process to exit, which may cause
your program to hang.  (Thus, this particular feature is of much less
use in practice than being able to close the @code{"to"} end.)

@quotation CAUTION
Normally,
it is a fatal error to write to the @code{"to"} end of a two-way
pipe which has been closed, and it is also a fatal error to read
from the @code{"from"} end of a two-way pipe that has been closed.

You may set @code{PROCINFO["@var{command}", "NONFATAL"]} to
make such operations become nonfatal. If you do so, you then need
to check @code{ERRNO} after each @code{print}, @code{printf},
or @code{getline}.
@xref{Nonfatal}, for more information.
@end quotation

@cindex @command{gawk}, @code{PROCINFO} array in
@cindex @code{PROCINFO} array, and communications via ptys
You may also use pseudo-ttys (ptys) for
two-way communication instead of pipes, if your system supports them.
This is done on a per-command basis, by setting a special element
in the @code{PROCINFO} array
(@pxref{Auto-set}),
like so:

@example
command = "sort -nr"           # command, save in convenience variable
PROCINFO[command, "pty"] = 1   # update PROCINFO
print @dots{} |& command           # start two-way pipe
@dots{}
@end example

@noindent
If your system does not have ptys, or if all the system's ptys are in use,
@command{gawk} automatically falls back to using regular pipes.

Using ptys usually avoids the buffer deadlock issues described earlier,
at some loss in performance. This is because the tty driver buffers
and sends data line-by-line.  On systems with the @command{stdbuf}
(part of the @uref{https://www.gnu.org/software/coreutils/coreutils.html,
GNU Coreutils package}), you can use that program instead of ptys.

Note also that ptys are not fully transparent. Certain binary control
codes, such @kbd{Ctrl-d} for end-of-file, are interpreted by the tty
driver and not passed through.

@quotation CAUTION
Finally, coprocesses open up the possibility of @dfn{deadlock} between
@command{gawk} and the program running in the coprocess. This can occur
if you send ``too much'' data to the coprocess before reading any back;
each process is blocked writing data with noone available to read what
they've already written.  There is no workaround for deadlock; careful
programming and knowledge of the behavior of the coprocess are required.
@end quotation

@node TCP/IP Networking
@section Using @command{gawk} for Network Programming
@cindex advanced features, network programming
@cindex networks, programming
@cindex TCP/IP
@cindex @code{/inet/@dots{}} special files (@command{gawk})
@cindex files, @code{/inet/@dots{}} (@command{gawk})
@cindex @code{/inet4/@dots{}} special files (@command{gawk})
@cindex files, @code{/inet4/@dots{}} (@command{gawk})
@cindex @code{/inet6/@dots{}} special files (@command{gawk})
@cindex files, @code{/inet6/@dots{}} (@command{gawk})
@cindex @code{EMRED}
@ifnotdocbook
@quotation
@code{EMRED}:@*
@ @ @ @ @i{A host is a host from coast to coast,@*
@ @ @ @ and nobody talks to a host that's close,@*
@ @ @ @ unless the host that isn't close@*
@ @ @ @ is busy, hung, or dead.}
@author Mike O'Brien (aka Mr.@: Protocol)
@end quotation
@end ifnotdocbook

@docbook
<blockquote>
<attribution>Mike O'Brien (aka Mr.&nbsp;Protocol)</attribution>
<literallayout class="normal"><literal>EMRED</literal>:
&nbsp;&nbsp;&nbsp;&nbsp;<emphasis>A host is a host from coast to coast,</emphasis>
&nbsp;&nbsp;&nbsp;&nbsp;<emphasis>and no-one can talk to host that's close,</emphasis>
&nbsp;&nbsp;&nbsp;&nbsp;<emphasis>unless the host that isn't close</emphasis>
&nbsp;&nbsp;&nbsp;&nbsp;<emphasis>is busy, hung, or dead.</emphasis></literallayout>
</blockquote>
@end docbook

In addition to being able to open a two-way pipeline to a coprocess
on the same system
(@pxref{Two-way I/O}),
it is possible to make a two-way connection to
another process on another system across an IP network connection.

You can think of this as just a @emph{very long} two-way pipeline to
a coprocess.
The way @command{gawk} decides that you want to use TCP/IP networking is
by recognizing special @value{FN}s that begin with one of @samp{/inet/},
@samp{/inet4/}, or @samp{/inet6/}.

The full syntax of the special @value{FN} is
@file{/@var{net-type}/@var{protocol}/@var{local-port}/@var{remote-host}/@var{remote-port}}.
The components are:

@table @var
@item net-type
Specifies the kind of Internet connection to make.
Use @samp{/inet4/} to force IPv4, and
@samp{/inet6/} to force IPv6.
Plain @samp{/inet/} (which used to be the only option) uses
the system default, most likely IPv4.

@item protocol
The protocol to use over IP.  This must be either @samp{tcp}, or
@samp{udp}, for a TCP or UDP IP connection,
respectively.  TCP should be used for most applications.

@item local-port
@cindex @code{getaddrinfo()} function (C library)
The local TCP or UDP port number to use.  Use a port number of @samp{0}
when you want the system to pick a port. This is what you should do
when writing a TCP or UDP client.
You may also use a well-known service name, such as @samp{smtp}
or @samp{http}, in which case @command{gawk} attempts to determine
the predefined port number using the C @code{getaddrinfo()} function.

@item remote-host
The IP address or fully qualified domain name of the Internet
host to which you want to connect.

@item remote-port
The TCP or UDP port number to use on the given @var{remote-host}.
Again, use @samp{0} if you don't care, or else a well-known
service name.
@end table

@cindex @command{gawk}, @code{ERRNO} variable in
@cindex @code{ERRNO} variable
@quotation NOTE
Failure in opening a two-way socket will result in a nonfatal error
being returned to the calling code. The value of @code{ERRNO} indicates
the error (@pxref{Auto-set}).
@end quotation

Consider the following very simple example:

@example
BEGIN @{
    Service = "/inet/tcp/0/localhost/daytime"
    Service |& getline
    print $0
    close(Service)
@}
@end example

This program reads the current date and time from the local system's
TCP @code{daytime} server.
It then prints the results and closes the connection.

Because this topic is extensive, the use of @command{gawk} for
TCP/IP programming is documented separately.
@ifinfo
See
@inforef{Top, , General Introduction, gawkinet, @value{GAWKINETTITLE}},
@end ifinfo
@ifnotinfo
See
@uref{https://www.gnu.org/software/gawk/manual/gawkinet/,
@cite{@value{GAWKINETTITLE}}},
which comes as part of the @command{gawk} distribution,
@end ifnotinfo
for a much more complete introduction and discussion, as well as
extensive examples.

@quotation NOTE
@command{gawk} can only open direct sockets. There is currently
no way to access services available over Secure Socket Layer
(SSL); this includes any web service whose URL starts with @samp{https://}.
@end quotation


@node Profiling
@section Profiling Your @command{awk} Programs
@cindex @command{awk} programs, profiling
@cindex profiling @command{awk} programs
@cindex @code{awkprof.out} file
@cindex files, @code{awkprof.out}

You may produce execution traces of your @command{awk} programs.
This is done by passing the option @option{--profile} to @command{gawk}.
When @command{gawk} has finished running, it creates a profile of your program in a file
named @file{awkprof.out}. Because it is profiling, it also executes up to 45% slower than
@command{gawk} normally does.

@cindex @option{--profile} option
As shown in the following example,
the @option{--profile} option can be used to change the name of the file
where @command{gawk} will write the profile:

@example
gawk --profile=myprog.prof -f myprog.awk data1 data2
@end example

@noindent
In the preceding example, @command{gawk} places the profile in
@file{myprog.prof} instead of in @file{awkprof.out}.

Here is a sample session showing a simple @command{awk} program,
its input data, and the results from running @command{gawk} with the
@option{--profile} option.  First, the @command{awk} program:

@example
BEGIN @{ print "First BEGIN rule" @}

END @{ print "First END rule" @}

/foo/ @{
    print "matched /foo/, gosh"
    for (i = 1; i <= 3; i++)
        sing()
@}

@{
    if (/foo/)
        print "if is true"
    else
        print "else is true"
@}

BEGIN @{ print "Second BEGIN rule" @}

END @{ print "Second END rule" @}

function sing(    dummy)
@{
    print "I gotta be me!"
@}
@end example

Following is the input data:

@example
foo
bar
baz
foo
junk
@end example

Here is the @file{awkprof.out} that results from running the
@command{gawk} profiler on this program and data (this example also
illustrates that @command{awk} programmers sometimes get up very early
in the morning to work):

@cindex @code{BEGIN} pattern, and profiling
@cindex @code{END} pattern, and profiling
@example
    # gawk profile, created Mon Sep 29 05:16:21 2014

    # BEGIN rule(s)

    BEGIN @{
 1          print "First BEGIN rule"
    @}

    BEGIN @{
 1          print "Second BEGIN rule"
    @}

    # Rule(s)

 5  /foo/ @{ # 2
 2          print "matched /foo/, gosh"
 6          for (i = 1; i <= 3; i++) @{
 6                  sing()
            @}
    @}

 5  @{
 5          if (/foo/) @{ # 2
 2                  print "if is true"
 3          @} else @{
 3                  print "else is true"
            @}
    @}

    # END rule(s)

    END @{
 1          print "First END rule"
    @}

    END @{
 1          print "Second END rule"
    @}


    # Functions, listed alphabetically

 6  function sing(dummy)
    @{
 6          print "I gotta be me!"
    @}
@end example

This example illustrates many of the basic features of profiling output.
They are as follows:

@itemize @value{BULLET}
@item
The program is printed in the order @code{BEGIN} rules,
@code{BEGINFILE} rules,
pattern--action rules,
@code{ENDFILE} rules, @code{END} rules, and functions, listed
alphabetically.
Multiple @code{BEGIN} and @code{END} rules retain their
separate identities, as do
multiple @code{BEGINFILE} and @code{ENDFILE} rules.

@cindex patterns, counts, in a profile
@item
Pattern--action rules have two counts.
The first count, to the left of the rule, shows how many times
the rule's pattern was @emph{tested}.
The second count, to the right of the rule's opening left brace
in a comment,
shows how many times the rule's action was @emph{executed}.
The difference between the two indicates how many times the rule's
pattern evaluated to false.

@item
Similarly,
the count for an @code{if}-@code{else} statement shows how many times
the condition was tested.
To the right of the opening left brace for the @code{if}'s body
is a count showing how many times the condition was true.
The count for the @code{else}
indicates how many times the test failed.

@cindex loops, count for header, in a profile
@item
The count for a loop header (such as @code{for}
or @code{while}) shows how many times the loop test was executed.
(Because of this, you can't just look at the count on the first
statement in a rule to determine how many times the rule was executed.
If the first statement is a loop, the count is misleading.)

@cindex functions, user-defined, counts, in a profile
@cindex user-defined, functions, counts, in a profile
@item
For user-defined functions, the count next to the @code{function}
keyword indicates how many times the function was called.
The counts next to the statements in the body show how many times
those statements were executed.

@cindex @code{@{@}} (braces)
@cindex braces (@code{@{@}})
@item
The layout uses ``K&R'' style with TABs.
Braces are used everywhere, even when
the body of an @code{if}, @code{else}, or loop is only a single statement.

@cindex @code{()} (parentheses), in a profile
@cindex parentheses @code{()}, in a profile
@item
Parentheses are used only where needed, as indicated by the structure
of the program and the precedence rules.
For example, @samp{(3 + 5) * 4} means add three and five, then multiply
the total by four.  However, @samp{3 + 5 * 4} has no parentheses, and
means @samp{3 + (5 * 4)}.

@ignore
@item
All string concatenations are parenthesized too.
(This could be made a bit smarter.)
@end ignore

@item
Parentheses are used around the arguments to @code{print}
and @code{printf} only when
the @code{print} or @code{printf} statement is followed by a redirection.
Similarly, if
the target of a redirection isn't a scalar, it gets parenthesized.

@item
@command{gawk} supplies leading comments in
front of the @code{BEGIN} and @code{END} rules,
the @code{BEGINFILE} and @code{ENDFILE} rules,
the pattern--action rules, and the functions.

@end itemize

The profiled version of your program may not look exactly like what you
typed when you wrote it.  This is because @command{gawk} creates the
profiled version by ``pretty-printing'' its internal representation of
the program.  The advantage to this is that @command{gawk} can produce
a standard representation.
Also, things such as:

@example
/foo/
@end example

@noindent
come out as:

@example
/foo/   @{
    print $0
@}
@end example

@noindent
which is correct, but possibly unexpected.

@cindex profiling @command{awk} programs, dynamically
@cindex @command{gawk} program, dynamic profiling
@cindex dynamic profiling
Besides creating profiles when a program has completed,
@command{gawk} can produce a profile while it is running.
This is useful if your @command{awk} program goes into an
infinite loop and you want to see what has been executed.
To use this feature, run @command{gawk} with the @option{--profile}
option in the background:

@example
$ @kbd{gawk --profile -f myprog &}
[1] 13992
@end example

@cindex @command{kill} command@comma{} dynamic profiling
@cindex @code{USR1} signal, for dynamic profiling
@cindex @code{SIGUSR1} signal, for dynamic profiling
@cindex signals, @code{USR1}/@code{SIGUSR1}, for profiling
@noindent
The shell prints a job number and process ID number; in this case, 13992.
Use the @command{kill} command to send the @code{USR1} signal
to @command{gawk}:

@example
$ @kbd{kill -USR1 13992}
@end example

@noindent
As usual, the profiled version of the program is written to
@file{awkprof.out}, or to a different file if one was specified with
the @option{--profile} option.

Along with the regular profile, as shown earlier, the profile file
includes a trace of any active functions:

@example
# Function Call Stack:

#   3. baz
#   2. bar
#   1. foo
# -- main --
@end example

You may send @command{gawk} the @code{USR1} signal as many times as you like.
Each time, the profile and function call trace are appended to the output
profile file.

@cindex @code{HUP} signal, for dynamic profiling
@cindex @code{SIGHUP} signal, for dynamic profiling
@cindex signals, @code{HUP}/@code{SIGHUP}, for profiling
If you use the @code{HUP} signal instead of the @code{USR1} signal,
@command{gawk} produces the profile and the function call trace and then exits.

@cindex @code{INT} signal (MS-Windows)
@cindex @code{SIGINT} signal (MS-Windows)
@cindex signals, @code{INT}/@code{SIGINT} (MS-Windows)
@cindex @code{QUIT} signal (MS-Windows)
@cindex @code{SIGQUIT} signal (MS-Windows)
@cindex signals, @code{QUIT}/@code{SIGQUIT} (MS-Windows)
When @command{gawk} runs on MS-Windows systems, it uses the
@code{INT} and @code{QUIT} signals for producing the profile, and in
the case of the @code{INT} signal, @command{gawk} exits.  This is
because these systems don't support the @command{kill} command, so the
only signals you can deliver to a program are those generated by the
keyboard.  The @code{INT} signal is generated by the
@kbd{Ctrl-c} or @kbd{Ctrl-BREAK} key, while the
@code{QUIT} signal is generated by the @kbd{Ctrl-\} key.

Finally, @command{gawk} also accepts another option, @option{--pretty-print}.
When called this way, @command{gawk} ``pretty-prints'' the program into
@file{awkprof.out}, without any execution counts.

@quotation NOTE
Once upon a time, the @option{--pretty-print} option would also run
your program.  This is is no longer the case.
@end quotation

There is a significant difference between the output created when
profiling, and that created when pretty-printing.  Pretty-printed output
preserves the original comments that were in the program, although their
placement may not correspond exactly to their original locations in the
source code.@footnote{@command{gawk} does the best it can to preserve
the distinction between comments at the end of a statement and comments
on lines by themselves. Due to implementation constraints, it does not
always do so correctly, particularly for @code{switch} statements. The
@command{gawk} maintainers hope to improve this in a subsequent
release.}

However, as a deliberate design decision, profiling output @emph{omits}
the original program's comments. This allows you to focus on the
execution count data and helps you avoid the temptation to use the
profiler for pretty-printing.

Additionally, pretty-printed output does not have the leading indentation
that the profiling output does. This makes it easy to pretty-print your
code once development is completed, and then use the result as the final
version of your program.

Because the internal representation of your program is formatted to
recreate an @command{awk} program, profiling and pretty-printing
automatically disable @command{gawk}'s default optimizations.

Pretty printing also preserves the original format of numeric
constants; if you used an octal or hexadecimal value in your source
code, it will appear that way in the output.

@node Advanced Features Summary
@section Summary

@itemize @value{BULLET}
@item
The @option{--non-decimal-data} option causes @command{gawk} to treat
octal- and hexadecimal-looking input data as octal and hexadecimal.
This option should be used with caution or not at all; use of @code{strtonum()}
is preferable.
Note that this option may disappear in a future version of @command{gawk}.

@item
You can take over complete control of sorting in @samp{for (@var{indx} in @var{array})}
array traversal by setting @code{PROCINFO["sorted_in"]} to the name of a user-defined
function that does the comparison of array elements based on index and value.

@item
Similarly, you can supply the name of a user-defined comparison function as the
third argument to either @code{asort()} or @command{asorti()} to control how
those functions sort arrays. Or you may provide one of the predefined control
strings that work for @code{PROCINFO["sorted_in"]}.

@item
You can use the @samp{|&} operator to create a two-way pipe to a coprocess.
You read from the coprocess with @code{getline} and write to it with @code{print}
or @code{printf}. Use @code{close()} to close off the coprocess completely, or
optionally, close off one side of the two-way communications.

@item
By using special @value{FN}s with the @samp{|&} operator, you can open a
TCP/IP (or UDP/IP) connection to remote hosts on the Internet. @command{gawk}
supports both IPv4 and IPv6.

@item
You can generate statement count profiles of your program. This can help you
determine which parts of your program may be taking the most time and let
you tune them more easily.  Sending the @code{USR1} signal while profiling causes
@command{gawk} to dump the profile and keep going, including a function call stack.

@item
You can also just ``pretty-print'' the program.

@end itemize


@node Internationalization
@chapter Internationalization with @command{gawk}

Once upon a time, computer makers
wrote software that worked only in English.
Eventually, hardware and software vendors noticed that if their
systems worked in the native languages of non-English-speaking
countries, they were able to sell more systems.
As a result, internationalization and localization
of programs and software systems became a common practice.

@cindex internationalization, localization
@cindex @command{gawk}, internationalization and, See internationalization
@cindex internationalization, localization, @command{gawk} and
For many years, the ability to provide internationalization
was largely restricted to programs written in C and C++.
This @value{CHAPTER} describes the underlying library @command{gawk}
uses for internationalization, as well as how
@command{gawk} makes internationalization
features available at the @command{awk} program level.
Having internationalization available at the @command{awk} level
gives software developers additional flexibility---they are no
longer forced to write in C or C++ when internationalization is
a requirement.

@menu
* I18N and L10N::               Internationalization and Localization.
* Explaining gettext::          How GNU @command{gettext} works.
* Programmer i18n::             Features for the programmer.
* Translator i18n::             Features for the translator.
* I18N Example::                A simple i18n example.
* Gawk I18N::                   @command{gawk} is also internationalized.
* I18N Summary::                Summary of I18N stuff.
@end menu

@node I18N and L10N
@section Internationalization and Localization

@cindex internationalization
@cindex localization, See internationalization@comma{} localization
@cindex localization
@dfn{Internationalization} means writing (or modifying) a program once,
in such a way that it can use multiple languages without requiring
further source code changes.
@dfn{Localization} means providing the data necessary for an
internationalized program to work in a particular language.
Most typically, these terms refer to features such as the language
used for printing error messages, the language used to read
responses, and information related to how numerical and
monetary values are printed and read.

@node Explaining gettext
@section GNU @command{gettext}

@cindex internationalizing a program
@cindex @command{gettext} library
@command{gawk} uses GNU @command{gettext} to provide its internationalization
features.
The facilities in GNU @command{gettext} focus on messages: strings printed
by a program, either directly or via formatting with @code{printf} or
@code{sprintf()}.@footnote{For some operating systems, the @command{gawk}
port doesn't support GNU @command{gettext}.
Therefore, these features are not available
if you are using one of those operating systems. Sorry.}

@cindex portability, @command{gettext} library and
When using GNU @command{gettext}, each application has its own
@dfn{text domain}.  This is a unique name, such as @samp{kpilot} or @samp{gawk},
that identifies the application.
A complete application may have multiple components---programs written
in C or C++, as well as scripts written in @command{sh} or @command{awk}.
All of the components use the same text domain.

To make the discussion concrete, assume we're writing an application
named @command{guide}.  Internationalization consists of the
following steps, in this order:

@enumerate
@item
The programmer reviews the source for all of @command{guide}'s components
and marks each string that is a candidate for translation.
For example, @code{"`-F': option required"} is a good candidate for translation.
A table with strings of option names is not (e.g., @command{gawk}'s
@option{--profile} option should remain the same, no matter what the local
language).

@cindex @code{textdomain()} function (C library)
@item
The programmer indicates the application's text domain
(@command{"guide"}) to the @command{gettext} library,
by calling the @code{textdomain()} function.

@cindex @code{.pot} files
@cindex files, @code{.pot}
@cindex portable object template files
@cindex files, portable object template
@item
Messages from the application are extracted from the source code and
collected into a portable object template file (@file{guide.pot}),
which lists the strings and their translations.
The translations are initially empty.
The original (usually English) messages serve as the key for
lookup of the translations.

@cindex @code{.po} files
@cindex files, @code{.po}
@cindex portable object files
@cindex files, portable object
@item
For each language with a translator, @file{guide.pot}
is copied to a portable object file (@code{.po})
and translations are created and shipped with the application.
For example, there might be a @file{fr.po} for a French translation.

@cindex @code{.gmo} files
@cindex files, @code{.gmo}
@cindex message object files
@cindex files, message object
@item
Each language's @file{.po} file is converted into a binary
message object (@file{.gmo}) file.
A message object file contains the original messages and their
translations in a binary format that allows fast lookup of translations
at runtime.

@item
When @command{guide} is built and installed, the binary translation files
are installed in a standard place.

@cindex @code{bindtextdomain()} function (C library)
@item
For testing and development, it is possible to tell @command{gettext}
to use @file{.gmo} files in a different directory than the standard
one by using the @code{bindtextdomain()} function.

@cindex @code{.gmo} files, specifying directory of
@cindex files, @code{.gmo}, specifying directory of
@cindex message object files, specifying directory of
@cindex files, message object, specifying directory of
@item
At runtime, @command{guide} looks up each string via a call
to @code{gettext()}.  The returned string is the translated string
if available, or the original string if not.

@item
If necessary, it is possible to access messages from a different
text domain than the one belonging to the application, without
having to switch the application's default text domain back
and forth.
@end enumerate

@cindex @code{gettext()} function (C library)
In C (or C++), the string marking and dynamic translation lookup
are accomplished by wrapping each string in a call to @code{gettext()}:

@example
printf("%s", gettext("Don't Panic!\n"));
@end example

The tools that extract messages from source code pull out all
strings enclosed in calls to @code{gettext()}.

@cindex @code{_} (underscore), C macro
@cindex underscore (@code{_}), C macro
The GNU @command{gettext} developers, recognizing that typing
@samp{gettext(@dots{})} over and over again is both painful and ugly to look
at, use the macro @samp{_} (an underscore) to make things easier:

@example
/* In the standard header file: */
#define _(str) gettext(str)

/* In the program text: */
printf("%s", _("Don't Panic!\n"));
@end example

@cindex internationalization, localization, locale categories
@cindex @command{gettext} library, locale categories
@cindex locale categories
@noindent
This reduces the typing overhead to just three extra characters per string
and is considerably easier to read as well.

There are locale @dfn{categories}
for different types of locale-related information.
The defined locale categories that @command{gettext} knows about are:

@table @code
@cindex @code{LC_MESSAGES} locale category
@item LC_MESSAGES
Text messages.  This is the default category for @command{gettext}
operations, but it is possible to supply a different one explicitly,
if necessary.  (It is almost never necessary to supply a different category.)

@cindex sorting characters in different languages
@cindex @code{LC_COLLATE} locale category
@item LC_COLLATE
Text-collation information (i.e., how different characters
and/or groups of characters sort in a given language).

@cindex @code{LC_CTYPE} locale category
@item LC_CTYPE
Character-type information (alphabetic, digit, upper- or lowercase, and
so on) as well as character encoding.
@ignore
In June 2001 Bruno Haible wrote:
- Description of LC_CTYPE: It determines both
  1. character encoding,
  2. character type information.
  (For example, in both KOI8-R and ISO-8859-5 the character type information
  is the same - cyrillic letters could as 'alpha' - but the encoding is
  different.)
@end ignore
This information is accessed via the
POSIX character classes in regular expressions,
such as @code{/[[:alnum:]]/}
(@pxref{Bracket Expressions}).

@cindex monetary information, localization
@cindex currency symbols, localization
@cindex @code{LC_MONETARY} locale category
@item LC_MONETARY
Monetary information, such as the currency symbol, and whether the
symbol goes before or after a number.

@cindex @code{LC_NUMERIC} locale category
@item LC_NUMERIC
Numeric information, such as which characters to use for the decimal
point and the thousands separator.@footnote{Americans
use a comma every three decimal places and a period for the decimal
point, while many Europeans do exactly the opposite:
1,234.56 versus 1.234,56.}

@cindex time, localization and
@cindex dates, information related to@comma{} localization
@cindex @code{LC_TIME} locale category
@item LC_TIME
Time- and date-related information, such as 12- or 24-hour clock, month printed
before or after the day in a date, local month abbreviations, and so on.

@cindex @code{LC_ALL} locale category
@item LC_ALL
All of the above.  (Not too useful in the context of @command{gettext}.)
@end table

@quotation NOTE
@cindex @env{LANGUAGE} environment variable
As described in @ref{Locales}, environment variables with the same
name as the locale categories (@env{LC_CTYPE}, @env{LC_ALL}, etc.)
influence @command{gawk}'s behavior (and that of other utilities).

Normally, these variables also affect how the @code{gettext} library
finds translations.  However, the @env{LANGUAGE} environment variable
overrides the @env{LC_@var{xxx}} variables. Many GNU/Linux systems
may define this variable without your knowledge, causing @command{gawk}
to not find the correct translations.  If this happens to you,
look to see if @env{LANGUAGE} is defined, and if so, use the shell's
@command{unset} command to remove it.
@end quotation

For testing translations of @command{gawk} itself, you can set
the @env{GAWK_LOCALE_DIR} environment variable. See the documentation
for the C @code{bindtextdomain()} function and also see
@ref{Other Environment Variables}.

@node Programmer i18n
@section Internationalizing @command{awk} Programs
@cindex @command{awk} programs, internationalizing

@command{gawk} provides the following variables for
internationalization:

@table @code
@cindex @code{TEXTDOMAIN} variable
@item TEXTDOMAIN
This variable indicates the application's text domain.
For compatibility with GNU @command{gettext}, the default
value is @code{"messages"}.

@cindex internationalization, localization, marked strings
@cindex strings, for localization
@item _"your message here"
String constants marked with a leading underscore
are candidates for translation at runtime.
String constants without a leading underscore are not translated.
@end table

@command{gawk} provides the following functions for
internationalization:

@table @code
@cindexgawkfunc{dcgettext}
@item @code{dcgettext(@var{string}} [@code{,} @var{domain} [@code{,} @var{category}]]@code{)}
Return the translation of @var{string} in
text domain @var{domain} for locale category @var{category}.
The default value for @var{domain} is the current value of @code{TEXTDOMAIN}.
The default value for @var{category} is @code{"LC_MESSAGES"}.

If you supply a value for @var{category}, it must be a string equal to
one of the known locale categories described in
@ifnotinfo
the previous @value{SECTION}.
@end ifnotinfo
@ifinfo
@ref{Explaining gettext}.
@end ifinfo
You must also supply a text domain.  Use @code{TEXTDOMAIN} if
you want to use the current domain.

@quotation CAUTION
The order of arguments to the @command{awk} version
of the @code{dcgettext()} function is purposely different from the order for
the C version.  The @command{awk} version's order was
chosen to be simple and to allow for reasonable @command{awk}-style
default arguments.
@end quotation

@cindexgawkfunc{dcngettext}
@item @code{dcngettext(@var{string1}, @var{string2}, @var{number}} [@code{,} @var{domain} [@code{,} @var{category}]]@code{)}
Return the plural form used for @var{number} of the
translation of @var{string1} and @var{string2} in text domain
@var{domain} for locale category @var{category}. @var{string1} is the
English singular variant of a message, and @var{string2} is the English plural
variant of the same message.
The default value for @var{domain} is the current value of @code{TEXTDOMAIN}.
The default value for @var{category} is @code{"LC_MESSAGES"}.

The same remarks about argument order as for the @code{dcgettext()} function apply.

@cindex @code{.gmo} files, specifying directory of
@cindex files, @code{.gmo}, specifying directory of
@cindex message object files, specifying directory of
@cindex files, message object, specifying directory of
@cindexgawkfunc{bindtextdomain}
@item @code{bindtextdomain(@var{directory}} [@code{,} @var{domain} ]@code{)}
Change the directory in which
@command{gettext} looks for @file{.gmo} files, in case they
will not or cannot be placed in the standard locations
(e.g., during testing).
Return the directory in which @var{domain} is ``bound.''

The default @var{domain} is the value of @code{TEXTDOMAIN}.
If @var{directory} is the null string (@code{""}), then
@code{bindtextdomain()} returns the current binding for the
given @var{domain}.
@end table

To use these facilities in your @command{awk} program, follow these steps:

@enumerate
@cindex @code{BEGIN} pattern, @code{TEXTDOMAIN} variable and
@cindex @code{TEXTDOMAIN} variable, @code{BEGIN} pattern and
@item
Set the variable @code{TEXTDOMAIN} to the text domain of
your program.  This is best done in a @code{BEGIN} rule
(@pxref{BEGIN/END}),
or it can also be done via the @option{-v} command-line
option (@pxref{Options}):

@example
BEGIN @{
    TEXTDOMAIN = "guide"
    @dots{}
@}
@end example

@cindex @code{_} (underscore), translatable string
@cindex underscore (@code{_}), translatable string
@item
Mark all translatable strings with a leading underscore (@samp{_})
character.  It @emph{must} be adjacent to the opening
quote of the string.  For example:

@example
print _"hello, world"
x = _"you goofed"
printf(_"Number of users is %d\n", nusers)
@end example

@item
If you are creating strings dynamically, you can
still translate them, using the @code{dcgettext()}
built-in function:@footnote{Thanks to Bruno Haible for this
example.}

@example
if (groggy)
    message = dcgettext("%d customers disturbing me\n", "adminprog")
else
    message = dcgettext("enjoying %d customers\n", "adminprog")
printf(message, ncustomers)
@end example

Here, the call to @code{dcgettext()} supplies a different
text domain (@code{"adminprog"}) in which to find the
message, but it uses the default @code{"LC_MESSAGES"} category.

The previous example only works if @code{ncustomers} is greater than one.
This example would be better done with @code{dcngettext()}:

@example
if (groggy)
    message = dcngettext("%d customer disturbing me\n",
                         "%d customers disturbing me\n",
                         ncustomers, "adminprog")
else
    message = dcngettext("enjoying %d customer\n",
                         "enjoying %d customers\n",
                         ncustomers, "adminprog")
printf(message, ncustomers)
@end example


@cindex @code{LC_MESSAGES} locale category, @code{bindtextdomain()} function (@command{gawk})
@item
During development, you might want to put the @file{.gmo}
file in a private directory for testing.  This is done
with the @code{bindtextdomain()} built-in function:

@example
BEGIN @{
   TEXTDOMAIN = "guide"   # our text domain
   if (Testing) @{
       # where to find our files
       bindtextdomain("testdir")
       # joe is in charge of adminprog
       bindtextdomain("../joe/testdir", "adminprog")
   @}
   @dots{}
@}
@end example

@end enumerate

@xref{I18N Example}
for an example program showing the steps to create
and use translations from @command{awk}.

@node Translator i18n
@section Translating @command{awk} Programs

@cindex @code{.po} files
@cindex files, @code{.po}
@cindex portable object files
@cindex files, portable object
Once a program's translatable strings have been marked, they must
be extracted to create the initial @file{.pot} file.
As part of translation, it is often helpful to rearrange the order
in which arguments to @code{printf} are output.

@command{gawk}'s @option{--gen-pot} command-line option extracts
the messages and is discussed next.
After that, @code{printf}'s ability to
rearrange the order for @code{printf} arguments at runtime
is covered.

@menu
* String Extraction::           Extracting marked strings.
* Printf Ordering::             Rearranging @code{printf} arguments.
* I18N Portability::            @command{awk}-level portability issues.
@end menu

@node String Extraction
@subsection Extracting Marked Strings
@cindex strings, extracting
@cindex marked strings@comma{} extracting
@cindex @option{--gen-pot} option
@cindex command-line options, string extraction
@cindex string extraction (internationalization)
@cindex marked string extraction (internationalization)
@cindex extraction, of marked strings (internationalization)

@cindex @option{--gen-pot} option
Once your @command{awk} program is working, and all the strings have
been marked and you've set (and perhaps bound) the text domain,
it is time to produce translations.
First, use the @option{--gen-pot} command-line option to create
the initial @file{.pot} file:

@example
gawk --gen-pot -f guide.awk > guide.pot
@end example

@cindex @code{xgettext} utility
When run with @option{--gen-pot}, @command{gawk} does not execute your
program.  Instead, it parses it as usual and prints all marked strings
to standard output in the format of a GNU @command{gettext} Portable Object
file.  Also included in the output are any constant strings that
appear as the first argument to @code{dcgettext()} or as the first and
second argument to @code{dcngettext()}.@footnote{The
@command{xgettext} utility that comes with GNU
@command{gettext} can handle @file{.awk} files.}
You should distribute the generated @file{.pot} file with
your @command{awk} program; translators will eventually use it
to provide you translations that you can also then distribute.
@xref{I18N Example}
for the full list of steps to go through to create and test
translations for @command{guide}.

@node Printf Ordering
@subsection Rearranging @code{printf} Arguments

@cindex @code{printf} statement, positional specifiers
@cindex positional specifiers, @code{printf} statement
Format strings for @code{printf} and @code{sprintf()}
(@pxref{Printf})
present a special problem for translation.
Consider the following:@footnote{This example is borrowed
from the GNU @command{gettext} manual.}

@example
printf(_"String `%s' has %d characters\n",
          string, length(string)))
@end example

A possible German translation for this might be:

@example
"%d Zeichen lang ist die Zeichenkette `%s'\n"
@end example

The problem should be obvious: the order of the format
specifications is different from the original!
Even though @code{gettext()} can return the translated string
at runtime,
it cannot change the argument order in the call to @code{printf}.

To solve this problem, @code{printf} format specifiers may have
an additional optional element, which we call a @dfn{positional specifier}.
For example:

@example
"%2$d Zeichen lang ist die Zeichenkette `%1$s'\n"
@end example

Here, the positional specifier consists of an integer count, which indicates which
argument to use, and a @samp{$}. Counts are one-based, and the
format string itself is @emph{not} included.  Thus, in the following
example, @samp{string} is the first argument and @samp{length(string)} is the second:

@example
$ @kbd{gawk 'BEGIN @{}
>     @kbd{string = "Don\47t Panic"}
>     @kbd{printf "%2$d characters live in \"%1$s\"\n",}
>                         @kbd{string, length(string)}
> @kbd{@}'}
@print{} 11 characters live in "Don't Panic"
@end example

If present, positional specifiers come first in the format specification,
before the flags, the field width, and/or the precision.

Positional specifiers can be used with the dynamic field width and
precision capability:

@example
$ @kbd{gawk 'BEGIN @{}
>    @kbd{printf("%*.*s\n", 10, 20, "hello")}
>    @kbd{printf("%3$*2$.*1$s\n", 20, 10, "hello")}
> @kbd{@}'}
@print{}      hello
@print{}      hello
@end example

@quotation NOTE
When using @samp{*} with a positional specifier, the @samp{*}
comes first, then the integer position, and then the @samp{$}.
This is somewhat counterintuitive.
@end quotation

@cindex @code{printf} statement, positional specifiers, mixing with regular formats
@cindex positional specifiers, @code{printf} statement, mixing with regular formats
@cindex format specifiers, mixing regular with positional specifiers
@command{gawk} does not allow you to mix regular format specifiers
and those with positional specifiers in the same string:

@example
@group
$ @kbd{gawk 'BEGIN @{ printf "%d %3$s\n", 1, 2, "hi" @}'}
@error{} gawk: cmd. line:1: fatal: must use `count$' on all formats or none
@end group
@end example

@quotation NOTE
There are some pathological cases that @command{gawk} may fail to
diagnose.  In such cases, the output may not be what you expect.
It's still a bad idea to try mixing them, even if @command{gawk}
doesn't detect it.
@end quotation

Although positional specifiers can be used directly in @command{awk} programs,
their primary purpose is to help in producing correct translations of
format strings into languages different from the one in which the program
is first written.

@node I18N Portability
@subsection @command{awk} Portability Issues

@cindex portability, internationalization and
@cindex internationalization, localization, portability and
@command{gawk}'s internationalization features were purposely chosen to
have as little impact as possible on the portability of @command{awk}
programs that use them to other versions of @command{awk}.
Consider this program:

@example
BEGIN @{
    TEXTDOMAIN = "guide"
    if (Test_Guide)   # set with -v
        bindtextdomain("/test/guide/messages")
    print _"don't panic!"
@}
@end example

@noindent
As written, it won't work on other versions of @command{awk}.
However, it is actually almost portable, requiring very little
change:

@itemize @value{BULLET}
@cindex @code{TEXTDOMAIN} variable, portability and
@item
Assignments to @code{TEXTDOMAIN} won't have any effect,
because @code{TEXTDOMAIN} is not special in other @command{awk} implementations.

@item
Non-GNU versions of @command{awk} treat marked strings
as the concatenation of a variable named @code{_} with the string
following it.@footnote{This is good fodder for an ``Obfuscated
@command{awk}'' contest.} Typically, the variable @code{_} has
the null string (@code{""}) as its value, leaving the original string constant as
the result.

@item
By defining ``dummy'' functions to replace @code{dcgettext()}, @code{dcngettext()},
and @code{bindtextdomain()}, the @command{awk} program can be made to run, but
all the messages are output in the original language.
For example:

@cindex @code{bindtextdomain()} function (@command{gawk}), portability and
@cindex @code{dcgettext()} function (@command{gawk}), portability and
@cindex @code{dcngettext()} function (@command{gawk}), portability and
@example
@c file eg/lib/libintl.awk
function bindtextdomain(dir, domain)
@{
    return dir
@}

function dcgettext(string, domain, category)
@{
    return string
@}

function dcngettext(string1, string2, number, domain, category)
@{
    return (number == 1 ? string1 : string2)
@}
@c endfile
@end example

@item
The use of positional specifications in @code{printf} or
@code{sprintf()} is @emph{not} portable.
To support @code{gettext()} at the C level, many systems' C versions of
@code{sprintf()} do support positional specifiers.  But it works only if
enough arguments are supplied in the function call.  Many versions of
@command{awk} pass @code{printf} formats and arguments unchanged to the
underlying C library version of @code{sprintf()}, but only one format and
argument at a time.  What happens if a positional specification is
used is anybody's guess.
However, because the positional specifications are primarily for use in
@emph{translated} format strings, and because non-GNU @command{awk}s never
retrieve the translated string, this should not be a problem in practice.
@end itemize

@node I18N Example
@section A Simple Internationalization Example

Now let's look at a step-by-step example of how to internationalize and
localize a simple @command{awk} program, using @file{guide.awk} as our
original source:

@example
@c file eg/prog/guide.awk
BEGIN @{
    TEXTDOMAIN = "guide"
    bindtextdomain(".")  # for testing
    print _"Don't Panic"
    print _"The Answer Is", 42
    print "Pardon me, Zaphod who?"
@}
@c endfile
@end example

@noindent
Run @samp{gawk --gen-pot} to create the @file{.pot} file:

@example
$ @kbd{gawk --gen-pot -f guide.awk > guide.pot}
@end example

@noindent
This produces:

@example
@c file eg/data/guide.po
#: guide.awk:4
msgid "Don't Panic"
msgstr ""

#: guide.awk:5
msgid "The Answer Is"
msgstr ""

@c endfile
@end example

This original portable object template file is saved and reused for each language
into which the application is translated.  The @code{msgid}
is the original string and the @code{msgstr} is the translation.

@quotation NOTE
Strings not marked with a leading underscore do not
appear in the @file{guide.pot} file.
@end quotation

Next, the messages must be translated.
Here is a translation to a hypothetical dialect of English,
called ``Mellow'':@footnote{Perhaps it would be better if it were
called ``Hippy.'' Ah, well.}

@example
@group
$ @kbd{cp guide.pot guide-mellow.po}
@var{Add translations to} guide-mellow.po @dots{}
@end group
@end example

@noindent
Following are the translations:

@example
@c file eg/data/guide-mellow.po
#: guide.awk:4
msgid "Don't Panic"
msgstr "Hey man, relax!"

#: guide.awk:5
msgid "The Answer Is"
msgstr "Like, the scoop is"

@c endfile
@end example

@cindex Linux
@cindex GNU/Linux
The next step is to make the directory to hold the binary message object
file and then to create the @file{guide.mo} file.
We pretend that our file is to be used in the @code{en_US.UTF-8} locale,
because we have to use a locale name known to the C @command{gettext} routines.
The directory layout shown here is standard for GNU @command{gettext} on
GNU/Linux systems.  Other versions of @command{gettext} may use a different
layout:

@example
$ @kbd{mkdir en_US.UTF-8 en_US.UTF-8/LC_MESSAGES}
@end example

@cindex @code{.po} files, converting to @code{.mo}
@cindex files, @code{.po}, converting to @code{.mo}
@cindex @code{.mo} files, converting from @code{.po}
@cindex files, @code{.mo}, converting from @code{.po}
@cindex portable object files, converting to message object files
@cindex files, portable object, converting to message object files
@cindex message object files, converting from portable object files
@cindex files, message object, converting from portable object files
@cindex @command{msgfmt} utility
The @command{msgfmt} utility does the conversion from human-readable
@file{.po} file to machine-readable @file{.mo} file.
By default, @command{msgfmt} creates a file named @file{messages}.
This file must be renamed and placed in the proper directory (using
the @option{-o} option) so that @command{gawk} can find it:

@example
$ @kbd{msgfmt guide-mellow.po -o en_US.UTF-8/LC_MESSAGES/guide.mo}
@end example

Finally, we run the program to test it:

@example
$ @kbd{gawk -f guide.awk}
@print{} Hey man, relax!
@print{} Like, the scoop is 42
@print{} Pardon me, Zaphod who?
@end example

If the three replacement functions for @code{dcgettext()}, @code{dcngettext()},
and @code{bindtextdomain()}
(@pxref{I18N Portability})
are in a file named @file{libintl.awk},
then we can run @file{guide.awk} unchanged as follows:

@example
$ @kbd{gawk --posix -f guide.awk -f libintl.awk}
@print{} Don't Panic
@print{} The Answer Is 42
@print{} Pardon me, Zaphod who?
@end example

@node Gawk I18N
@section @command{gawk} Can Speak Your Language

@command{gawk} itself has been internationalized
using the GNU @command{gettext} package.
(GNU @command{gettext} is described in
complete detail in
@ifinfo
@inforef{Top, , GNU @command{gettext} utilities, gettext, GNU @command{gettext} utilities}.)
@end ifinfo
@ifnotinfo
@uref{https://www.gnu.org/software/gettext/manual/,
@cite{GNU @command{gettext} utilities}}.)
@end ifnotinfo
As of this writing, the latest version of GNU @command{gettext} is
@uref{ftp://ftp.gnu.org/gnu/gettext/gettext-0.19.4.tar.gz,
@value{PVERSION} 0.19.4}.

If a translation of @command{gawk}'s messages exists,
then @command{gawk} produces usage messages, warnings,
and fatal errors in the local language.

@node I18N Summary
@section Summary

@itemize @value{BULLET}
@item
Internationalization means writing a program such that it can use multiple
languages without requiring source code changes.  Localization means
providing the data necessary for an internationalized program to work
in a particular language.

@item
@command{gawk} uses GNU @command{gettext} to let you internationalize
and localize @command{awk} programs.  A program's text domain identifies
the program for grouping all messages and other data together.

@item
You mark a program's strings for translation by preceding them with
an underscore. Once that is done, the strings are extracted into a
@file{.pot} file.  This file is copied for each language into a @file{.po}
file, and the @file{.po} files are compiled into @file{.gmo} files for
use at runtime.

@item
You can use positional specifications with @code{sprintf()} and
@code{printf} to rearrange the placement of argument values in formatted
strings and output. This is useful for the translation of format
control strings.

@item
The internationalization features have been designed so that they
can be easily worked around in a standard @command{awk}.

@item
@command{gawk} itself has been internationalized and ships with
a number of translations for its messages.

@end itemize


@node Debugger
@chapter Debugging @command{awk} Programs
@cindex debugging @command{awk} programs

@c The original text for this chapter was contributed by Efraim Yawitz.
@c FIXME: Add more indexing.

It would be nice if computer programs worked perfectly the first time they
were run, but in real life, this rarely happens for programs of
any complexity.  Thus, most programming languages have facilities available
for ``debugging'' programs, and now @command{awk} is no exception.

The @command{gawk} debugger is purposely modeled after
@uref{https://www.gnu.org/software/gdb/, the GNU Debugger (GDB)}
command-line debugger.  If you are familiar with GDB, learning
how to use @command{gawk} for debugging your program is easy.

@menu
* Debugging::                   Introduction to @command{gawk} debugger.
* Sample Debugging Session::    Sample debugging session.
* List of Debugger Commands::   Main debugger commands.
* Readline Support::            Readline support.
* Limitations::                 Limitations and future plans.
* Debugging Summary::           Debugging summary.
@end menu

@node Debugging
@section Introduction to the @command{gawk} Debugger

This @value{SECTION} introduces debugging in general and begins
the discussion of debugging in @command{gawk}.

@menu
* Debugging Concepts::          Debugging in General.
* Debugging Terms::             Additional Debugging Concepts.
* Awk Debugging::               Awk Debugging.
@end menu

@node Debugging Concepts
@subsection Debugging in General

(If you have used debuggers in other languages, you may want to skip
ahead to @ref{Awk Debugging}.)

Of course, a debugging program cannot remove bugs for you, because it has
no way of knowing what you or your users consider a ``bug'' versus a
``feature.''  (Sometimes, we humans have a hard time with this ourselves.)
In that case, what can you expect from such a tool?  The answer to that
depends on the language being debugged, but in general, you can expect at
least the following:

@itemize @value{BULLET}
@item
The ability to watch a program execute its instructions one by one,
giving you, the programmer, the opportunity to think about what is happening
on a time scale of seconds, minutes, or hours, rather than the nanosecond
time scale at which the code usually runs.

@item
The opportunity to not only passively observe the operation of your
program, but to control it and try different paths of execution, without
having to change your source files.

@item
The chance to see the values of data in the program at any point in
execution, and also to change that data on the fly, to see how that
affects what happens afterward.  (This often includes the ability
to look at internal data structures besides the variables you actually
defined in your code.)

@item
The ability to obtain additional information about your program's state
or even its internal structure.
@end itemize

All of these tools provide a great amount of help in using your own
skills and understanding of the goals of your program to find where it
is going wrong (or, for that matter, to better comprehend a perfectly
functional program that you or someone else wrote).

@node Debugging Terms
@subsection Debugging Concepts

Before diving in to the details, we need to introduce several
important concepts that apply to just about all debuggers.
The following list defines terms used throughout the rest of
this @value{CHAPTER}:

@table @dfn
@cindex stack frame
@item Stack frame
Programs generally call functions during the course of their execution.
One function can call another, or a function can call itself (recursion).
You can view the chain of called functions (main program calls A, which
calls B, which calls C), as a stack of executing functions: the currently
running function is the topmost one on the stack, and when it finishes
(returns), the next one down then becomes the active function.
Such a stack is termed a @dfn{call stack}.

For each function on the call stack, the system maintains a data area
that contains the function's parameters, local variables, and return value,
as well as any other ``bookkeeping'' information needed to manage the
call stack.  This data area is termed a @dfn{stack frame}.

@command{gawk} also follows this model, and gives you
access to the call stack and to each stack frame. You can see the
call stack, as well as from where each function on the stack was
invoked. Commands that print the call stack print information about
each stack frame (as detailed later on).

@item Breakpoint
@cindex breakpoint
During debugging, you often wish to let the program run until it
reaches a certain point, and then continue execution from there one
statement (or instruction) at a time.  The way to do this is to set
a @dfn{breakpoint} within the program.  A breakpoint is where the
execution of the program should break off (stop), so that you can
take over control of the program's execution.  You can add and remove
as many breakpoints as you like.

@item Watchpoint
@cindex watchpoint
A watchpoint is similar to a breakpoint.  The difference is that
breakpoints are oriented around the code: stop when a certain point in the
code is reached.  A watchpoint, however, specifies that program execution
should stop when a @emph{data value} is changed.  This is useful, as
sometimes it happens that a variable receives an erroneous value, and it's
hard to track down where this happens just by looking at the code.
By using a watchpoint, you can stop whenever a variable is assigned to,
and usually find the errant code quite quickly.
@end table

@node Awk Debugging
@subsection @command{awk} Debugging

Debugging an @command{awk} program has some specific aspects that are
not shared with programs written in other languages.

First of all, the fact that @command{awk} programs usually take input
line by line from a file or files and operate on those lines using specific
rules makes it especially useful to organize viewing the execution of
the program in terms of these rules.  As we will see, each @command{awk}
rule is treated almost like a function call, with its own specific block
of instructions.

In addition, because @command{awk} is by design a very concise language,
it is easy to lose sight of everything that is going on ``inside''
each line of @command{awk} code.  The debugger provides the opportunity
to look at the individual primitive instructions carried out
by the higher-level @command{awk} commands.

@node Sample Debugging Session
@section Sample @command{gawk} Debugging Session
@cindex sample debugging session

In order to illustrate the use of @command{gawk} as a debugger, let's look at a sample
debugging session.  We will use the @command{awk} implementation of the
POSIX @command{uniq} command described earlier (@pxref{Uniq Program})
as our example.

@menu
* Debugger Invocation::         How to Start the Debugger.
* Finding The Bug::             Finding the Bug.
@end menu

@node Debugger Invocation
@subsection How to Start the Debugger
@cindex starting the debugger
@cindex debugger, how to start

Starting the debugger is almost exactly like running @command{gawk} normally,
except you have to pass an additional option, @option{--debug}, or the
corresponding short option, @option{-D}.  The file(s) containing the
program and any supporting code are given on the command line as arguments
to one or more @option{-f} options. (@command{gawk} is not designed
to debug command-line programs, only programs contained in files.)
In our case, we invoke the debugger like this:

@example
$ @kbd{gawk -D -f getopt.awk -f join.awk -f uniq.awk -1 inputfile}
@end example

@noindent
where both @file{getopt.awk} and @file{uniq.awk} are in @env{$AWKPATH}.
(Experienced users of GDB or similar debuggers should note that
this syntax is slightly different from what you are used to.
With the @command{gawk} debugger, you give the arguments for running the program
in the command line to the debugger rather than as part of the @code{run}
command at the debugger prompt.)
The @option{-1} is an option to @file{uniq.awk}.

Instead of immediately running the program on @file{inputfile}, as
@command{gawk} would ordinarily do, the debugger merely loads all
the program source files, compiles them internally, and then gives
us a prompt:

@example
gawk>
@end example

@noindent
from which we can issue commands to the debugger.  At this point, no
code has been executed.

@node Finding The Bug
@subsection Finding the Bug

Let's say that we are having a problem using (a faulty version of)
@file{uniq.awk} in the ``field-skipping'' mode, and it doesn't seem to be
catching lines which should be identical when skipping the first field,
such as:

@example
awk is a wonderful program!
gawk is a wonderful program!
@end example

This could happen if we were thinking (C-like) of the fields in a record
as being numbered in a zero-based fashion, so instead of the lines:

@example
clast = join(alast, fcount+1, n)
cline = join(aline, fcount+1, m)
@end example

@noindent
we wrote:

@example
clast = join(alast, fcount, n)
cline = join(aline, fcount, m)
@end example

The first thing we usually want to do when trying to investigate a
problem like this is to put a breakpoint in the program so that we can
watch it at work and catch what it is doing wrong.  A reasonable spot for
a breakpoint in @file{uniq.awk} is at the beginning of the function
@code{are_equal()}, which compares the current line with the previous one. To set
the breakpoint, use the @code{b} (breakpoint) command:

@example
gawk> @kbd{b are_equal}
@print{} Breakpoint 1 set at file `awklib/eg/prog/uniq.awk', line 63
@end example

The debugger tells us the file and line number where the breakpoint is.
Now type @samp{r} or @samp{run} and the program runs until it hits
the breakpoint for the first time:

@example
gawk> @kbd{r}
@print{} Starting program:
@print{} Stopping in Rule ...
@print{} Breakpoint 1, are_equal(n, m, clast, cline, alast, aline)
         at `awklib/eg/prog/uniq.awk':63
@print{} 63          if (fcount == 0 && charcount == 0)
gawk>
@end example

Now we can look at what's going on inside our program.  First of all,
let's see how we got to where we are.  At the prompt, we type @samp{bt}
(short for ``backtrace''), and the debugger responds with a
listing of the current stack frames:

@example
gawk> @kbd{bt}
@print{} #0  are_equal(n, m, clast, cline, alast, aline)
         at `awklib/eg/prog/uniq.awk':68
@print{} #1  in main() at `awklib/eg/prog/uniq.awk':88
@end example

This tells us that @code{are_equal()} was called by the main program at
line 88 of @file{uniq.awk}.  (This is not a big surprise, because this
is the only call to @code{are_equal()} in the program, but in more complex
programs, knowing who called a function and with what parameters can be
the key to finding the source of the problem.)

Now that we're in @code{are_equal()}, we can start looking at the values
of some variables.  Let's say we type @samp{p n}
(@code{p} is short for ``print'').  We would expect to see the value of
@code{n}, a parameter to @code{are_equal()}.  Actually, the debugger
gives us:

@example
gawk> @kbd{p n}
@print{} n = untyped variable
@end example

@noindent
In this case, @code{n} is an uninitialized local variable, because the
function was called without arguments (@pxref{Function Calls}).

A more useful variable to display might be the current record:

@example
gawk> @kbd{p $0}
@print{} $0 = "gawk is a wonderful program!"
@end example

@noindent
This might be a bit puzzling at first, as this is the second line of
our test input.  Let's look at @code{NR}:

@example
gawk> @kbd{p NR}
@print{} NR = 2
@end example

@noindent
So we can see that @code{are_equal()} was only called for the second record
of the file.  Of course, this is because our program contains a rule for
@samp{NR == 1}:

@example
NR == 1 @{
    last = $0
    next
@}
@end example

OK, let's just check that that rule worked correctly:

@example
gawk> @kbd{p last}
@print{} last = "awk is a wonderful program!"
@end example

Everything we have done so far has verified that the program has worked as
planned, up to and including the call to @code{are_equal()}, so the problem must
be inside this function.  To investigate further, we must begin
``stepping through'' the lines of @code{are_equal()}.  We start by typing
@samp{n} (for ``next''):

@example
@group
gawk> @kbd{n}
@print{} 66          if (fcount > 0) @{
@end group
@end example

This tells us that @command{gawk} is now ready to execute line 66, which
decides whether to give the lines the special ``field-skipping'' treatment
indicated by the @option{-1} command-line option.  (Notice that we skipped
from where we were before, at line 63, to here, because the condition
in line 63, @samp{if (fcount == 0 && charcount == 0)}, was false.)

Continuing to step, we now get to the splitting of the current and
last records:

@example
gawk> @kbd{n}
@print{} 67              n = split(last, alast)
gawk> @kbd{n}
@print{} 68              m = split($0, aline)
@end example

At this point, we should be curious to see what our records were split
into, so we try to look:

@example
gawk> @kbd{p n m alast aline}
@print{} n = 5
@print{} m = untyped variable
@print{} alast = array, 5 elements
@print{} aline = untyped variable
@end example

@noindent
(The @code{p} command can take more than one argument, similar to
@command{awk}'s @code{print} statement.)

This is kind of disappointing, though.  All we found out is that there
are five elements in @code{alast}; @code{m} and @code{aline} don't have
values because we are at line 68 but haven't executed it yet.
This information is useful enough (we now know that
none of the words were accidentally left out), but what if we want to see
inside the array?

The first choice would be to use subscripts:

@example
gawk> @kbd{p alast[0]}
@print{} "0" not in array `alast'
@end example

@noindent
Oops!

@example
gawk> @kbd{p alast[1]}
@print{} alast["1"] = "awk"
@end example

This would be kind of slow for a 100-member array, though, so
@command{gawk} provides a shortcut (reminiscent of another language
not to be mentioned):

@example
gawk> @kbd{p @@alast}
@print{} alast["1"] = "awk"
@print{} alast["2"] = "is"
@print{} alast["3"] = "a"
@print{} alast["4"] = "wonderful"
@print{} alast["5"] = "program!"
@end example

It looks like we got this far OK.  Let's take another step
or two:

@example
gawk> @kbd{n}
@print{} 69              clast = join(alast, fcount, n)
gawk> @kbd{n}
@print{} 70              cline = join(aline, fcount, m)
@end example

Well, here we are at our error (sorry to spoil the suspense).  What we
had in mind was to join the fields starting from the second one to make
the virtual record to compare, and if the first field were numbered zero,
this would work.  Let's look at what we've got:

@example
gawk> @kbd{p cline clast}
@print{} cline = "gawk is a wonderful program!"
@print{} clast = "awk is a wonderful program!"
@end example

Hey, those look pretty familiar!  They're just our original, unaltered
input records.  A little thinking (the human brain is still the best
debugging tool), and we realize that we were off by one!

We get out of the debugger:

@example
gawk> @kbd{q}
@print{} The program is running. Exit anyway (y/n)? @kbd{y}
@end example

@noindent
Then we get into an editor:

@example
clast = join(alast, fcount+1, n)
cline = join(aline, fcount+1, m)
@end example

@noindent
and problem solved!

@node List of Debugger Commands
@section Main Debugger Commands

The @command{gawk} debugger command set can be divided into the
following categories:

@itemize @value{BULLET}

@item
Breakpoint control

@item
Execution control

@item
Viewing and changing data

@item
Working with the stack

@item
Getting information

@item
Miscellaneous
@end itemize

Each of these are discussed in the following subsections.
In the following descriptions, commands that may be abbreviated
show the abbreviation on a second description line.
A debugger command name may also be truncated if that partial
name is unambiguous. The debugger has the built-in capability to
automatically repeat the previous command just by hitting @kbd{Enter}.
This works for the commands @code{list}, @code{next}, @code{nexti},
@code{step}, @code{stepi}, and @code{continue} executed without any
argument.

@menu
* Breakpoint Control::          Control of Breakpoints.
* Debugger Execution Control::  Control of Execution.
* Viewing And Changing Data::   Viewing and Changing Data.
* Execution Stack::             Dealing with the Stack.
* Debugger Info::               Obtaining Information about the Program and
                                the Debugger State.
* Miscellaneous Debugger Commands:: Miscellaneous Commands.
@end menu

@node Breakpoint Control
@subsection Control of Breakpoints

As we saw earlier, the first thing you probably want to do in a debugging
session is to get your breakpoints set up, because your program
will otherwise just run as if it was not under the debugger.  The commands for
controlling breakpoints are:

@table @asis
@cindex debugger commands, @code{b} (@code{break})
@cindex debugger commands, @code{break}
@cindex @code{break} debugger command
@cindex @code{b} debugger command (alias for @code{break})
@cindex set breakpoint
@cindex breakpoint, setting
@item @code{break} [[@var{filename}@code{:}]@var{n} | @var{function}] [@code{"@var{expression}"}]
@itemx @code{b} [[@var{filename}@code{:}]@var{n} | @var{function}] [@code{"@var{expression}"}]
Without any argument, set a breakpoint at the next instruction
to be executed in the selected stack frame.
Arguments can be one of the following:

@c @asis for docbook
@c nested table
@table @asis
@item @var{n}
Set a breakpoint at line number @var{n} in the current source file.

@item @var{filename}@code{:}@var{n}
Set a breakpoint at line number @var{n} in source file @var{filename}.

@item @var{function}
Set a breakpoint at entry to (the first instruction of)
function @var{function}.
@end table

Each breakpoint is assigned a number that can be used to delete it from
the breakpoint list using the @code{delete} command.

With a breakpoint, you may also supply a condition.  This is an
@command{awk} expression (enclosed in double quotes) that the debugger
evaluates whenever the breakpoint is reached. If the condition is true,
then the debugger stops execution and prompts for a command. Otherwise,
it continues executing the program.

@cindex debugger commands, @code{clear}
@cindex @code{clear} debugger command
@cindex delete breakpoint at location
@cindex breakpoint at location, how to delete
@item @code{clear} [[@var{filename}@code{:}]@var{n} | @var{function}]
Without any argument, delete any breakpoint at the next instruction
to be executed in the selected stack frame. If the program stops at
a breakpoint, this deletes that breakpoint so that the program
does not stop at that location again.  Arguments can be one of the following:

@c nested table
@table @asis
@item @var{n}
Delete breakpoint(s) set at line number @var{n} in the current source file.

@item @var{filename}@code{:}@var{n}
Delete breakpoint(s) set at line number @var{n} in source file @var{filename}.

@item @var{function}
Delete breakpoint(s) set at entry to function @var{function}.
@end table

@cindex debugger commands, @code{condition}
@cindex @code{condition} debugger command
@cindex breakpoint condition
@item @code{condition} @var{n} @code{"@var{expression}"}
Add a condition to existing breakpoint or watchpoint @var{n}. The
condition is an @command{awk} expression @emph{enclosed in double quotes}
that the debugger evaluates
whenever the breakpoint or watchpoint is reached. If the condition is true, then
the debugger stops execution and prompts for a command. Otherwise,
the debugger continues executing the program. If the condition expression is
not specified, any existing condition is removed (i.e., the breakpoint or
watchpoint is made unconditional).

@cindex debugger commands, @code{d} (@code{delete})
@cindex debugger commands, @code{delete}
@cindex @code{delete} debugger command
@cindex @code{d} debugger command (alias for @code{delete})
@cindex delete breakpoint by number
@cindex breakpoint, delete by number
@item @code{delete} [@var{n1 n2} @dots{}] [@var{n}--@var{m}]
@itemx @code{d} [@var{n1 n2} @dots{}] [@var{n}--@var{m}]
Delete specified breakpoints or a range of breakpoints. Delete
all defined breakpoints if no argument is supplied.

@cindex debugger commands, @code{disable}
@cindex @code{disable} debugger command
@cindex disable breakpoint
@cindex breakpoint, how to disable or enable
@item @code{disable} [@var{n1 n2} @dots{} | @var{n}--@var{m}]
Disable specified breakpoints or a range of breakpoints. Without
any argument, disable all breakpoints.

@cindex debugger commands, @code{e} (@code{enable})
@cindex debugger commands, @code{enable}
@cindex @code{enable} debugger command
@cindex @code{e} debugger command (alias for @code{enable})
@cindex enable breakpoint
@item @code{enable} [@code{del} | @code{once}] [@var{n1 n2} @dots{}] [@var{n}--@var{m}]
@itemx @code{e} [@code{del} | @code{once}] [@var{n1 n2} @dots{}] [@var{n}--@var{m}]
Enable specified breakpoints or a range of breakpoints. Without
any argument, enable all breakpoints.
Optionally, you can specify how to enable the breakpoints:

@c nested table
@table @code
@item del
Enable the breakpoints temporarily, then delete each one when
the program stops at it.

@item once
Enable the breakpoints temporarily, then disable each one when
the program stops at it.
@end table

@cindex debugger commands, @code{ignore}
@cindex @code{ignore} debugger command
@cindex ignore breakpoint
@item @code{ignore} @var{n} @var{count}
Ignore breakpoint number @var{n} the next @var{count} times it is
hit.

@cindex debugger commands, @code{t} (@code{tbreak})
@cindex debugger commands, @code{tbreak}
@cindex @code{tbreak} debugger command
@cindex @code{t} debugger command (alias for @code{tbreak})
@cindex temporary breakpoint
@item @code{tbreak} [[@var{filename}@code{:}]@var{n} | @var{function}]
@itemx @code{t} [[@var{filename}@code{:}]@var{n} | @var{function}]
Set a temporary breakpoint (enabled for only one stop).
The arguments are the same as for @code{break}.
@end table

@node Debugger Execution Control
@subsection Control of Execution

Now that your breakpoints are ready, you can start running the program
and observing its behavior.  There are more commands for controlling
execution of the program than we saw in our earlier example:

@table @asis
@cindex debugger commands, @code{commands}
@cindex @code{commands} debugger command
@cindex debugger commands, @code{silent}
@cindex @code{silent} debugger command
@cindex debugger commands, @code{end}
@cindex @code{end} debugger command
@cindex breakpoint commands
@cindex commands to execute at breakpoint
@item @code{commands} [@var{n}]
@itemx @code{silent}
@itemx @dots{}
@itemx @code{end}
Set a list of commands to be executed upon stopping at
a breakpoint or watchpoint. @var{n} is the breakpoint or watchpoint number.
Without a number, the last one set is used. The actual commands follow,
starting on the next line, and terminated by the @code{end} command.
If the command @code{silent} is in the list, the usual messages about
stopping at a breakpoint and the source line are not printed. Any command
in the list that resumes execution (e.g., @code{continue}) terminates the list
(an implicit @code{end}), and subsequent commands are ignored.
For example:

@example
gawk> @kbd{commands}
> @kbd{silent}
> @kbd{printf "A silent breakpoint; i = %d\n", i}
> @kbd{info locals}
> @kbd{set i = 10}
> @kbd{continue}
> @kbd{end}
gawk>
@end example

@cindex debugger commands, @code{c} (@code{continue})
@cindex debugger commands, @code{continue}
@cindex continue program, in debugger
@item @code{continue} [@var{count}]
@itemx @code{c} [@var{count}]
Resume program execution. If continued from a breakpoint and @var{count} is
specified, ignore the breakpoint at that location the next @var{count} times
before stopping.

@cindex debugger commands, @code{finish}
@cindex @code{finish} debugger command
@item @code{finish}
Execute until the selected stack frame returns.
Print the returned value.

@cindex debugger commands, @code{n} (@code{next})
@cindex debugger commands, @code{next}
@cindex @code{next} debugger command
@cindex @code{n} debugger command (alias for @code{next})
@cindex single-step execution, in the debugger
@item @code{next} [@var{count}]
@itemx @code{n} [@var{count}]
Continue execution to the next source line, stepping over function calls.
The argument @var{count} controls how many times to repeat the action, as
in @code{step}.

@cindex debugger commands, @code{ni} (@code{nexti})
@cindex debugger commands, @code{nexti}
@cindex @code{nexti} debugger command
@cindex @code{ni} debugger command (alias for @code{nexti})
@item @code{nexti} [@var{count}]
@itemx @code{ni} [@var{count}]
Execute one (or @var{count}) instruction(s), stepping over function calls.

@cindex debugger commands, @code{return}
@cindex @code{return} debugger command
@item @code{return} [@var{value}]
Cancel execution of a function call. If @var{value} (either a string or a
number) is specified, it is used as the function's return value. If used in a
frame other than the innermost one (the currently executing function; i.e.,
frame number 0), discard all inner frames in addition to the selected one,
and the caller of that frame becomes the innermost frame.

@cindex debugger commands, @code{r} (@code{run})
@cindex debugger commands, @code{run}
@cindex @code{run} debugger command
@cindex @code{r} debugger command (alias for @code{run})
@item @code{run}
@itemx @code{r}
Start/restart execution of the program. When restarting, the debugger
retains the current breakpoints, watchpoints, command history,
automatic display variables, and debugger options.

@cindex debugger commands, @code{s} (@code{step})
@cindex debugger commands, @code{step}
@cindex @code{step} debugger command
@cindex @code{s} debugger command (alias for @code{step})
@item @code{step} [@var{count}]
@itemx @code{s} [@var{count}]
Continue execution until control reaches a different source line in the
current stack frame, stepping inside any function called within
the line.  If the argument @var{count} is supplied, steps that many times before
stopping, unless it encounters a breakpoint or watchpoint.

@cindex debugger commands, @code{si} (@code{stepi})
@cindex debugger commands, @code{stepi}
@cindex @code{stepi} debugger command
@cindex @code{si} debugger command (alias for @code{stepi})
@item @code{stepi} [@var{count}]
@itemx @code{si} [@var{count}]
Execute one (or @var{count}) instruction(s), stepping inside function calls.
(For illustration of what is meant by an ``instruction'' in @command{gawk},
see the output shown under @code{dump} in @ref{Miscellaneous Debugger Commands}.)

@cindex debugger commands, @code{u} (@code{until})
@cindex debugger commands, @code{until}
@cindex @code{until} debugger command
@cindex @code{u} debugger command (alias for @code{until})
@item @code{until} [[@var{filename}@code{:}]@var{n} | @var{function}]
@itemx @code{u} [[@var{filename}@code{:}]@var{n} | @var{function}]
Without any argument, continue execution until a line past the current
line in the current stack frame is reached. With an argument,
continue execution until the specified location is reached, or the current
stack frame returns.
@end table

@node Viewing And Changing Data
@subsection Viewing and Changing Data

The commands for viewing and changing variables inside of @command{gawk} are:

@table @asis
@cindex debugger commands, @code{display}
@cindex @code{display} debugger command
@item @code{display} [@var{var} | @code{$}@var{n}]
Add variable @var{var} (or field @code{$@var{n}}) to the display list.
The value of the variable or field is displayed each time the program stops.
Each variable added to the list is identified by a unique number:

@example
gawk> @kbd{display x}
@print{} 10: x = 1
@end example

@noindent
This displays the assigned item number, the variable name, and its current value.
If the display variable refers to a function parameter, it is silently
deleted from the list as soon as the execution reaches a context where
no such variable of the given name exists.
Without argument, @code{display} displays the current values of
items on the list.

@cindex debugger commands, @code{eval}
@cindex @code{eval} debugger command
@cindex evaluate expressions, in debugger
@item @code{eval "@var{awk statements}"}
Evaluate @var{awk statements} in the context of the running program.
You can do anything that an @command{awk} program would do: assign
values to variables, call functions, and so on.

@item @code{eval} @var{param}, @dots{}
@itemx @var{awk statements}
@itemx @code{end}
This form of @code{eval} is similar, but it allows you to define
``local variables'' that exist in the context of the
@var{awk statements}, instead of using variables or function
parameters defined by the program.

@cindex debugger commands, @code{p} (@code{print})
@cindex debugger commands, @code{print}
@cindex @code{print} debugger command
@cindex @code{p} debugger command (alias for @code{print})
@cindex print variables, in debugger
@item @code{print} @var{var1}[@code{,} @var{var2} @dots{}]
@itemx @code{p} @var{var1}[@code{,} @var{var2} @dots{}]
Print the value of a @command{gawk} variable or field.
Fields must be referenced by constants:

@example
gawk> @kbd{print $3}
@end example

@noindent
This prints the third field in the input record (if the specified field does not
exist, it prints @samp{Null field}). A variable can be an array element, with
the subscripts being constant string values. To print the contents of an array,
prefix the name of the array with the @samp{@@} symbol:

@example
gawk> @kbd{print @@a}
@end example

@noindent
This prints the indices and the corresponding values for all elements in
the array @code{a}.

@cindex debugger commands, @code{printf}
@cindex @code{printf} debugger command
@item @code{printf} @var{format} [@code{,} @var{arg} @dots{}]
Print formatted text. The @var{format} may include escape sequences,
such as @samp{\n}
(@pxref{Escape Sequences}).
No newline is printed unless one is specified.

@cindex debugger commands, @code{set}
@cindex @code{set} debugger command
@cindex assign values to variables, in debugger
@item @code{set} @var{var}@code{=}@var{value}
Assign a constant (number or string) value to an @command{awk} variable
or field.
String values must be enclosed between double quotes (@code{"}@dots{}@code{"}).

You can also set special @command{awk} variables, such as @code{FS},
@code{NF}, @code{NR}, and so on.

@cindex debugger commands, @code{w} (@code{watch})
@cindex debugger commands, @code{watch}
@cindex @code{watch} debugger command
@cindex @code{w} debugger command (alias for @code{watch})
@cindex set watchpoint
@item @code{watch} @var{var} | @code{$}@var{n} [@code{"@var{expression}"}]
@itemx @code{w} @var{var} | @code{$}@var{n} [@code{"@var{expression}"}]
Add variable @var{var} (or field @code{$@var{n}}) to the watch list.
The debugger then stops whenever
the value of the variable or field changes. Each watched item is assigned a
number that can be used to delete it from the watch list using the
@code{unwatch} command.

With a watchpoint, you may also supply a condition.  This is an
@command{awk} expression (enclosed in double quotes) that the debugger
evaluates whenever the watchpoint is reached. If the condition is true,
then the debugger stops execution and prompts for a command. Otherwise,
@command{gawk} continues executing the program.

@cindex debugger commands, @code{undisplay}
@cindex @code{undisplay} debugger command
@cindex stop automatic display, in debugger
@item @code{undisplay} [@var{n}]
Remove item number @var{n} (or all items, if no argument) from the
automatic display list.

@cindex debugger commands, @code{unwatch}
@cindex @code{unwatch} debugger command
@cindex delete watchpoint
@item @code{unwatch} [@var{n}]
Remove item number @var{n} (or all items, if no argument) from the
watch list.

@end table

@node Execution Stack
@subsection Working with the Stack

Whenever you run a program that contains any function calls,
@command{gawk} maintains a stack of all of the function calls leading up
to where the program is right now.  You can see how you got to where you are,
and also move around in the stack to see what the state of things was in the
functions that called the one you are in.  The commands for doing this are:

@table @asis
@cindex debugger commands, @code{bt} (@code{backtrace})
@cindex debugger commands, @code{backtrace}
@cindex debugger commands, @code{where} (@code{backtrace})
@cindex @code{backtrace} debugger command
@cindex @code{bt} debugger command (alias for @code{backtrace})
@cindex @code{where} debugger command
@cindex @code{where} debugger command (alias for @code{backtrace})
@cindex call stack, display in debugger
@cindex traceback, display in debugger
@item @code{backtrace} [@var{count}]
@itemx @code{bt} [@var{count}]
@itemx @code{where} [@var{count}]
Print a backtrace of all function calls (stack frames), or innermost @var{count}
frames if @var{count} > 0. Print the outermost @var{count} frames if
@var{count} < 0.  The backtrace displays the name and arguments to each
function, the source @value{FN}, and the line number.
The alias @code{where} for @code{backtrace} is provided for longtime
GDB users who may be used to that command.

@cindex debugger commands, @code{down}
@cindex @code{down} debugger command
@item @code{down} [@var{count}]
Move @var{count} (default 1) frames down the stack toward the innermost frame.
Then select and print the frame.

@cindex debugger commands, @code{f} (@code{frame})
@cindex debugger commands, @code{frame}
@cindex @code{frame} debugger command
@cindex @code{f} debugger command (alias for @code{frame})
@item @code{frame} [@var{n}]
@itemx @code{f} [@var{n}]
Select and print stack frame @var{n}.  Frame 0 is the currently executing,
or @dfn{innermost}, frame (function call); frame 1 is the frame that
called the innermost one. The highest-numbered frame is the one for the
main program.  The printed information consists of the frame number,
function and argument names, source file, and the source line.

@cindex debugger commands, @code{up}
@cindex @code{up} debugger command
@item @code{up} [@var{count}]
Move @var{count} (default 1) frames up the stack toward the outermost frame.
Then select and print the frame.
@end table

@node Debugger Info
@subsection Obtaining Information About the Program and the Debugger State

Besides looking at the values of variables, there is often a need to get
other sorts of information about the state of your program and of the
debugging environment itself.  The @command{gawk} debugger has one command that
provides this information, appropriately called @code{info}.  @code{info}
is used with one of a number of arguments that tell it exactly what
you want to know:

@table @asis
@cindex debugger commands, @code{i} (@code{info})
@cindex debugger commands, @code{info}
@cindex @code{info} debugger command
@cindex @code{i} debugger command (alias for @code{info})
@item @code{info} @var{what}
@itemx @code{i} @var{what}
The value for @var{what} should be one of the following:

@c nested table
@table @code
@item args
@cindex show function arguments, in debugger
List arguments of the selected frame.

@item break
@cindex show breakpoints
List all currently set breakpoints.

@item display
@cindex automatic displays, in debugger
List all items in the automatic display list.

@item frame
@cindex describe call stack frame, in debugger
Give a description of the selected stack frame.

@item functions
@cindex list function definitions, in debugger
List all function definitions including source @value{FN}s and
line numbers.

@item locals
@cindex show local variables, in debugger
List local variables of the selected frame.

@item source
@cindex show name of current source file, in debugger
Print the name of the current source file. Each time the program stops, the
current source file is the file containing the current instruction.
When the debugger first starts, the current source file is the first file
included via the @option{-f} option. The
@samp{list @var{filename}:@var{lineno}} command can
be used at any time to change the current source.

@item sources
@cindex show all source files, in debugger
List all program sources.

@item variables
@cindex list all global variables, in debugger
List all global variables.

@item watch
@cindex show watchpoints
List all items in the watch list.
@end table
@end table

Additional commands give you control over the debugger, the ability to
save the debugger's state, and the ability to run debugger commands
from a file.  The commands are:

@table @asis
@cindex debugger commands, @code{o} (@code{option})
@cindex debugger commands, @code{option}
@cindex @code{option} debugger command
@cindex @code{o} debugger command (alias for @code{option})
@cindex display debugger options
@cindex debugger options
@item @code{option} [@var{name}[@code{=}@var{value}]]
@itemx @code{o} [@var{name}[@code{=}@var{value}]]
Without an argument, display the available debugger options
and their current values. @samp{option @var{name}} shows the current
value of the named option. @samp{option @var{name}=@var{value}} assigns
a new value to the named option.
The available options are:

@c nested table
@c asis for docbook
@table @asis
@item @code{history_size}
@cindex debugger history size
Set the maximum number of lines to keep in the history file
@file{./.gawk_history}.  The default is 100.

@item @code{listsize}
@cindex debugger default list amount
Specify the number of lines that @code{list} prints. The default is 15.

@item @code{outfile}
@cindex redirect @command{gawk} output, in debugger
Send @command{gawk} output to a file; debugger output still goes
to standard output. An empty string (@code{""}) resets output to
standard output.

@item @code{prompt}
@cindex debugger prompt
Change the debugger prompt. The default is @samp{@w{gawk> }}.

@item @code{save_history} [@code{on} | @code{off}]
@cindex debugger history file
Save command history to file @file{./.gawk_history}.
The default is @code{on}.

@item @code{save_options} [@code{on} | @code{off}]
@cindex save debugger options
Save current options to file @file{./.gawkrc} upon exit.
The default is @code{on}.
Options are read back into the next session upon startup.

@item @code{trace} [@code{on} | @code{off}]
@cindex instruction tracing, in debugger
Turn instruction tracing on or off. The default is @code{off}.
@end table

@item @code{save} @var{filename}
Save the commands from the current session to the given @value{FN},
so that they can be replayed using the @command{source} command.

@item @code{source} @var{filename}
@cindex debugger, read commands from a file
Run command(s) from a file; an error in any command does not
terminate execution of subsequent commands. Comments (lines starting
with @samp{#}) are allowed in a command file.
Empty lines are ignored; they do @emph{not}
repeat the last command.
You can't restart the program by having more than one @code{run}
command in the file. Also, the list of commands may include additional
@code{source} commands; however, the @command{gawk} debugger will not source the
same file more than once in order to avoid infinite recursion.

In addition to, or instead of, the @code{source} command, you can use
the @option{-D @var{file}} or @option{--debug=@var{file}} command-line
options to execute commands from a file non-interactively
(@pxref{Options}).
@end table

@node Miscellaneous Debugger Commands
@subsection Miscellaneous Commands

There are a few more commands that do not fit into the
previous categories, as follows:

@table @asis
@cindex debugger commands, @code{dump}
@cindex @code{dump} debugger command
@item @code{dump} [@var{filename}]
Dump byte code of the program to standard output or to the file
named in @var{filename}.  This prints a representation of the internal
instructions that @command{gawk} executes to implement the @command{awk}
commands in a program.  This can be very enlightening, as the following
partial dump of Davide Brini's obfuscated code
(@pxref{Signature Program}) demonstrates:

@c FIXME: This will need updating if num-handler branch is ever merged in.
@smallexample
@group
gawk> @kbd{dump}
@print{}        # BEGIN
@print{}
@print{} [  1:0xfcd340] Op_rule           : [in_rule = BEGIN] [source_file = brini.awk]
@end group
@print{} [  1:0xfcc240] Op_push_i         : "~" [MALLOC|STRING|STRCUR]
@print{} [  1:0xfcc2a0] Op_push_i         : "~" [MALLOC|STRING|STRCUR]
@print{} [  1:0xfcc280] Op_match          :
@print{} [  1:0xfcc1e0] Op_store_var      : O
@print{} [  1:0xfcc2e0] Op_push_i         : "==" [MALLOC|STRING|STRCUR]
@print{} [  1:0xfcc340] Op_push_i         : "==" [MALLOC|STRING|STRCUR]
@print{} [  1:0xfcc320] Op_equal          :
@print{} [  1:0xfcc200] Op_store_var      : o
@print{} [  1:0xfcc380] Op_push           : o
@print{} [  1:0xfcc360] Op_plus_i         : 0 [MALLOC|NUMCUR|NUMBER]
@print{} [  1:0xfcc220] Op_push_lhs       : o [do_reference = true]
@print{} [  1:0xfcc300] Op_assign_plus    :
@print{} [   :0xfcc2c0] Op_pop            :
@print{} [  1:0xfcc400] Op_push           : O
@print{} [  1:0xfcc420] Op_push_i         : "" [MALLOC|STRING|STRCUR]
@print{} [   :0xfcc4a0] Op_no_op          :
@print{} [  1:0xfcc480] Op_push           : O
@print{} [   :0xfcc4c0] Op_concat         : [expr_count = 3] [concat_flag = 0]
@print{} [  1:0xfcc3c0] Op_store_var      : x
@print{} [  1:0xfcc440] Op_push_lhs       : X [do_reference = true]
@print{} [  1:0xfcc3a0] Op_postincrement  :
@print{} [  1:0xfcc4e0] Op_push           : x
@print{} [  1:0xfcc540] Op_push           : o
@print{} [  1:0xfcc500] Op_plus           :
@print{} [  1:0xfcc580] Op_push           : o
@print{} [  1:0xfcc560] Op_plus           :
@print{} [  1:0xfcc460] Op_leq            :
@print{} [   :0xfcc5c0] Op_jmp_false      : [target_jmp = 0xfcc5e0]
@print{} [  1:0xfcc600] Op_push_i         : "%c" [MALLOC|STRING|STRCUR]
@print{} [   :0xfcc660] Op_no_op          :
@print{} [  1:0xfcc520] Op_assign_concat  : c
@print{} [   :0xfcc620] Op_jmp            : [target_jmp = 0xfcc440]
@dots{}
@print{} [     2:0xfcc5a0] Op_K_printf         : [expr_count = 17] [redir_type = ""]
@print{} [      :0xfcc140] Op_no_op            :
@print{} [      :0xfcc1c0] Op_atexit           :
@print{} [      :0xfcc640] Op_stop             :
@print{} [      :0xfcc180] Op_no_op            :
@print{} [      :0xfcd150] Op_after_beginfile  :
@group
@print{} [      :0xfcc160] Op_no_op            :
@print{} [      :0xfcc1a0] Op_after_endfile    :
gawk>
@end group
@end smallexample

@cindex @code{exit} debugger command
@cindex exit the debugger
@item @code{exit}
Exit the debugger.
See the entry for @samp{quit}, later in this list.

@cindex debugger commands, @code{h} (@code{help})
@cindex debugger commands, @code{help}
@cindex @code{help} debugger command
@cindex @code{h} debugger command (alias for @code{help})
@item @code{help}
@itemx @code{h}
Print a list of all of the @command{gawk} debugger commands with a short
summary of their usage.  @samp{help @var{command}} prints the information
about the command @var{command}.

@cindex debugger commands, @code{l} (@code{list})
@cindex debugger commands, @code{list}
@cindex @code{list} debugger command
@cindex @code{l} debugger command (alias for @code{list})
@item @code{list} [@code{-} | @code{+} | @var{n} | @var{filename}@code{:}@var{n} | @var{n}--@var{m} | @var{function}]
@itemx @code{l} [@code{-} | @code{+} | @var{n} | @var{filename}@code{:}@var{n} | @var{n}--@var{m} | @var{function}]
Print the specified lines (default 15) from the current source file
or the file named @var{filename}. The possible arguments to @code{list}
are as follows:

@c nested table
@table @asis
@item @code{-} (Minus)
Print lines before the lines last printed.

@item @code{+}
Print lines after the lines last printed.
@code{list} without any argument does the same thing.

@item @var{n}
Print lines centered around line number @var{n}.

@item  @var{n}--@var{m}
Print lines from @var{n} to @var{m}.

@item @var{filename}@code{:}@var{n}
Print lines centered around line number @var{n} in
source file @var{filename}. This command may change the current source file.

@item @var{function}
Print lines centered around the beginning of the
function @var{function}. This command may change the current source file.
@end table

@cindex debugger commands, @code{q} (@code{quit})
@cindex debugger commands, @code{quit}
@cindex @code{quit} debugger command
@cindex @code{q} debugger command (alias for @code{quit})
@cindex exit the debugger
@item @code{quit}
@itemx @code{q}
Exit the debugger.  Debugging is great fun, but sometimes we all have
to tend to other obligations in life, and sometimes we find the bug
and are free to go on to the next one!  As we saw earlier, if you are
running a program, the debugger warns you when you type
@samp{q} or @samp{quit}, to make sure you really want to quit.

@cindex debugger commands, @code{trace}
@cindex @code{trace} debugger command
@item @code{trace} [@code{on} | @code{off}]
Turn on or off continuous printing of the instructions that are about to
be executed, along with the @command{awk} lines they
implement.  The default is @code{off}.

It is to be hoped that most of the ``opcodes'' in these instructions are
fairly self-explanatory, and using @code{stepi} and @code{nexti} while
@code{trace} is on will make them into familiar friends.

@end table

@node Readline Support
@section Readline Support
@cindex command completion, in debugger
@cindex history expansion, in debugger

If @command{gawk} is compiled with
@uref{http://cnswww.cns.cwru.edu/php/chet/readline/readline.html,
the GNU Readline library}, you can take advantage of that library's
command completion and history expansion features. The following types
of completion are available:

@table @asis
@item Command completion
Command names.

@item Source @value{FN} completion
Source @value{FN}s. Relevant commands are
@code{break},
@code{clear},
@code{list},
@code{tbreak},
and
@code{until}.

@item Argument completion
Non-numeric arguments to a command.
Relevant commands are @code{enable} and @code{info}.

@item Variable name completion
Global variable names, and function arguments in the current context
if the program is running. Relevant commands are
@code{display},
@code{print},
@code{set},
and
@code{watch}.

@end table

@node Limitations
@section Limitations

We hope you find the @command{gawk} debugger useful and enjoyable to work with,
but as with any program, especially in its early releases, it still has
some limitations.  A few that it's worth being aware of are:

@itemize @value{BULLET}
@item
At this point, the debugger does not give a detailed explanation of
what you did wrong when you type in something it doesn't like. Rather, it just
responds @samp{syntax error}.  When you do figure out what your mistake was,
though, you'll feel like a real guru.

@item
@c NOTE: no comma after the ref{} on purpose, due to following
@c parenthetical remark.
If you perused the dump of opcodes in @ref{Miscellaneous Debugger Commands}
(or if you are already familiar with @command{gawk} internals),
you will realize that much of the internal manipulation of data
in @command{gawk}, as in many interpreters, is done on a stack.
@code{Op_push}, @code{Op_pop}, and the like are the ``bread and butter'' of
most @command{gawk} code.

Unfortunately, as of now, the @command{gawk}
debugger does not allow you to examine the stack's contents.
That is, the intermediate results of expression evaluation are on the
stack, but cannot be printed.  Rather, only variables that are defined
in the program can be printed.  Of course, a workaround for
this is to use more explicit variables at the debugging stage and then
change back to obscure, perhaps more optimal code later.

@item
There is no way to look ``inside'' the process of compiling
regular expressions to see if you got it right.  As an @command{awk}
programmer, you are expected to know the meaning of
@code{/[^[:alnum:][:blank:]]/}.

@item
The @command{gawk} debugger is designed to be used by running a program (with all its
parameters) on the command line, as described in @ref{Debugger Invocation}.
There is no way (as of now) to attach or ``break into'' a running program.
This seems reasonable for a language that is used mainly for quickly
executing, short programs.

@item
The @command{gawk} debugger only accepts source code supplied with the @option{-f} option.
@end itemize

@ignore
@c 11/2016: This no longer applies after all the type cleanup work that's been done.
One other point is worth discussing.  Conventional debuggers run in a
separate process (and thus address space) from the programs that they
debug (the @dfn{debuggee}, if you will).

The @command{gawk} debugger is different; it is an integrated part
of @command{gawk} itself.  This makes it possible, in rare cases,
for @command{gawk} to become an excellent demonstrator of Heisenberg
Uncertainty physics, where the mere act of observing something can change
it. Consider the following:@footnote{Thanks to Hermann Peifer for
this example.}

@example
$ @kbd{cat test.awk}
@print{} @{ print typeof($1), typeof($2) @}
$ @kbd{cat test.data}
@print{} abc 123
$ @kbd{gawk -f test.awk test.data}
@print{} strnum strnum
@end example

This is all as expected: field data has the STRNUM attribute
(@pxref{Variable Typing}).  Now watch what happens when we run
this program under the debugger:

@example
$ @kbd{gawk -D -f test.awk test.data}
gawk> @kbd{w $1}                        @ii{Set watchpoint on} $1
@print{} Watchpoint 1: $1
gawk> @kbd{w $2}                        @ii{Set watchpoint on} $2
@print{} Watchpoint 2: $2
gawk> @kbd{r}                           @ii{Start the program}
@print{} Starting program:
@print{} Stopping in Rule ...
@print{} Watchpoint 1: $1               @ii{Watchpoint fires}
@print{}   Old value: ""
@print{}   New value: "abc"
@print{} main() at `test.awk':1
@print{} 1       @{ print typeof($1), typeof($2) @}
gawk> @kbd{n}                           @ii{Keep going @dots{}}
@print{} Watchpoint 2: $2               @ii{Watchpoint fires}
@print{}   Old value: ""
@print{}   New value: "123"
@print{} main() at `test.awk':1
@print{} 1       @{ print typeof($1), typeof($2) @}
gawk> @kbd{n}                           @ii{Get result from} typeof()
@print{} strnum number                  @ii{Result for} $2 @ii{isn't right}
@print{} Program exited normally with exit value: 0
gawk> @kbd{quit}
@end example

In this case, the act of comparing the new value of @code{$2}
with the old one caused @command{gawk} to evaluate it and determine that it
is indeed a number, and this is reflected in the result of
@code{typeof()}.

Cases like this where the debugger is not transparent to the program's
execution should be rare. If you encounter one, please report it
(@pxref{Bugs}).
@end ignore

@ignore
Look forward to a future release when these and other missing features may
be added, and of course feel free to try to add them yourself!
@end ignore

@node Debugging Summary
@section Summary

@itemize @value{BULLET}
@item
Programs rarely work correctly the first time.  Finding bugs
is called debugging, and a program that helps you find bugs is a
debugger.  @command{gawk} has a built-in debugger that works very
similarly to the GNU Debugger, GDB.

@item
Debuggers let you step through your program one statement at a time,
examine and change variable and array values, and do a number of other
things that let you understand what your program is actually doing (as
opposed to what it is supposed to do).

@item
Like most debuggers, the @command{gawk} debugger works in terms of stack
frames, and lets you set both breakpoints (stop at a point in the code)
and watchpoints (stop when a data value changes).

@item
The debugger command set is fairly complete, providing control over
breakpoints, execution, viewing and changing data, working with the stack,
getting information, and other tasks.

@item
If the GNU Readline library is available when @command{gawk} is
compiled, it is used by the debugger to provide command-line history
and editing.

@item
Usually, the debugger does not not affect the
program being debugged, but occasionally it can.

@end itemize

@node Arbitrary Precision Arithmetic
@chapter Arithmetic and Arbitrary-Precision Arithmetic with @command{gawk}
@cindex arbitrary precision
@cindex multiple precision
@cindex infinite precision
@cindex floating-point, numbers@comma{} arbitrary precision

This @value{CHAPTER} introduces some basic concepts relating to
how computers do arithmetic and defines some important terms.
It then proceeds to describe floating-point arithmetic,
which is what @command{awk} uses for all its computations, including a
discussion of arbitrary-precision floating-point arithmetic, which is
a feature available only in @command{gawk}. It continues on to present
arbitrary-precision integers, and concludes with a description of some
points where @command{gawk} and the POSIX standard are not quite in
agreement.

@quotation NOTE
Most users of @command{gawk} can safely skip this chapter.
But if you want to do scientific calculations with @command{gawk},
this is the place to be.
@end quotation

@menu
* Computer Arithmetic::           A quick intro to computer math.
* Math Definitions::              Defining terms used.
* MPFR features::                 The MPFR features in @command{gawk}.
* FP Math Caution::               Things to know.
* Arbitrary Precision Integers::  Arbitrary Precision Integer Arithmetic with
                                  @command{gawk}.
* Checking for MPFR::             How to check if MPFR is available.
* POSIX Floating Point Problems:: Standards Versus Existing Practice.
* Floating point summary::        Summary of floating point discussion.
@end menu

@node Computer Arithmetic
@section A General Description of Computer Arithmetic

Until now, we have worked with data as either numbers or
strings.  Ultimately, however, computers represent everything in terms
of @dfn{binary digits}, or @dfn{bits}.  A decimal digit can take on any
of 10 values: zero through nine.  A binary digit can take on any of two
values, zero or one.  Using binary, computers (and computer software)
can represent and manipulate numerical and character data.  In general,
the more bits you can use to represent a particular thing, the greater
the range of possible values it can take on.

Modern computers support at least two, and often more, ways to do
arithmetic.  Each kind of arithmetic uses a different representation
(organization of the bits) for the numbers.  The kinds of arithmetic
that interest us are:

@table @asis
@item Decimal arithmetic
This is the kind of arithmetic you learned in elementary school, using
paper and pencil (and/or a calculator). In theory, numbers can have an
arbitrary number of digits on either side (or both sides) of the decimal
point, and the results of a computation are always exact.

Some modern systems can do decimal arithmetic in hardware, but usually you
need a special software library to provide access to these instructions.
There are also libraries that do decimal arithmetic entirely in software.

Despite the fact that some users expect @command{gawk} to be performing
decimal arithmetic,@footnote{We don't know why they expect this, but
they do.} it does not do so.

@item Integer arithmetic
In school, integer values were referred to as ``whole'' numbers---that
is, numbers without any fractional part, such as 1, 42, or @minus{}17.
The advantage to integer numbers is that they represent values exactly.
The disadvantage is that their range is limited.

@cindex unsigned integers
@cindex integers, unsigned
In computers, integer values come in two flavors: @dfn{signed} and
@dfn{unsigned}.  Signed values may be negative or positive, whereas
unsigned values are always greater than or equal
to zero.

In computer systems, integer arithmetic is exact, but the possible
range of values is limited.  Integer arithmetic is generally faster than
floating-point arithmetic.

@cindex floating-point, numbers
@item Floating-point arithmetic
Floating-point numbers represent what were called in school ``real''
numbers (i.e., those that have a fractional part, such as 3.1415927).
The advantage to floating-point numbers is that they can represent a
much larger range of values than can integers.  The disadvantage is that
there are numbers that they cannot represent exactly.

Modern systems support floating-point arithmetic in hardware, with a
limited range of values.  There are software libraries that allow
the use of arbitrary-precision floating-point calculations.

@cindex floating-point, numbers@comma{} single-precision
@cindex floating-point, numbers@comma{} double-precision
@cindex floating-point, numbers@comma{} arbitrary-precision
@cindex single-precision
@cindex double-precision
@cindex arbitrary-precision
POSIX @command{awk} uses @dfn{double-precision} floating-point numbers, which
can hold more digits than @dfn{single-precision} floating-point numbers.
@command{gawk} has facilities for performing arbitrary-precision
floating-point arithmetic, which we describe in more detail shortly.
@end table

Computers work with integer and floating-point values of different
ranges. Integer values are usually either 32 or 64 bits in size.
Single-precision floating-point values occupy 32 bits, whereas double-precision
floating-point values occupy 64 bits.
(Quadruple-precision floating point values also exist. They occupy 128 bits,
but such numbers are not available in @command{awk}.)
Floating-point values are always
signed. The possible ranges of values are shown in @ref{table-numeric-ranges}
and @ref{table-floating-point-ranges}.

@float Table,table-numeric-ranges
@caption{Value ranges for integer representations}
@multitable @columnfractions .34 .33 .33
@headitem Representation @tab Minimum value @tab Maximum value
@item 32-bit signed integer @tab @minus{}2,147,483,648 @tab 2,147,483,647
@item 32-bit unsigned integer @tab 0 @tab 4,294,967,295
@item 64-bit signed integer @tab @minus{}9,223,372,036,854,775,808 @tab 9,223,372,036,854,775,807
@item 64-bit unsigned integer @tab 0 @tab 18,446,744,073,709,551,615
@end multitable
@end float

@float Table,table-floating-point-ranges
@caption{Approximate value ranges for floating-point number representations}
@multitable @columnfractions .38 .22 .22 .23
@iftex
@headitem Representation @tab @w{Minimum positive} @w{nonzero value} @tab Minimum @w{finite value} @tab Maximum @w{finite value}
@end iftex
@ifnottex
@headitem Representation @tab Minimum positive nonzero value @tab Minimum finite value @tab Maximum finite value
@end ifnottex
@iftex
@item @w{Single-precision floating-point} @tab @math{1.175494 @cdot 10^{-38}} @tab @math{-3.402823 @cdot 10^{38}} @tab @math{3.402823 @cdot 10^{38}}
@item @w{Double-precision floating-point} @tab @math{2.225074 @cdot 10^{-308}} @tab @math{-1.797693 @cdot 10^{308}} @tab @math{1.797693 @cdot 10^{308}}
@item @w{Quadruple-precision floating-point} @tab @math{3.362103 @cdot 10^{-4932}} @tab @math{-1.189731 @cdot 10^{4932}} @tab @math{1.189731 @cdot 10^{4932}}
@end iftex
@ifinfo
@item Single-precision floating-point @tab 1.175494e-38 @tab -3.402823e+38 @tab 3.402823e+38
@item Double-precision floating-point @tab 2.225074e-308 @tab -1.797693e+308 @tab 1.797693e+308
@item Quadruple-precision floating-point @tab 3.362103e-4932 @tab -1.189731e+4932 @tab 1.189731e+4932
@end ifinfo
@ifnottex
@ifnotinfo
@item Single-precision floating-point @tab 1.175494*10@sup{-38} @tab -3.402823*10@sup{38} @tab 3.402823*10@sup{38}
@item Double-precision floating-point @tab 2.225074*10@sup{-308} @tab -1.797693*10@sup{308} @tab 1.797693*10@sup{308}
@item Quadruple-precision floating-point @tab 3.362103*10@sup{-4932} @tab -1.189731*10@sup{4932} @tab 1.189731*10@sup{4932}
@end ifnotinfo
@end ifnottex
@end multitable
@end float

@node Math Definitions
@section Other Stuff to Know

The rest of this @value{CHAPTER} uses a number of terms. Here are some
informal definitions that should help you work your way through the material
here:

@table @dfn
@item Accuracy
A floating-point calculation's accuracy is how close it comes
to the real (paper and pencil) value.

@item Error
The difference between what the result of a computation ``should be''
and what it actually is.  It is best to minimize error as much
as possible.

@item Exponent
The order of magnitude of a value;
some number of bits in a floating-point value store the exponent.

@item Inf
A special value representing infinity. Operations involving another
number and infinity produce infinity.

@item NaN
``Not a number.''@footnote{Thanks to Michael Brennan for this description,
which we have paraphrased, and for the examples.} A special value that
results from attempting a calculation that has no answer as a real number.
In such a case, programs can either receive a floating-point exception,
or get @code{NaN} back as the result. The IEEE 754 standard recommends
that systems return @code{NaN}. Some examples:

@table @code
@item sqrt(-1)
This makes sense in the range of complex numbers, but not in the
range of real numbers, so the result is @code{NaN}.

@item log(-8)
@minus{}8 is out of the domain of @code{log()}, so the result is @code{NaN}.
@end table

@item Normalized
How the significand (see later in this list) is usually stored. The
value is adjusted so that the first bit is one, and then that leading
one is assumed instead of physically stored.  This provides one
extra bit of precision.

@item Precision
The number of bits used to represent a floating-point number.
The more bits, the more digits you can represent.
Binary and decimal precisions are related approximately, according to the
formula:

@display
@iftex
@math{prec = 3.322 @cdot dps}
@end iftex
@ifnottex
@ifnotdocbook
@var{prec} = 3.322 * @var{dps}
@end ifnotdocbook
@end ifnottex
@docbook
<emphasis>prec</emphasis> = 3.322 &sdot; <emphasis>dps</emphasis>
@end docbook
@end display

@noindent
Here, @emph{prec} denotes the binary precision
(measured in bits) and @emph{dps} (short for decimal places)
is the decimal digits.

@item Rounding mode
How numbers are rounded up or down when necessary.
More details are provided later.

@item Significand
A floating-point value consists of the significand multiplied by 10
to the power of the exponent. For example, in @code{1.2345e67},
the significand is @code{1.2345}.

@item Stability
From @uref{https://en.wikipedia.org/wiki/Numerical_stability,
the Wikipedia article on numerical stability}:
``Calculations that can be proven not to magnify approximation errors
are called @dfn{numerically stable}.''
@end table

See @uref{https://en.wikipedia.org/wiki/Accuracy_and_precision,
the Wikipedia article on accuracy and precision} for more information
on some of those terms.

On modern systems, floating-point hardware uses the representation and
operations defined by the IEEE 754 standard.
Three of the standard IEEE 754 types are 32-bit single precision,
64-bit double precision, and 128-bit quadruple precision.
The standard also specifies extended precision formats
to allow greater precisions and larger exponent ranges.
(@command{awk} uses only the 64-bit double-precision format.)

@ref{table-ieee-formats} lists the precision and exponent
field values for the basic IEEE 754 binary formats.

@float Table,table-ieee-formats
@caption{Basic IEEE format values}
@multitable @columnfractions .20 .20 .20 .20 .20
@headitem Name @tab Total bits @tab Precision @tab Minimum exponent @tab Maximum exponent
@item Single @tab 32 @tab 24 @tab @minus{}126 @tab +127
@item Double @tab 64 @tab 53 @tab @minus{}1022 @tab +1023
@item Quadruple @tab 128 @tab 113 @tab @minus{}16382 @tab +16383
@end multitable
@end float

@quotation NOTE
The precision numbers include the implied leading one that gives them
one extra bit of significand.
@end quotation

@node MPFR features
@section Arbitrary-Precision Arithmetic Features in @command{gawk}

By default, @command{gawk} uses the double-precision floating-point values
supplied by the hardware of the system it runs on.  However, if it was
compiled to do so, and the @option{-M} command-line option is supplied,
@command{gawk} uses the @uref{http://www.mpfr.org,
GNU MPFR} and @uref{https://gmplib.org, GNU MP} (GMP) libraries for
arbitrary-precision arithmetic on numbers.  You can see if MPFR support
is available like so:

@example
$ @kbd{gawk --version}
@print{} GNU Awk 4.1.2, API: 1.1 (GNU MPFR 3.1.0-p3, GNU MP 5.0.2)
@print{} Copyright (C) 1989, 1991-2015 Free Software Foundation.
@dots{}
@end example

@noindent
(You may see different version numbers than what's shown here. That's OK;
what's important is to see that GNU MPFR and GNU MP are listed in
the output.)

Additionally, there are a few elements available in the @code{PROCINFO}
array to provide information about the MPFR and GMP libraries
(@pxref{Auto-set}).

The MPFR library provides precise control over precisions and rounding
modes, and gives correctly rounded, reproducible, platform-independent
results.  With the @option{-M} command-line option,
all floating-point arithmetic operators and numeric functions
can yield results to any desired precision level supported by MPFR.

Two predefined variables, @code{PREC} and @code{ROUNDMODE},
provide control over the working precision and the rounding mode.
The precision and the rounding mode are set globally for every operation
to follow.
@xref{Setting precision} and @ref{Setting the rounding mode}
for more information.

@node FP Math Caution
@section Floating-Point Arithmetic: Caveat Emptor!

@quotation
@i{Math class is tough!}
@author Teen Talk Barbie, July 1992
@end quotation

This @value{SECTION} provides a high-level overview of the issues
involved when doing lots of floating-point arithmetic.@footnote{There
is a very nice @uref{http://www.validlab.com/goldberg/paper.pdf,
paper on floating-point arithmetic} by David Goldberg, ``What Every
Computer Scientist Should Know About Floating-Point Arithmetic,''
@cite{ACM Computing Surveys} @strong{23}, 1 (1991-03): 5-48.  This is
worth reading if you are interested in the details, but it does require
a background in computer science.}
The discussion applies to both hardware and arbitrary-precision
floating-point arithmetic.

@quotation CAUTION
The material here is purposely general. If you need to do serious
computer arithmetic, you should do some research first, and not
rely just on what we tell you.
@end quotation

@menu
* Inexactness of computations:: Floating point math is not exact.
* Getting Accuracy::            Getting more accuracy takes some work.
* Try To Round::                Add digits and round.
* Setting precision::           How to set the precision.
* Setting the rounding mode::   How to set the rounding mode.
@end menu

@node Inexactness of computations
@subsection Floating-Point Arithmetic Is Not Exact

Binary floating-point representations and arithmetic are inexact.
Simple values like 0.1 cannot be precisely represented using
binary floating-point numbers, and the limited precision of
floating-point numbers means that slight changes in
the order of operations or the precision of intermediate storage
can change the result. To make matters worse, with arbitrary-precision
floating-point arithmetic, you can set the precision before starting a
computation, but then you cannot be sure of the number of significant
decimal places in the final result.

@menu
* Inexact representation::      Numbers are not exactly represented.
* Comparing FP Values::         How to compare floating point values.
* Errors accumulate::           Errors get bigger as they go.
@end menu

@node Inexact representation
@subsubsection Many Numbers Cannot Be Represented Exactly

So, before you start to write any code, you should think
about what you really want and what's really happening. Consider the
two numbers in the following example:

@example
x = 0.875             # 1/2 + 1/4 + 1/8
y = 0.425
@end example

Unlike the number in @code{y}, the number stored in @code{x}
is exactly representable
in binary because it can be written as a finite sum of one or
more fractions whose denominators are all powers of two.
When @command{gawk} reads a floating-point number from
program source, it automatically rounds that number to whatever
precision your machine supports. If you try to print the numeric
content of a variable using an output format string of @code{"%.17g"},
it may not produce the same number as you assigned to it:

@example
$ @kbd{gawk 'BEGIN @{ x = 0.875; y = 0.425}
> @kbd{              printf("%0.17g, %0.17g\n", x, y) @}'}
@print{} 0.875, 0.42499999999999999
@end example

Often the error is so small you do not even notice it, and if you do,
you can always specify how much precision you would like in your output.
Usually this is a format string like @code{"%.15g"}, which, when
used in the previous example, produces an output identical to the input.

@node Comparing FP Values
@subsubsection Be Careful Comparing Values

Because the underlying representation can be a little bit off from the exact value,
comparing floating-point values to see if they are exactly equal is generally a bad idea.
Here is an example where it does not work like you would expect:

@example
$ @kbd{gawk 'BEGIN @{ print (0.1 + 12.2 == 12.3) @}'}
@print{} 0
@end example

The general wisdom when comparing floating-point values is to see if
they are within some small range of each other (called a @dfn{delta},
or @dfn{tolerance}).
You have to decide how small a delta is important to you. Code to do
this looks something like the following:

@example
@group
delta = 0.00001                 # for example
difference = abs(a) - abs(b)    # subtract the two values
if (difference < delta)
    # all ok
else
    # not ok
@end group
@end example

@noindent
(We assume that you have a simple absolute value function named
@code{abs()} defined elsewhere in your program.)

@node Errors accumulate
@subsubsection Errors Accumulate

The loss of accuracy during a single computation with floating-point
numbers usually isn't enough to worry about. However, if you compute a
value that is the result of a sequence of floating-point operations,
the error can accumulate and greatly affect the computation itself.
Here is an attempt to compute the value of @value{PI} using one of its
many series representations:

@example
BEGIN @{
    x = 1.0 / sqrt(3.0)
    n = 6
    for (i = 1; i < 30; i++) @{
        n = n * 2.0
        x = (sqrt(x * x + 1) - 1) / x
        printf("%.15f\n", n * x)
    @}
@}
@end example

When run, the early errors propagate through later computations,
causing the loop to terminate prematurely after attempting to divide by zero:

@example
$ @kbd{gawk -f pi.awk}
@print{} 3.215390309173475
@print{} 3.159659942097510
@print{} 3.146086215131467
@print{} 3.142714599645573
@dots{}
@print{} 3.224515243534819
@print{} 2.791117213058638
@print{} 0.000000000000000
@error{} gawk: pi.awk:6: fatal: division by zero attempted
@end example

Here is an additional example where the inaccuracies in internal representations
yield an unexpected result:

@example
$ @kbd{gawk 'BEGIN @{}
>   @kbd{for (d = 1.1; d <= 1.5; d += 0.1)    # loop five times (?)}
>       @kbd{i++}
>   @kbd{print i}
> @kbd{@}'}
@print{} 4
@end example

@node Getting Accuracy
@subsection Getting the Accuracy You Need

Can arbitrary-precision arithmetic give exact results? There are
no easy answers. The standard rules of algebra often do not apply
when using floating-point arithmetic.
Among other things, the distributive and associative laws
do not hold completely, and order of operation may be important
for your computation. Rounding error, cumulative precision loss,
and underflow are often troublesome.

When @command{gawk} tests the expressions @samp{0.1 + 12.2} and
@samp{12.3} for equality using the machine double-precision arithmetic,
it decides that they are not equal!  (@xref{Comparing FP Values}.)
You can get the result you want by increasing the precision; 56 bits in
this case does the job:

@example
$ @kbd{gawk -M -v PREC=56 'BEGIN @{ print (0.1 + 12.2 == 12.3) @}'}
@print{} 1
@end example

If adding more bits is good, perhaps adding even more bits of
precision is better?
Here is what happens if we use an even larger value of @code{PREC}:

@example
$ @kbd{gawk -M -v PREC=201 'BEGIN @{ print (0.1 + 12.2 == 12.3) @}'}
@print{} 0
@end example

This is not a bug in @command{gawk} or in the MPFR library.
It is easy to forget that the finite number of bits used to store the value
is often just an approximation after proper rounding.
The test for equality succeeds if and only if @emph{all} bits in the two operands
are exactly the same. Because this is not necessarily true after floating-point
computations with a particular precision and effective rounding mode,
a straight test for equality may not work. Instead, compare the
two numbers to see if they are within the desirable delta of each other.

In applications where 15 or fewer decimal places suffice,
hardware double-precision arithmetic can be adequate, and is usually much faster.
But you need to keep in mind that every floating-point operation
can suffer a new rounding error with catastrophic consequences, as illustrated
by our earlier attempt to compute the value of @value{PI}.
Extra precision can greatly enhance the stability and the accuracy
of your computation in such cases.

Additionally, you should understand that
repeated addition is not necessarily equivalent to multiplication
in floating-point arithmetic. In the example in
@ref{Errors accumulate}:

@example
$ @kbd{gawk 'BEGIN @{}
>   @kbd{for (d = 1.1; d <= 1.5; d += 0.1)    # loop five times (?)}
>       @kbd{i++}
>   @kbd{print i}
> @kbd{@}'}
@print{} 4
@end example

@noindent
you may or may not succeed in getting the correct result by choosing
an arbitrarily large value for @code{PREC}. Reformulation of
the problem at hand is often the correct approach in such situations.

@node Try To Round
@subsection Try a Few Extra Bits of Precision and Rounding

Instead of arbitrary-precision floating-point arithmetic,
often all you need is an adjustment of your logic
or a different order for the operations in your calculation.
The stability and the accuracy of the computation of @value{PI}
in the earlier example can be enhanced by using the following
simple algebraic transformation:

@example
(sqrt(x * x + 1) - 1) / x @equiv{} x / (sqrt(x * x + 1) + 1)
@end example

@noindent
After making this change, the program converges to
@value{PI} in under 30 iterations:

@example
$ @kbd{gawk -f pi2.awk}
@print{} 3.215390309173473
@print{} 3.159659942097501
@print{} 3.146086215131436
@print{} 3.142714599645370
@print{} 3.141873049979825
@dots{}
@print{} 3.141592653589797
@print{} 3.141592653589797
@end example

@node Setting precision
@subsection Setting the Precision

@command{gawk} uses a global working precision; it does not keep track of
the precision or accuracy of individual numbers. Performing an arithmetic
operation or calling a built-in function rounds the result to the current
working precision. The default working precision is 53 bits, which you can
modify using the predefined variable @code{PREC}. You can also set the
value to one of the predefined case-insensitive strings
shown in @ref{table-predefined-precision-strings},
to emulate an IEEE 754 binary format.

@float Table,table-predefined-precision-strings
@caption{Predefined precision strings for @code{PREC}}
@multitable {@code{"double"}} {12345678901234567890123456789012345}
@headitem @code{PREC} @tab IEEE 754 binary format
@item @code{"half"} @tab 16-bit half-precision
@item @code{"single"} @tab Basic 32-bit single precision
@item @code{"double"} @tab Basic 64-bit double precision
@item @code{"quad"} @tab Basic 128-bit quadruple precision
@item @code{"oct"} @tab 256-bit octuple precision
@end multitable
@end float

The following example illustrates the effects of changing precision
on arithmetic operations:

@example
$ @kbd{gawk -M -v PREC=100 'BEGIN @{ x = 1.0e-400; print x + 0}
>   @kbd{PREC = "double"; print x + 0 @}'}
@print{} 1e-400
@print{} 0
@end example

@quotation CAUTION
Be wary of floating-point constants! When reading a floating-point
constant from program source code, @command{gawk} uses the default
precision (that of a C @code{double}), unless overridden by an assignment
to the special variable @code{PREC} on the command line, to store it
internally as an MPFR number.  Changing the precision using @code{PREC}
in the program text does @emph{not} change the precision of a constant.

If you need to represent a floating-point constant at a higher precision
than the default and cannot use a command-line assignment to @code{PREC},
you should either specify the constant as a string, or as a rational
number, whenever possible. The following example illustrates the
differences among various ways to print a floating-point constant:

@example
$ @kbd{gawk -M 'BEGIN @{ PREC = 113; printf("%0.25f\n", 0.1) @}'}
@print{} 0.1000000000000000055511151
$ @kbd{gawk -M -v PREC=113 'BEGIN @{ printf("%0.25f\n", 0.1) @}'}
@print{} 0.1000000000000000000000000
$ @kbd{gawk -M 'BEGIN @{ PREC = 113; printf("%0.25f\n", "0.1") @}'}
@print{} 0.1000000000000000000000000
$ @kbd{gawk -M 'BEGIN @{ PREC = 113; printf("%0.25f\n", 1/10) @}'}
@print{} 0.1000000000000000000000000
@end example
@end quotation

@node Setting the rounding mode
@subsection Setting the Rounding Mode

@cindex @code{ROUNDMODE} variable
The @code{ROUNDMODE} variable provides
program-level control over the rounding mode.
The correspondence between @code{ROUNDMODE} and the IEEE
rounding modes is shown in @ref{table-gawk-rounding-modes}.

@float Table,table-gawk-rounding-modes
@caption{@command{gawk} rounding modes}
@multitable @columnfractions .45 .30 .25
@headitem Rounding mode @tab IEEE name @tab @code{ROUNDMODE}
@item Round to nearest, ties to even @tab @code{roundTiesToEven} @tab @code{"N"} or @code{"n"}
@item Round toward positive infinity @tab @code{roundTowardPositive} @tab @code{"U"} or @code{"u"}
@item Round toward negative infinity @tab @code{roundTowardNegative} @tab @code{"D"} or @code{"d"}
@item Round toward zero @tab @code{roundTowardZero} @tab @code{"Z"} or @code{"z"}
@item Round away from zero @tab @tab @code{"A"} or @code{"a"}
@end multitable
@end float

@code{ROUNDMODE} has the default value @code{"N"}, which
selects the IEEE 754 rounding mode @code{roundTiesToEven}.
In @ref{table-gawk-rounding-modes}, the value @code{"A"} selects
rounding away from zero. This is only available if your version of the
MPFR library supports it; otherwise, setting @code{ROUNDMODE} to @code{"A"}
has no effect.

The default mode @code{roundTiesToEven} is the most preferred,
but the least intuitive. This method does the obvious thing for most values,
by rounding them up or down to the nearest digit.
For example, rounding 1.132 to two digits yields 1.13,
and rounding 1.157 yields 1.16.

However, when it comes to rounding a value that is exactly halfway between,
things do not work the way you probably learned in school.
In this case, the number is rounded to the nearest even digit.
So rounding 0.125 to two digits rounds down to 0.12,
but rounding 0.6875 to three digits rounds up to 0.688.
You probably have already encountered this rounding mode when
using @code{printf} to format floating-point numbers.
For example:

@example
BEGIN @{
    x = -4.5
    for (i = 1; i < 10; i++) @{
        x += 1.0
        printf("%4.1f => %2.0f\n", x, x)
    @}
@}
@end example

@noindent
produces the following output when run on the author's system:@footnote{It
is possible for the output to be completely different if the
C library in your system does not use the IEEE 754 even-rounding
rule to round halfway cases for @code{printf}.}

@example
-3.5 => -4
-2.5 => -2
-1.5 => -2
-0.5 => 0
 0.5 => 0
 1.5 => 2
 2.5 => 2
 3.5 => 4
 4.5 => 4
@end example

The theory behind @code{roundTiesToEven} is that it more or less evenly
distributes upward and downward rounds of exact halves, which might
cause any accumulating round-off error to cancel itself out. This is the
default rounding mode for IEEE 754 computing functions and operators.

@c January 2018. Thanks to nethox@gmail.com for the example.
@sidebar Rounding Modes and Conversion
It's important to understand that, along with @code{CONVFMT} and
@code{OFMT}, the rounding mode affects how numbers are converted to strings.
For example, consider the following program:

@example
BEGIN @{
    pi = 3.1416
    OFMT = "%.f"        # Print value as integer
    print pi            # ROUNDMODE = "N" by default.
    ROUNDMODE = "U"     # Now change ROUNDMODE
    print pi
@}
@end example

@noindent
Running this program produces this output:

@example
$ @kbd{gawk -M -f roundmode.awk}
@print{} 3
@print{} 4
@end example
@end sidebar

The other rounding modes are rarely used.  Rounding toward positive infinity
(@code{roundTowardPositive}) and toward negative infinity
(@code{roundTowardNegative}) are often used to implement interval
arithmetic, where you adjust the rounding mode to calculate upper and
lower bounds for the range of output. The @code{roundTowardZero} mode can
be used for converting floating-point numbers to integers.  When rounding
away from zero, the nearest number with magnitude greater than or equal to
the value is selected.

Some numerical analysts will tell you that your choice of rounding
style has tremendous impact on the final outcome, and advise you to
wait until final output for any rounding. Instead, you can often avoid
round-off error problems by setting the precision initially to some
value sufficiently larger than the final desired precision, so that
the accumulation of round-off error does not influence the outcome.
If you suspect that results from your computation are sensitive to
accumulation of round-off error, look for a significant difference in
output when you change the rounding mode to be sure.

@node Arbitrary Precision Integers
@section Arbitrary-Precision Integer Arithmetic with @command{gawk}
@cindex integers, arbitrary precision
@cindex arbitrary precision integers

When given the @option{-M} option,
@command{gawk} performs all integer arithmetic using GMP arbitrary-precision
integers.  Any number that looks like an integer in a source
or @value{DF} is stored as an arbitrary-precision integer.  The size
of the integer is limited only by the available memory.  For example,
the following computes
@iftex
@math{5^{4^{3^{2}}}},
@end iftex
@ifinfo
5^4^3^2,
@end ifinfo
@ifnottex
@ifnotinfo
5@sup{4@sup{3@sup{2}}},
@end ifnotinfo
@end ifnottex
the result of which is beyond the
limits of ordinary hardware double-precision floating-point values:

@example
$ @kbd{gawk -M 'BEGIN @{}
>   @kbd{x = 5^4^3^2}
>   @kbd{print "number of digits =", length(x)}
>   @kbd{print substr(x, 1, 20), "...", substr(x, length(x) - 19, 20)}
> @kbd{@}'}
@print{} number of digits = 183231
@print{} 62060698786608744707 ... 92256259918212890625
@end example

If instead you were to compute the same value using arbitrary-precision
floating-point values, the precision needed for correct output (using
the formula
@iftex
@math{prec = 3.322 @cdot dps})
would be @math{3.322 @cdot 183231},
@end iftex
@ifnottex
@ifnotdocbook
@samp{prec = 3.322 * dps})
would be 3.322 x 183231,
@end ifnotdocbook
@end ifnottex
@docbook
<emphasis>prec</emphasis> = 3.322 &sdot; <emphasis>dps</emphasis>)
would be
<emphasis>prec</emphasis> = 3.322 &sdot; 183231,
@end docbook
or 608693.

The result from an arithmetic operation with an integer and a floating-point value
is a floating-point value with a precision equal to the working precision.
The following program calculates the eighth term in
Sylvester's sequence@footnote{Weisstein, Eric W.
@cite{Sylvester's Sequence}. From MathWorld---A Wolfram Web Resource
@w{(@url{http://mathworld.wolfram.com/SylvestersSequence.html}).}}
using a recurrence:

@example
$ @kbd{gawk -M 'BEGIN @{}
>   @kbd{s = 2.0}
>   @kbd{for (i = 1; i <= 7; i++)}
>       @kbd{s = s * (s - 1) + 1}
>   @kbd{print s}
> @kbd{@}'}
@print{} 113423713055421845118910464
@end example

The output differs from the actual number, 113,423,713,055,421,844,361,000,443,
because the default precision of 53 bits is not enough to represent the
floating-point results exactly. You can either increase the precision
(100 bits is enough in this case), or replace the floating-point constant
@samp{2.0} with an integer, to perform all computations using integer
arithmetic to get the correct output.

Sometimes @command{gawk} must implicitly convert an arbitrary-precision
integer into an arbitrary-precision floating-point value.  This is
primarily because the MPFR library does not always provide the relevant
interface to process arbitrary-precision integers or mixed-mode numbers
as needed by an operation or function.  In such a case, the precision is
set to the minimum value necessary for exact conversion, and the working
precision is not used for this purpose.  If this is not what you need or
want, you can employ a subterfuge and convert the integer to floating
point first, like this:

@example
gawk -M 'BEGIN @{ n = 13; print (n + 0.0) % 2.0 @}'
@end example

You can avoid this issue altogether by specifying the number as a floating-point value
to begin with:

@example
gawk -M 'BEGIN @{ n = 13.0; print n % 2.0 @}'
@end example

Note that for this particular example, it is likely best
to just use the following:

@example
gawk -M 'BEGIN @{ n = 13; print n % 2 @}'
@end example

When dividing two arbitrary precision integers with either
@samp{/} or @samp{%}, the result is typically an arbitrary
precision floating point value (unless the denominator evenly
divides into the numerator).
@ifset INTDIV
In order to do integer division
or remainder with arbitrary precision integers, use the built-in
@code{intdiv0()} function (@pxref{Numeric Functions}).

You can simulate the @code{intdiv0()} function in standard @command{awk}
using this user-defined function:

@example
@c file eg/lib/intdiv0.awk
# intdiv0 --- do integer division

@c endfile
@ignore
@c file eg/lib/intdiv0.awk
#
# Arnold Robbins, arnold@@skeeve.com, Public Domain
# July, 2014
#
# Name changed from div() to intdiv()
# April, 2015
#
# Changed to intdiv0()
# April, 2016

@c endfile

@end ignore
@c file eg/lib/intdiv0.awk
function intdiv0(numerator, denominator, result)
@{
    split("", result)

    numerator = int(numerator)
    denominator = int(denominator)
    result["quotient"] = int(numerator / denominator)
    result["remainder"] = int(numerator % denominator)

    return 0.0
@}
@c endfile
@end example

The following example program, contributed by Katie Wasserman,
uses @code{intdiv0()} to
compute the digits of @value{PI} to as many places as you
choose to set:

@example
@c file eg/prog/pi.awk
@group
# pi.awk --- compute the digits of pi
@c endfile
@c endfile
@ignore
@c file eg/prog/pi.awk
#
# Katie Wasserman, katie@@wass.net
# August 2014
@c endfile
@end ignore
@c file eg/prog/pi.awk

BEGIN @{
    digits = 100000
    two = 2 * 10 ^ digits
@end group
    pi = two
    for (m = digits * 4; m > 0; --m) @{
        d = m * 2 + 1
        x = pi * m
        intdiv0(x, d, result)
        pi = result["quotient"]
        pi = pi + two
    @}
    print pi
@}
@c endfile
@end example

@ignore
Date: Wed, 20 Aug 2014 10:19:11 -0400
To: arnold@skeeve.com
From: Katherine Wasserman <katie@wass.net>
Subject: Re: computation of digits of pi?

Arnold,

>The program that you sent to compute the digits of pi using div(). Is
>that some standard algorithm that every math student knows? If so,
>what's it called?

It's not that well known but it's not that obscure either

It's Euler's modification to Newton's method for calculating pi.

Take a look at lines (23) - (25)  here: http://mathworld.wolfram.com/PiFormulas.htm

The algorithm I wrote simply expands the multiply by 2 and works from the innermost expression outwards.  I used this to program HP calculators because it's quite easy to modify for tiny memory devices with smallish word sizes.   

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/articles.cgi?read=899

-Katie
@end ignore

When asked about the algorithm used, Katie replied:

@quotation
It's not that well known but it's not that obscure either.
It's Euler's modification to Newton's method for calculating pi.
Take a look at lines (23) - (25) here: @uref{http://mathworld.wolfram.com/PiFormulas.html}.

The algorithm I wrote simply expands the multiply by 2 and works from
the innermost expression outwards.  I used this to program HP calculators
because it's quite easy to modify for tiny memory devices with smallish
word sizes. See
@uref{http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/articles.cgi?read=899}.
@end quotation
@end ifset

@node Checking for MPFR
@section How To Check If MPFR Is Available

@cindex MPFR, checking availability of
@cindex checking for MPFR
Occasionally, you might like to be able to check if @command{gawk}
was invoked with the @option{-M} option, enabling arbitrary-precision
arithmetic.  You can do so with the following function, contributed
by Andrew Schorr:

@example
@c file eg/lib/have_mpfr.awk
# adequate_math_precision --- return true if we have enough bits
@c endfile
@ignore
@c file eg/lib/have_mpfr.awk
#
# Andrew Schorr, aschorr@@telemetry-investments.com, Public Domain
# May 2017
@c endfile
@end ignore
@c file eg/lib/have_mpfr.awk

function adequate_math_precision(n)
@{
    return (1 != (1+(1/(2^(n-1)))))
@}
@c endfile
@end example

Here is code that invokes the function in order to check
if arbitrary-precision arithmetic is available:

@example
BEGIN @{
    # How many bits of mantissa precision are required
    # for this program to function properly?
    fpbits = 123

    # We hope that we were invoked with MPFR enabled. If so, the
    # following statement should configure calculations to our desired
    # precision.
    PREC = fpbits

    if (! adequate_math_precision(fpbits)) @{
        print("Error: insufficient computation precision available.\n" \
              "Try again with the -M argument?") > "/dev/stderr"
        # Note: you may need to set a flag here to bail out of END rules
        exit 1
    @}
@}
@end example

Please be aware that @code{exit} will jump to the @code{END} rules, if present (@pxref{Exit Statement}).

@node POSIX Floating Point Problems
@section Standards Versus Existing Practice

Historically, @command{awk} has converted any nonnumeric-looking string
to the numeric value zero, when required.  Furthermore, the original
definition of the language and the original POSIX standards specified that
@command{awk} only understands decimal numbers (base 10), and not octal
(base 8) or hexadecimal numbers (base 16).

Changes in the language of the
2001 and 2004 POSIX standards can be interpreted to imply that @command{awk}
should support additional features.  These features are:

@itemize @value{BULLET}
@item
Interpretation of floating-point data values specified in hexadecimal
notation (e.g., @code{0xDEADBEEF}). (Note: data values, @emph{not}
source code constants.)

@item
Support for the special IEEE 754 floating-point values ``not a number''
(NaN), positive infinity (``inf''), and negative infinity (``@minus{}inf'').
In particular, the format for these values is as specified by the ISO 1999
C standard, which ignores case and can allow implementation-dependent additional
characters after the @samp{nan} and allow either @samp{inf} or @samp{infinity}.
@end itemize

The first problem is that both of these are clear changes to historical
practice:

@itemize @value{BULLET}
@item
The @command{gawk} maintainer feels that supporting hexadecimal
floating-point values, in particular, is ugly, and was never intended by the
original designers to be part of the language.

@item
Allowing completely alphabetic strings to have valid numeric
values is also a very severe departure from historical practice.
@end itemize

The second problem is that the @command{gawk} maintainer feels that this
interpretation of the standard, which required a certain amount of
``language lawyering'' to arrive at in the first place, was not even
intended by the standard developers.  In other words, ``We see how you
got where you are, but we don't think that that's where you want to be.''

Recognizing these issues, but attempting to provide compatibility
with the earlier versions of the standard,
the 2008 POSIX standard added explicit wording to allow, but not require,
that @command{awk} support hexadecimal floating-point values and
special values for ``not a number'' and infinity.

Although the @command{gawk} maintainer continues to feel that
providing those features is inadvisable,
nevertheless, on systems that support IEEE floating point, it seems
reasonable to provide @emph{some} way to support NaN and infinity values.
The solution implemented in @command{gawk} is as follows:

@itemize @value{BULLET}
@item
With the @option{--posix} command-line option, @command{gawk} becomes
``hands off.'' String values are passed directly to the system library's
@code{strtod()} function, and if it successfully returns a numeric value,
that is what's used.@footnote{You asked for it, you got it.}
By definition, the results are not portable across
different systems.  They are also a little surprising:

@example
$ @kbd{echo nanny | gawk --posix '@{ print $1 + 0 @}'}
@print{} nan
$ @kbd{echo 0xDeadBeef | gawk --posix '@{ print $1 + 0 @}'}
@print{} 3735928559
@end example

@item
Without @option{--posix}, @command{gawk} interprets the four string values
@samp{+inf},
@samp{-inf},
@samp{+nan},
and
@samp{-nan}
specially, producing the corresponding special numeric values.
The leading sign acts a signal to @command{gawk} (and the user)
that the value is really numeric.  Hexadecimal floating point is
not supported (unless you also use @option{--non-decimal-data},
which is @emph{not} recommended). For example:

@example
$ @kbd{echo nanny | gawk '@{ print $1 + 0 @}'}
@print{} 0
$ @kbd{echo +nan | gawk '@{ print $1 + 0 @}'}
@print{} nan
$ @kbd{echo 0xDeadBeef | gawk '@{ print $1 + 0 @}'}
@print{} 0
@end example

@command{gawk} ignores case in the four special values.
Thus, @samp{+nan} and @samp{+NaN} are the same.
@end itemize

@node Floating point summary
@section Summary

@itemize @value{BULLET}
@item
Most computer arithmetic is done using either integers or floating-point
values.  Standard @command{awk} uses double-precision
floating-point values.

@item
In the early 1990s Barbie mistakenly said, ``Math class is tough!''
Although math isn't tough, floating-point arithmetic isn't the same
as pencil-and-paper math, and care must be taken:

@c nested list
@itemize @value{MINUS}
@item
Not all numbers can be represented exactly.

@item
Comparing values should use a delta, instead of being done directly
with @samp{==} and @samp{!=}.

@item
Errors accumulate.

@item
Operations are not always truly associative or distributive.
@end itemize

@item
Increasing the accuracy can help, but it is not a panacea.

@item
Often, increasing the accuracy and then rounding to the desired
number of digits produces reasonable results.

@item
Use @option{-M} (or @option{--bignum}) to enable MPFR
arithmetic. Use @code{PREC} to set the precision in bits, and
@code{ROUNDMODE} to set the IEEE 754 rounding mode.

@item
With @option{-M}, @command{gawk} performs
arbitrary-precision integer arithmetic using the GMP library.
This is faster and more space-efficient than using MPFR for
the same calculations.

@item
There are several areas with respect to floating-point
numbers where @command{gawk} disagrees with the POSIX standard.
It pays to be aware of them.

@item
Overall, there is no need to be unduly suspicious about the results from
floating-point arithmetic. The lesson to remember is that floating-point
arithmetic is always more complex than arithmetic using pencil and
paper. In order to take advantage of the power of floating-point arithmetic,
you need to know its limitations and work within them. For most casual
use of floating-point arithmetic, you will often get the expected result
if you simply round the display of your final results to the correct number
of significant decimal digits.

@item
As general advice, avoid presenting numerical data in a manner that
implies better precision than is actually the case.

@end itemize

@node Dynamic Extensions
@chapter Writing Extensions for @command{gawk}
@cindex dynamically loaded extensions

It is possible to add new functions written in C or C++ to @command{gawk} using
dynamically loaded libraries. This facility is available on systems
that support the C @code{dlopen()} and @code{dlsym()}
functions.  This @value{CHAPTER} describes how to create extensions
using code written in C or C++.

If you don't know anything about C programming, you can safely skip this
@value{CHAPTER}, although you may wish to review the documentation on the
extensions that come with @command{gawk} (@pxref{Extension Samples}),
and the information on the @code{gawkextlib} project (@pxref{gawkextlib}).
The sample extensions are automatically built and installed when
@command{gawk} is.

@quotation NOTE
When @option{--sandbox} is specified, extensions are disabled
(@pxref{Options}).
@end quotation

@menu
* Extension Intro::             What is an extension.
* Plugin License::              A note about licensing.
* Extension Mechanism Outline:: An outline of how it works.
* Extension API Description::   A full description of the API.
* Finding Extensions::          How @command{gawk} finds compiled extensions.
* Extension Example::           Example C code for an extension.
* Extension Samples::           The sample extensions that ship with
                                @command{gawk}.
* gawkextlib::                  The @code{gawkextlib} project.
* Extension summary::           Extension summary.
* Extension Exercises::         Exercises.
@end menu

@node Extension Intro
@section Introduction

@cindex plug-in
An @dfn{extension} (sometimes called a @dfn{plug-in}) is a piece of
external compiled code that @command{gawk} can load at runtime to
provide additional functionality, over and above the built-in capabilities
described in the rest of this @value{DOCUMENT}.

Extensions are useful because they allow you (of course) to extend
@command{gawk}'s functionality. For example, they can provide access to
system calls (such as @code{chdir()} to change directory) and to other
C library routines that could be of use.  As with most software,
``the sky is the limit''; if you can imagine something that you might
want to do and can write in C or C++, you can write an extension to do it!

Extensions are written in C or C++, using the @dfn{application programming
interface} (API) defined for this purpose by the @command{gawk}
developers.  The rest of this @value{CHAPTER} explains
the facilities that the API provides and how to use
them, and presents a small example extension.  In addition, it documents
the sample extensions included in the @command{gawk} distribution
and describes the @code{gawkextlib} project.
@ifclear FOR_PRINT
@xref{Extension Design}, for a discussion of the extension mechanism
goals and design.
@end ifclear
@ifset FOR_PRINT
See @uref{https://www.gnu.org/software/gawk/manual/html_node/Extension-Design.html}
for a discussion of the extension mechanism
goals and design.
@end ifset

@node Plugin License
@section Extension Licensing

Every dynamic extension must be distributed under a license that is
compatible with the GNU GPL (@pxref{Copying}).

In order for the extension to tell @command{gawk} that it is
properly licensed, the extension must define the global symbol
@code{plugin_is_GPL_compatible}.  If this symbol does not exist,
@command{gawk} emits a fatal error and exits when it tries to load
your extension.

The declared type of the symbol should be @code{int}.  It does not need
to be in any allocated section, though.  The code merely asserts that
the symbol exists in the global scope.  Something like this is enough:

@example
int plugin_is_GPL_compatible;
@end example

@node Extension Mechanism Outline
@section How It Works at a High Level

Communication between
@command{gawk} and an extension is two-way.  First, when an extension
is loaded, @command{gawk} passes it a pointer to a @code{struct} whose fields are
function pointers.
@ifnotdocbook
This is shown in @ref{figure-load-extension}.
@end ifnotdocbook
@ifdocbook
This is shown in @inlineraw{docbook, <xref linkend="figure-load-extension"/>}.
@end ifdocbook

@ifnotdocbook
@float Figure,figure-load-extension
@caption{Loading the extension}
@center @image{api-figure1, , , Loading the extension}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-load-extension" float="0">
<title>Loading the extension</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="api-figure1.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

The extension can call functions inside @command{gawk} through these
function pointers, at runtime, without needing (link-time) access
to @command{gawk}'s symbols.  One of these function pointers is to a
function for ``registering'' new functions.
@ifnotdocbook
This is shown in @ref{figure-register-new-function}.
@end ifnotdocbook
@ifdocbook
This is shown in @inlineraw{docbook, <xref linkend="figure-register-new-function"/>}.
@end ifdocbook

@ifnotdocbook
@float Figure,figure-register-new-function
@caption{Registering a new function}
@center @image{api-figure2, , , Registering a new Function}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-register-new-function" float="0">
<title>Registering a new function</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="api-figure2.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

In the other direction, the extension registers its new functions
with @command{gawk} by passing function pointers to the functions that
provide the new feature (@code{do_chdir()}, for example).  @command{gawk}
associates the function pointer with a name and can then call it, using a
defined calling convention.
@ifnotdocbook
This is shown in @ref{figure-call-new-function}.
@end ifnotdocbook
@ifdocbook
This is shown in @inlineraw{docbook, <xref linkend="figure-call-new-function"/>}.
@end ifdocbook

@ifnotdocbook
@float Figure,figure-call-new-function
@caption{Calling the new function}
@center @image{api-figure3, , , Calling the new function}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-call-new-function" float="0">
<title>Calling the new function</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="api-figure3.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

The @code{do_@var{xxx}()} function, in turn, then uses the function
pointers in the API @code{struct} to do its work, such as updating
variables or arrays, printing messages, setting @code{ERRNO}, and so on.

Convenience macros make calling through the function pointers look
like regular function calls so that extension code is quite readable
and understandable.

Although all of this sounds somewhat complicated, the result is that
extension code is quite straightforward to write and to read. You can
see this in the sample extension @file{filefuncs.c} (@pxref{Extension
Example}) and also in the @file{testext.c} code for testing the APIs.

Some other bits and pieces:

@itemize @value{BULLET}
@item
The API provides access to @command{gawk}'s @code{do_@var{xxx}} values,
reflecting command-line options, like @code{do_lint}, @code{do_profiling},
and so on (@pxref{Extension API Variables}).
These are informational: an extension cannot affect their values
inside @command{gawk}.  In addition, attempting to assign to them
produces a compile-time error.

@item
The API also provides major and minor version numbers, so that an
extension can check if the @command{gawk} it is loaded with supports the
facilities it was compiled with.  (Version mismatches ``shouldn't''
happen, but we all know how @emph{that} goes.)
@xref{Extension Versioning} for details.
@end itemize

@node Extension API Description
@section API Description
@cindex extension API

C or C++ code for an extension must include the header file
@file{gawkapi.h}, which declares the functions and defines the data
types used to communicate with @command{gawk}.
This (rather large) @value{SECTION} describes the API in detail.

@menu
* Extension API Functions Introduction:: Introduction to the API functions.
* General Data Types::                   The data types.
* Memory Allocation Functions::          Functions for allocating memory.
* Constructor Functions::                Functions for creating values.
* Registration Functions::               Functions to register things with
                                         @command{gawk}.
* Printing Messages::                    Functions for printing messages.
* Updating @code{ERRNO}::                Functions for updating @code{ERRNO}.
* Requesting Values::                    How to get a value.
* Accessing Parameters::                 Functions for accessing parameters.
* Symbol Table Access::                  Functions for accessing global
                                         variables.
* Array Manipulation::                   Functions for working with arrays.
* Redirection API::                      How to access and manipulate
                                         redirections.
* Extension API Variables::              Variables provided by the API.
* Extension API Boilerplate::            Boilerplate code for using the API.
* Changes from API V1::                  Changes from V1 of the API.
@end menu

@node Extension API Functions Introduction
@subsection Introduction

Access to facilities within @command{gawk} is achieved
by calling through function pointers passed into your extension.

API function pointers are provided for the following kinds of operations:

@itemize @value{BULLET}
@item
Allocating, reallocating, and releasing memory.

@item
Registration functions. You may register:

@c nested list
@itemize @value{MINUS}
@item
Extension functions
@item
Exit callbacks
@item
A version string
@item
Input parsers
@item
Output wrappers
@item
Two-way processors
@end itemize

All of these are discussed in detail later in this @value{CHAPTER}.

@item
Printing fatal, warning, and ``lint'' warning messages.

@item
Updating @code{ERRNO}, or unsetting it.

@item
Accessing parameters, including converting an undefined parameter into
an array.

@item
Symbol table access: retrieving a global variable, creating one,
or changing one.

@item
Creating and releasing cached values; this provides an
efficient way to use values for multiple variables and
can be a big performance win.

@item
Manipulating arrays:

@itemize @value{MINUS}
@item
Retrieving, adding, deleting, and modifying elements

@item
Getting the count of elements in an array

@item
Creating a new array

@item
Clearing an array

@item
Flattening an array for easy C-style looping over all its indices and elements
@end itemize

@item
Accessing and manipulating redirections.

@end itemize

Some points about using the API:

@itemize @value{BULLET}
@item
The following types, macros, and/or functions are referenced
in @file{gawkapi.h}.  For correct use, you must therefore include the
corresponding standard header file @emph{before} including @file{gawkapi.h}:

@c FIXME: Make this is a float at some point.
@multitable {@code{memset()}, @code{memcpy()}} {@code{<sys/types.h>}}
@headitem C entity @tab Header file
@item @code{EOF} @tab @code{<stdio.h>}
@item Values for @code{errno} @tab @code{<errno.h>}
@item @code{FILE} @tab @code{<stdio.h>}
@item @code{NULL} @tab @code{<stddef.h>}
@item @code{memcpy()} @tab @code{<string.h>}
@item @code{memset()} @tab @code{<string.h>}
@item @code{size_t} @tab @code{<sys/types.h>}
@item @code{struct stat} @tab @code{<sys/stat.h>}
@end multitable

Due to portability concerns, especially to systems that are not
fully standards-compliant, it is your responsibility
to include the correct files in the correct way. This requirement
is necessary in order to keep @file{gawkapi.h} clean, instead of becoming
a portability hodge-podge as can be seen in some parts of
the @command{gawk} source code.

@item
If your extension uses MPFR facilities, and you wish to receive such
values from @command{gawk} and/or pass such values to it, you must include the
@code{<mpfr.h>} header before including @code{<gawkapi.h>}.

@item
The @file{gawkapi.h} file may be included more than once without ill effect.
Doing so, however, is poor coding practice.

@item
Although the API only uses ISO C 90 features, there is an exception; the
``constructor'' functions use the @code{inline} keyword. If your compiler
does not support this keyword, you should either place
@samp{-Dinline=''} on your command line or use the GNU Autotools and include a
@file{config.h} file in your extensions.

@item
All pointers filled in by @command{gawk} point to memory
managed by @command{gawk} and should be treated by the extension as
read-only.  Memory for @emph{all} strings passed into @command{gawk}
from the extension @emph{must} come from calling one of
@code{gawk_malloc()}, @code{gawk_calloc()}, or @code{gawk_realloc()},
and is managed by @command{gawk} from then on.

@item
The API defines several simple @code{struct}s that map values as seen
from @command{awk}.  A value can be a @code{double}, a string, or an
array (as in multidimensional arrays, or when creating a new array).

String values maintain both pointer and length, because embedded @sc{nul}
characters are allowed.

@quotation NOTE
By intent, @command{gawk} maintains strings using the current multibyte
encoding (as defined by @env{LC_@var{xxx}} environment variables)
and not using wide characters.  This matches how @command{gawk} stores
strings internally and also how characters are likely to be input into
and output from files.
@end quotation

@quotation NOTE
String values passed to an extension by @command{gawk} are always
@sc{nul}-terminated.  Thus it is safe to pass such string values to
standard library and system routines. However, because @command{gawk}
allows embedded @sc{nul} characters in string data, before using the data
as a regular C string, you should check that the length for that string
passed to the extension matches the return value of @code{strlen()}
for it.
@end quotation

@item
When retrieving a value (such as a parameter or that of a global variable
or array element), the extension requests a specific type (number, string,
scalar, value cookie, array, or ``undefined'').  When the request is
``undefined,'' the returned value will have the real underlying type.

However, if the request and actual type don't match, the access function
returns ``false'' and fills in the type of the actual value that is there,
so that the extension can, e.g., print an error message
(such as ``scalar passed where array expected'').

@c This is documented in the header file and needs some expanding upon.
@c The table there should be presented here
@end itemize

You may call the API functions by using the function pointers
directly, but the interface is not so pretty. To make extension code look
more like regular code, the @file{gawkapi.h} header file defines several
macros that you should use in your code.  This @value{SECTION} presents
the macros as if they were functions.

@node General Data Types
@subsection General-Purpose Data Types

@cindex Robbins, Arnold
@cindex Ramey, Chet
@quotation
@i{I have a true love/hate relationship with unions.}
@author Arnold Robbins
@end quotation

@quotation
@i{That's the thing about unions: the compiler will arrange things so they
can accommodate both love and hate.}
@author Chet Ramey
@end quotation

The extension API defines a number of simple types and structures for
general-purpose use. Additional, more specialized, data structures are
introduced in subsequent @value{SECTION}s, together with the functions
that use them.

The general-purpose types and structures are as follows:

@table @code
@item typedef void *awk_ext_id_t;
A value of this type is received from @command{gawk} when an extension is loaded.
That value must then be passed back to @command{gawk} as the first parameter of
each API function.

@item #define awk_const @dots{}
This macro expands to @samp{const} when compiling an extension,
and to nothing when compiling @command{gawk} itself.  This makes
certain fields in the API data structures unwritable from extension code,
while allowing @command{gawk} to use them as it needs to.

@item typedef enum awk_bool @{
@itemx @ @ @ @ awk_false = 0,
@itemx @ @ @ @ awk_true
@itemx @} awk_bool_t;
A simple Boolean type.

@item typedef struct awk_string @{
@itemx @ @ @ @ char *str;@ @ @ @ @ @ /* data */
@itemx @ @ @ @ size_t len;@ @ @ @ @ /* length thereof, in chars */
@itemx @} awk_string_t;
This represents a mutable string. @command{gawk}
owns the memory pointed to if it supplied
the value. Otherwise, it takes ownership of the memory pointed to.
@emph{Such memory must come from calling one of the
@code{gawk_malloc()}, @code{gawk_calloc()}, or
@code{gawk_realloc()} functions!}

As mentioned earlier, strings are maintained using the current
multibyte encoding.

@item typedef enum @{
@itemx @ @ @ @ AWK_UNDEFINED,
@itemx @ @ @ @ AWK_NUMBER,
@itemx @ @ @ @ AWK_STRING,
@itemx @ @ @ @ AWK_REGEX,
@itemx @ @ @ @ AWK_STRNUM,
@itemx @ @ @ @ AWK_ARRAY,
@itemx @ @ @ @ AWK_SCALAR,@ @ @ @ @ @ @ @ @ /* opaque access to a variable */
@itemx @ @ @ @ AWK_VALUE_COOKIE@ @ @ @ /* for updating a previously created value */
@itemx @} awk_valtype_t;
This @code{enum} indicates the type of a value.
It is used in the following @code{struct}.

@item typedef struct awk_value @{
@itemx @ @ @ @ awk_valtype_t   val_type;
@itemx @ @ @ @ union @{
@itemx @ @ @ @ @ @ @ @ awk_string_t@ @ @ @ @ @ @ s;
@itemx @ @ @ @ @ @ @ @ awknum_t@ @ @ @ @ @ @ @ @ @ @ n;
@itemx @ @ @ @ @ @ @ @ awk_array_t@ @ @ @ @ @ @ @ a;
@itemx @ @ @ @ @ @ @ @ awk_scalar_t@ @ @ @ @ @ @ scl;
@itemx @ @ @ @ @ @ @ @ awk_value_cookie_t@ vc;
@itemx @ @ @ @ @} u;
@itemx @} awk_value_t;
An ``@command{awk} value.''
The @code{val_type} member indicates what kind of value the
@code{union} holds, and each member is of the appropriate type.

@item #define str_value@ @ @ @ @ @ u.s
@itemx #define strnum_value@ @ @ str_value
@itemx #define regex_value@ @ @ @ str_value
@itemx #define num_value@ @ @ @ @ @ u.n.d
@itemx #define num_type@ @ @ @ @ @ @ u.n.type
@itemx #define num_ptr@ @ @ @ @ @ @ @ u.n.ptr
@itemx #define array_cookie@ @ @ u.a
@itemx #define scalar_cookie@ @ u.scl
@itemx #define value_cookie@ @ @ u.vc
Using these macros makes accessing the fields of the @code{awk_value_t} more
readable.

@item enum AWK_NUMBER_TYPE @{
@itemx @ @ @ @ AWK_NUMBER_TYPE_DOUBLE,
@itemx @ @ @ @ AWK_NUMBER_TYPE_MPFR,
@itemx @ @ @ @ AWK_NUMBER_TYPE_MPZ
@itemx @};
This @code{enum} is used in the following structure for defining the
type of numeric value that is being worked with.  It is declared at the
top level of the file so that it works correctly for C++ as well as for C.

@item typedef struct awk_number @{
@itemx @ @ @ @ double d;
@itemx @ @ @ @ enum AWK_NUMBER_TYPE type;
@itemx @ @ @ @ void *ptr;
@itemx @} awk_number_t;
This represents a numeric value.  Internally, @command{gawk} stores
every number as either a C @code{double}, a GMP integer, or an MPFR
arbitrary-precision floating-point value.  In order to allow extensions
to also support GMP and MPFR values, numeric values are passed in this
structure.

The double-precision @code{d} element is always populated
in data received from @command{gawk}. In addition, by examining the
@code{type} member, an extension can determine if the @code{ptr}
member is either a GMP integer (type @code{mpz_ptr}), or an MPFR
floating-point value (type @code{mpfr_ptr_t}), and cast it appropriately.

@item typedef void *awk_scalar_t;
Scalars can be represented as an opaque type. These values are obtained
from @command{gawk} and then passed back into it. This is discussed
in a general fashion in the text following this list, and in more detail in
@ref{Symbol table by cookie}.

@item typedef void *awk_value_cookie_t;
A ``value cookie'' is an opaque type representing a cached value.
This is also discussed in a general fashion in the text following this list,
and in more detail in @ref{Cached values}.

@end table

Scalar values in @command{awk} are numbers, strings, strnums, or typed regexps. The
@code{awk_value_t} struct represents values.  The @code{val_type} member
indicates what is in the @code{union}.

Representing numbers is easy---the API uses a C @code{double}.  Strings
require more work. Because @command{gawk} allows embedded @sc{nul} bytes
in string values, a string must be represented as a pair containing a
data pointer and length. This is the @code{awk_string_t} type.

A strnum (numeric string) value is represented as a string and consists
of user input data that appears to be numeric.
When an extension creates a strnum value, the result is a string flagged
as user input. Subsequent parsing by @command{gawk} then determines whether it
looks like a number and should be treated as a strnum, or as a regular string.

This is useful in cases where an extension function would like to do something
comparable to the @code{split()} function which sets the strnum attribute
on the array elements it creates.  For example, an extension that implements
CSV splitting would want to use this feature. This is also useful for a
function that retrieves a data item from a database. The PostgreSQL
@code{PQgetvalue()} function, for example, returns a string that may be numeric
or textual depending on the contents.

Typed regexp values (@pxref{Strong Regexp Constants}) are not of
much use to extension functions.  Extension functions can tell that
they've received them, and create them for scalar values. Otherwise,
they can examine the text of the regexp through @code{regex_value.str}
and @code{regex_value.len}.

Identifiers (i.e., the names of global variables) can be associated
with either scalar values or with arrays.  In addition, @command{gawk}
provides true arrays of arrays, where any given array element can
itself be an array.  Discussion of arrays is delayed until
@ref{Array Manipulation}.

The various macros listed earlier make it easier to use the elements
of the @code{union} as if they were fields in a @code{struct}; this
is a common coding practice in C.  Such code is easier to write and to
read, but it remains @emph{your} responsibility to make sure that
the @code{val_type} member correctly reflects the type of the value in
the @code{awk_value_t} struct.

Conceptually, the first three members of the @code{union} (number, string,
and array) are all that is needed for working with @command{awk} values.
However, because the API provides routines for accessing and changing
the value of a global scalar variable only by using the variable's name,
there is a performance penalty: @command{gawk} must find the variable
each time it is accessed and changed.  This turns out to be a real issue,
not just a theoretical one.

Thus, if you know that your extension will spend considerable time
reading and/or changing the value of one or more scalar variables, you
can obtain a @dfn{scalar cookie}@footnote{See
@uref{http://catb.org/jargon/html/C/cookie.html, the ``cookie'' entry in the Jargon file} for a
definition of @dfn{cookie}, and @uref{http://catb.org/jargon/html/M/magic-cookie.html,
the ``magic cookie'' entry in the Jargon file} for a nice example.
@ifclear FOR_PRINT
See also the entry for ``Cookie'' in the @ref{Glossary}.
@end ifclear
}
object for that variable, and then use
the cookie for getting the variable's value or for changing the variable's
value.
The @code{awk_scalar_t} type holds a scalar cookie, and the
@code{scalar_cookie} macro provides access to the value of that type
in the @code{awk_value_t} struct.
Given a scalar cookie, @command{gawk} can directly retrieve or
modify the value, as required, without having to find it first.

The @code{awk_value_cookie_t} type and @code{value_cookie} macro are similar.
If you know that you wish to
use the same numeric or string @emph{value} for one or more variables,
you can create the value once, retaining a @dfn{value cookie} for it,
and then pass in that value cookie whenever you wish to set the value of a
variable.  This saves storage space within the running @command{gawk}
process and reduces the time needed to create the value.

@node Memory Allocation Functions
@subsection Memory Allocation Functions and Convenience Macros
@cindex allocating memory for extensions
@cindex extensions, allocating memory

The API provides a number of @dfn{memory allocation} functions for
allocating memory that can be passed to @command{gawk}, as well as a number of
convenience macros.
This @value{SUBSECTION} presents them all as function prototypes, in
the way that extension code would use them:

@table @code
@item void *gawk_malloc(size_t size);
Call the correct version of @code{malloc()} to allocate storage that may
be passed to @command{gawk}.

@item void *gawk_calloc(size_t nmemb, size_t size);
Call the correct version of @code{calloc()} to allocate storage that may
be passed to @command{gawk}.

@item void *gawk_realloc(void *ptr, size_t size);
Call the correct version of @code{realloc()} to allocate storage that may
be passed to @command{gawk}.

@item void gawk_free(void *ptr);
Call the correct version of @code{free()} to release storage that was
allocated with @code{gawk_malloc()}, @code{gawk_calloc()}, or @code{gawk_realloc()}.
@end table

The API has to provide these functions because it is possible
for an extension to be compiled and linked against a different
version of the C library than was used for the @command{gawk}
executable.@footnote{This is more common on MS-Windows systems, but it
can happen on Unix-like systems as well.} If @command{gawk} were
to use its version of @code{free()} when the memory came from an
unrelated version of @code{malloc()}, unexpected behavior would
likely result.

Three convenience macros may be used for allocating storage
from @code{gawk_malloc()}, @code{gawk_calloc}, and
@code{gawk_realloc()}. If the allocation fails, they cause @command{gawk}
to exit with a fatal error message.  They should be used as if they were
procedure calls that do not return a value:

@table @code
@item #define emalloc(pointer, type, size, message) @dots{}
The arguments to this macro are as follows:

@c nested table
@table @code
@item pointer
The pointer variable to point at the allocated storage.

@item type
The type of the pointer variable.  This is used to create a cast for
the call to @code{gawk_malloc()}.

@item size
The total number of bytes to be allocated.

@item message
A message to be prefixed to the fatal error message. Typically this is the name
of the function using the macro.
@end table

@noindent
For example, you might allocate a string value like so:

@example
@group
awk_value_t result;
char *message;
const char greet[] = "Don't Panic!";

emalloc(message, char *, sizeof(greet), "myfunc");
strcpy(message, greet);
make_malloced_string(message, strlen(message), & result);
@end group
@end example

@sp 2
@item #define ezalloc(pointer, type, size, message) @dots{}
This is like @code{emalloc()}, but it calls @code{gawk_calloc()}
instead of @code{gawk_malloc()}.
The arguments are the same as for the @code{emalloc()} macro, but this
macro guarantees that the memory returned is initialized to zero.

@item #define erealloc(pointer, type, size, message) @dots{}
This is like @code{emalloc()}, but it calls @code{gawk_realloc()}
instead of @code{gawk_malloc()}.
The arguments are the same as for the @code{emalloc()} macro.
@end table

Two additional functions allocate MPFR and GMP objects for use
by extension functions that need to create and then return such
values:

@table @code
@item void *get_mpfr_ptr();
Allocate and initialize an MPFR object and return a pointer to it.
If the allocation fails, @command{gawk} exits with a fatal
``out of memory'' error.  If @command{gawk} was compiled without
MPFR support, calling this function causes a fatal error.

@item void *get_mpz_ptr();
Allocate and initialize a GMP object and return a pointer to it.
If the allocation fails, @command{gawk} exits with a fatal
``out of memory'' error.  If @command{gawk} was compiled without
MPFR support, calling this function causes a fatal error.
@end table

Both of these functions return @samp{void *}, since the @file{gawkapi.h}
header file should not have dependency upon @code{<mpfr.h>} (and @code{<gmp.h>},
which is included from @code{<mpfr.h>}).  The actual return values are of
types @code{mpfr_ptr} and @code{mpz_ptr} respectively, and you should cast
the return values appropriately before assigning the results to variables
of the correct types.

@node Constructor Functions
@subsection Constructor Functions

The API provides a number of @dfn{constructor} functions for creating
string and numeric values, as well as a number of convenience macros.
This @value{SUBSECTION} presents them all as function prototypes, in
the way that extension code would use them:

@table @code
@item static inline awk_value_t *
@itemx make_const_string(const char *string, size_t length, awk_value_t *result);
This function creates a string value in the @code{awk_value_t} variable
pointed to by @code{result}. It expects @code{string} to be a C string constant
(or other string data), and automatically creates a @emph{copy} of the data
for storage in @code{result}. It returns @code{result}.

@item static inline awk_value_t *
@itemx make_malloced_string(const char *string, size_t length, awk_value_t *result);
This function creates a string value in the @code{awk_value_t} variable
pointed to by @code{result}. It expects @code{string} to be a @samp{char *}
value pointing to data previously obtained from @code{gawk_malloc()}, @code{gawk_calloc()}, or @code{gawk_realloc()}. The idea here
is that the data is passed directly to @command{gawk}, which assumes
responsibility for it. It returns @code{result}.

@item static inline awk_value_t *
@itemx make_null_string(awk_value_t *result);
This specialized function creates a null string (the ``undefined'' value)
in the @code{awk_value_t} variable pointed to by @code{result}.
It returns @code{result}.

@item static inline awk_value_t *
@itemx make_number(double num, awk_value_t *result);
This function simply creates a numeric value in the @code{awk_value_t} variable
pointed to by @code{result}.

@item static inline awk_value_t *
@itemx make_number_mpz(void *mpz, awk_value_t *result);
This function creates a GMP number value in @code{result}.
The @code{mpz} must be from a call to @code{get_mpz_ptr()}
(and thus be of real underlying type @code{mpz_ptr}).
@command{gawk} takes ownership of this memory.

@item static inline awk_value_t *
@itemx make_number_mpfr(void *mpfr, awk_value_t *result);
This function creates an MPFR number value in @code{result}.
The @code{mpfr} must be from a call to @code{get_mpfr_ptr()}.
(and thus be of real underlying type @code{mpfr_ptr})
@command{gawk} takes ownership of this memory.

@item static inline awk_value_t *
@itemx make_const_user_input(const char *string, size_t length, awk_value_t *result);
This function is identical to @code{make_const_string()}, but the string is
flagged as user input that should be treated as a strnum value if the contents
of the string are numeric.

@item static inline awk_value_t *
@itemx make_malloced_user_input(const char *string, size_t length, awk_value_t *result);
This function is identical to @code{make_malloced_string()}, but the string is
flagged as user input that should be treated as a strnum value if the contents
of the string are numeric.

@item static inline awk_value_t *
@itemx make_const_regex(const char *string, size_t length, awk_value_t *result);
This function creates a strongly typed regexp value by allocating a copy of the string.
@code{string} is the regular expression of length @code{len}.

@item static inline awk_value_t *
@itemx make_malloced_regex(const char *string, size_t length, awk_value_t *result);
This function creates a strongly typed regexp value.  @code{string} is
the regular expression of length @code{len}.  It expects @code{string}
to be a @samp{char *} value pointing to data previously obtained from
@code{gawk_malloc()}, @code{gawk_calloc()}, or @code{gawk_realloc()}.

@end table

@node Registration Functions
@subsection Registration Functions
@cindex register extension
@cindex extension registration

This @value{SECTION} describes the API functions for
registering parts of your extension with @command{gawk}.

@menu
* Extension Functions::         Registering extension functions.
* Exit Callback Functions::     Registering an exit callback.
* Extension Version String::    Registering a version string.
* Input Parsers::               Registering an input parser.
* Output Wrappers::             Registering an output wrapper.
* Two-way processors::          Registering a two-way processor.
@end menu

@node Extension Functions
@subsubsection Registering An Extension Function

Extension functions are described by the following record:

@example
@group
typedef struct awk_ext_func @{
@ @ @ @ const char *name;
@ @ @ @ awk_value_t *(*const function)(int num_actual_args,
@ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result,
@ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ struct awk_ext_func *finfo);
@ @ @ @ const size_t max_expected_args;
@ @ @ @ const size_t min_required_args;
@ @ @ @ awk_bool_t suppress_lint;
@ @ @ @ void *data;        /* opaque pointer to any extra state */
@} awk_ext_func_t;
@end group
@end example

The fields are:

@table @code
@item const char *name;
The name of the new function.
@command{awk}-level code calls the function by this name.
This is a regular C string.

Function names must obey the rules for @command{awk}
identifiers. That is, they must begin with either an English letter
or an underscore, which may be followed by any number of
letters, digits, and underscores.
Letter case in function names is significant.

@item awk_value_t *(*const function)(int num_actual_args,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ struct awk_ext_func *finfo);
This is a pointer to the C function that provides the extension's
functionality.
The function must fill in @code{*result} with either a number,
a string, or a regexp.
@command{gawk} takes ownership of any string memory.
As mentioned earlier, string memory @emph{must} come from one of
@code{gawk_malloc()}, @code{gawk_calloc()}, or @code{gawk_realloc()}.

The @code{num_actual_args} argument tells the C function how many
actual parameters were passed from the calling @command{awk} code.

The @code{finfo} parameter is a pointer to the @code{awk_ext_func_t} for
this function. The called function may access data within it as desired, or not.

The function must return the value of @code{result}.
This is for the convenience of the calling code inside @command{gawk}.

@item const size_t max_expected_args;
This is the maximum number of arguments the function expects to receive.
If called with more arguments than this, and if lint checking has
been enabled, then @command{gawk} prints a warning message.  For more
information, see the entry for @code{suppress_lint}, later in this list.

@item const size_t min_required_args;
This is the minimum number of arguments the function expects to receive.
If called with fewer arguments, @command{gawk} prints a fatal error
message and exits.

@item awk_bool_t suppress_lint;
This flag tells @command{gawk} not to print a lint message if lint
checking has been enabled and if more arguments were supplied in the call
than expected.  An extension function can tell if @command{gawk} already
printed at least one such message by checking if @samp{num_actual_args >
finfo->max_expected_args}.  If so, and the function does not want more
lint messages to be printed, it should set @code{finfo->suppress_lint}
to @code{awk_true}.

@item void *data;
This is an opaque pointer to any data that an extension function may
wish to have available when called.  Passing the @code{awk_ext_func_t}
structure to the extension function, and having this pointer available
in it enable writing a single C or C++ function that implements multiple
@command{awk}-level extension functions.
@end table

Once you have a record representing your extension function, you register
it with @command{gawk} using this API function:

@table @code
@item awk_bool_t add_ext_func(const char *name_space, awk_ext_func_t *func);
This function returns true upon success, false otherwise.
The @code{name_space} parameter is currently not used; you should pass in an
empty string (@code{""}).  The @code{func} pointer is the address of a
@code{struct} representing your function, as just described.

@command{gawk} does not modify what @code{func} points to, but the
extension function itself receives this pointer and can modify what it
points to, thus it is purposely not declared to be @code{const}.
@end table

The combination of @code{min_required_args}, @code{max_expected_args},
and @code{suppress_lint} may be confusing. Here is how you should
set things up.

@table @asis
@item Any number of arguments is valid
Set @code{min_required_args} and @code{max_expected_args} to zero and
set @code{suppress_lint} to @code{awk_true}.

@item A minimum number of arguments is required, no limit on maximum number of arguments
Set @code{min_required_args} to the minimum required. Set
@code{max_expected_args} to zero and
set @code{suppress_lint} to @code{awk_true}.

@item A minimum number of arguments is required, a maximum number is expected
Set @code{min_required_args} to the minimum required. Set
@code{max_expected_args} to the maximum expected.
Set @code{suppress_lint} to @code{awk_false}.

@item A minimum number of arguments is required, and no more than a maximum is allowed
Set @code{min_required_args} to the minimum required. Set
@code{max_expected_args} to the maximum expected.
Set @code{suppress_lint} to @code{awk_false}.
In your extension function, check that @code{num_actual_args} does not
exceed @code{f->max_expected_args}. If it does, issue a fatal error message.
@end table

@node Exit Callback Functions
@subsubsection Registering An Exit Callback Function

An @dfn{exit callback} function is a function that
@command{gawk} calls before it exits.
Such functions are useful if you have general ``cleanup'' tasks
that should be performed in your extension (such as closing database
connections or other resource deallocations).
You can register such
a function with @command{gawk} using the following function:

@table @code
@item void awk_atexit(void (*funcp)(void *data, int exit_status),
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ void *arg0);
The parameters are:

@c nested table
@table @code
@item funcp
A pointer to the function to be called before @command{gawk} exits. The @code{data}
parameter will be the original value of @code{arg0}.
The @code{exit_status} parameter is the exit status value that
@command{gawk} intends to pass to the @code{exit()} system call.

@item arg0
A pointer to private data that @command{gawk} saves in order to pass to
the function pointed to by @code{funcp}.
@end table
@end table

Exit callback functions are called in last-in, first-out (LIFO)
order---that is, in the reverse order in which they are registered with
@command{gawk}.

@node Extension Version String
@subsubsection Registering An Extension Version String

You can register a version string that indicates the name and
version of your extension with @command{gawk}, as follows:

@table @code
@item void register_ext_version(const char *version);
Register the string pointed to by @code{version} with @command{gawk}.
Note that @command{gawk} does @emph{not} copy the @code{version} string, so
it should not be changed.
@end table

@command{gawk} prints all registered extension version strings when it
is invoked with the @option{--version} option.

@node Input Parsers
@subsubsection Customized Input Parsers
@cindex customized input parser

By default, @command{gawk} reads text files as its input. It uses the value
of @code{RS} to find the end of the record, and then uses @code{FS}
(or @code{FIELDWIDTHS} or @code{FPAT}) to split it into fields (@pxref{Reading Files}).
Additionally, it sets the value of @code{RT} (@pxref{Built-in Variables}).

If you want, you can provide your own custom input parser.  An input
parser's job is to return a record to the @command{gawk} record-processing
code, along with indicators for the value and length of the data to be
used for @code{RT}, if any.

To provide an input parser, you must first provide two functions
(where @var{XXX} is a prefix name for your extension):

@table @code
@item awk_bool_t @var{XXX}_can_take_file(const awk_input_buf_t *iobuf);
This function examines the information available in @code{iobuf}
(which we discuss shortly).  Based on the information there, it
decides if the input parser should be used for this file.
If so, it should return true. Otherwise, it should return false.
It should not change any state (variable values, etc.) within @command{gawk}.

@item awk_bool_t @var{XXX}_take_control_of(awk_input_buf_t *iobuf);
When @command{gawk} decides to hand control of the file over to the
input parser, it calls this function.  This function in turn must fill
in certain fields in the @code{awk_input_buf_t} structure and ensure
that certain conditions are true.  It should then return true. If an
error of some kind occurs, it should not fill in any fields and should
return false; then @command{gawk} will not use the input parser.
The details are presented shortly.
@end table

Your extension should package these functions inside an
@code{awk_input_parser_t}, which looks like this:

@example
@group
typedef struct awk_input_parser @{
    const char *name;   /* name of parser */
    awk_bool_t (*can_take_file)(const awk_input_buf_t *iobuf);
    awk_bool_t (*take_control_of)(awk_input_buf_t *iobuf);
    awk_const struct awk_input_parser *awk_const next;   /* for gawk */
@} awk_input_parser_t;
@end group
@end example

The fields are:

@table @code
@item const char *name;
The name of the input parser. This is a regular C string.

@item awk_bool_t (*can_take_file)(const awk_input_buf_t *iobuf);
A pointer to your @code{@var{XXX}_can_take_file()} function.

@item awk_bool_t (*take_control_of)(awk_input_buf_t *iobuf);
A pointer to your @code{@var{XXX}_take_control_of()} function.

@item awk_const struct input_parser *awk_const next;
This is for use by @command{gawk};
therefore it is marked @code{awk_const} so that the extension cannot
modify it.
@end table

The steps are as follows:

@enumerate
@item
Create a @code{static awk_input_parser_t} variable and initialize it
appropriately.

@item
When your extension is loaded, register your input parser with
@command{gawk} using the @code{register_input_parser()} API function
(described next).
@end enumerate

An @code{awk_input_buf_t} looks like this:

@example
typedef struct awk_input @{
    const char *name;       /* filename */
    int fd;                 /* file descriptor */
#define INVALID_HANDLE (-1)
    void *opaque;           /* private data for input parsers */
    int (*get_record)(char **out, struct awk_input *iobuf,
                      int *errcode, char **rt_start, size_t *rt_len,
                      const awk_fieldwidth_info_t **field_width);
    ssize_t (*read_func)();
    void (*close_func)(struct awk_input *iobuf);
    struct stat sbuf;       /* stat buf */
@} awk_input_buf_t;
@end example

The fields can be divided into two categories: those for use (initially,
at least) by @code{@var{XXX}_can_take_file()}, and those for use by
@code{@var{XXX}_take_control_of()}.  The first group of fields and their uses
are as follows:

@table @code
@item const char *name;
The name of the file.

@item int fd;
A file descriptor for the file.  If @command{gawk} was able to
open the file, then @code{fd} will @emph{not} be equal to
@code{INVALID_HANDLE}. Otherwise, it will.

@item struct stat sbuf;
If the file descriptor is valid, then @command{gawk} will have filled
in this structure via a call to the @code{fstat()} system call.
@end table

The @code{@var{XXX}_can_take_file()} function should examine these
fields and decide if the input parser should be used for the file.
The decision can be made based upon @command{gawk} state (the value
of a variable defined previously by the extension and set by
@command{awk} code), the name of the
file, whether or not the file descriptor is valid, the information
in the @code{struct stat}, or any combination of these factors.

Once @code{@var{XXX}_can_take_file()} has returned true, and
@command{gawk} has decided to use your input parser, it calls
@code{@var{XXX}_take_control_of()}.  That function then fills
either the @code{get_record} field or the @code{read_func} field in
the @code{awk_input_buf_t}.  It must also ensure that @code{fd} is @emph{not}
set to @code{INVALID_HANDLE}.  The following list describes the fields that
may be filled by @code{@var{XXX}_take_control_of()}:

@table @code
@item void *opaque;
This is used to hold any state information needed by the input parser
for this file.  It is ``opaque'' to @command{gawk}.  The input parser
is not required to use this pointer.

@item int@ (*get_record)(char@ **out,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ struct@ awk_input *iobuf,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ int *errcode,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ char **rt_start,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ size_t *rt_len,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const awk_fieldwidth_info_t **field_width);
This function pointer should point to a function that creates the input
records.  Said function is the core of the input parser.  Its behavior
is described in the text following this list.

@item ssize_t (*read_func)();
This function pointer should point to a function that has the
same behavior as the standard POSIX @code{read()} system call.
It is an alternative to the @code{get_record} pointer.  Its behavior
is also described in the text following this list.

@item void (*close_func)(struct awk_input *iobuf);
This function pointer should point to a function that does
the ``teardown.'' It should release any resources allocated by
@code{@var{XXX}_take_control_of()}.  It may also close the file. If it
does so, it should set the @code{fd} field to @code{INVALID_HANDLE}.

If @code{fd} is still not @code{INVALID_HANDLE} after the call to this
function, @command{gawk} calls the regular @code{close()} system call.

Having a ``teardown'' function is optional. If your input parser does
not need it, do not set this field.  Then, @command{gawk} calls the
regular @code{close()} system call on the file descriptor, so it should
be valid.
@end table

The @code{@var{XXX}_get_record()} function does the work of creating
input records.  The parameters are as follows:

@table @code
@item char **out
This is a pointer to a @code{char *} variable that is set to point
to the record.  @command{gawk} makes its own copy of the data, so
the extension must manage this storage.

@item struct awk_input *iobuf
This is the @code{awk_input_buf_t} for the file.  The fields should be
used for reading data (@code{fd}) and for managing private state
(@code{opaque}), if any.

@item int *errcode
If an error occurs, @code{*errcode} should be set to an appropriate
code from @code{<errno.h>}.

@item char **rt_start
@itemx size_t *rt_len
If the concept of a ``record terminator'' makes sense, then
@code{*rt_start} should be set to point to the data to be used for
@code{RT}, and @code{*rt_len} should be set to the length of the
data. Otherwise, @code{*rt_len} should be set to zero.
@command{gawk} makes its own copy of this data, so the
extension must manage this storage.

@item const awk_fieldwidth_info_t **field_width
If @code{field_width} is not @code{NULL}, then @code{*field_width} will be initialized
to @code{NULL}, and the function may set it to point to a structure
supplying field width information to override the default
field parsing mechanism. Note that this structure will not
be copied by @command{gawk}; it must persist at least until the next call
to @code{get_record} or @code{close_func}. Note also that @code{field_width} is
@code{NULL} when @code{getline} is assigning the results to a variable, thus
field parsing is not needed. If the parser does set @code{*field_width},
then @command{gawk} uses this layout to parse the input record,
and the @code{PROCINFO["FS"]} value will be @code{"API"} while this record
is active in @code{$0}.
The @code{awk_fieldwidth_info_t} data structure
is described below.
@end table

The return value is the length of the buffer pointed to by
@code{*out}, or @code{EOF} if end-of-file was reached or an
error occurred.

It is guaranteed that @code{errcode} is a valid pointer, so there is no
need to test for a @code{NULL} value.  @command{gawk} sets @code{*errcode}
to zero, so there is no need to set it unless an error occurs.

If an error does occur, the function should return @code{EOF} and set
@code{*errcode} to a value greater than zero.  In that case, if @code{*errcode}
does not equal zero, @command{gawk} automatically updates
the @code{ERRNO} variable based on the value of @code{*errcode}.
(In general, setting @samp{*errcode = errno} should do the right thing.)

As an alternative to supplying a function that returns an input record,
you may instead supply a function that simply reads bytes, and let
@command{gawk} parse the data into records.  If you do so, the data
should be returned in the multibyte encoding of the current locale.
Such a function should follow the same behavior as the @code{read()}
system call, and you fill in the @code{read_func} pointer with its
address in the @code{awk_input_buf_t} structure.

By default, @command{gawk} sets the @code{read_func} pointer to
point to the @code{read()} system call. So your extension need not
set this field explicitly.

@quotation NOTE
You must choose one method or the other: either a function that
returns a record, or one that returns raw data.  In particular,
if you supply a function to get a record, @command{gawk} will
call it, and will never call the raw read function.
@end quotation

@command{gawk} ships with a sample extension that reads directories,
returning records for each entry in a directory (@pxref{Extension
Sample Readdir}).  You may wish to use that code as a guide for writing
your own input parser.

When writing an input parser, you should think about (and document)
how it is expected to interact with @command{awk} code.  You may want
it to always be called, and to take effect as appropriate (as the
@code{readdir} extension does).  Or you may want it to take effect
based upon the value of an @command{awk} variable, as the XML extension
from the @code{gawkextlib} project does (@pxref{gawkextlib}).
In the latter case, code in a @code{BEGINFILE} rule
can look at @code{FILENAME} and @code{ERRNO} to decide whether or
not to activate an input parser (@pxref{BEGINFILE/ENDFILE}).

You register your input parser with the following function:

@table @code
@item void register_input_parser(awk_input_parser_t *input_parser);
Register the input parser pointed to by @code{input_parser} with
@command{gawk}.
@end table

If you would like to override the default field parsing mechanism for a given
record, then you must populate an @code{awk_fieldwidth_info_t} structure,
which looks like this:

@example
typedef struct @{
        awk_bool_t     use_chars; /* false ==> use bytes */
        size_t         nf;        /* number of fields in record (NF) */
        struct awk_field_info @{
                size_t skip;      /* amount to skip before field starts */
                size_t len;       /* length of field */
        @} fields[1];              /* actual dimension should be nf */
@} awk_fieldwidth_info_t;
@end example

The fields are:

@table @code
@item awk_bool_t use_chars;
Set this to @code{awk_true} if the field lengths are specified in terms
of potentially multi-byte characters, and set it to @code{awk_false} if
the lengths are in terms of bytes.
Performance will be better if the values are supplied in
terms of bytes. 

@item size_t nf;
Set this to the number of fields in the input record, i.e. @code{NF}.

@item struct awk_field_info fields[nf];
This is a variable-length array whose actual dimension should be @code{nf}.
For each field, the @code{skip} element should be set to the number
of characters or bytes, as controlled by the @code{use_chars} flag,
to skip before the start of this field. The @code{len} element provides
the length of the field. The values in @code{fields[0]} provide the information
for @code{$1}, and so on through the @code{fields[nf-1]} element containing the information for @code{$NF}.
@end table

A convenience macro @code{awk_fieldwidth_info_size(numfields)} is provided to
calculate the appropriate size of a variable-length 
@code{awk_fieldwidth_info_t} structure containing @code{numfields} fields. This can
be used as an argument to @code{malloc()} or in a union to allocate space
statically. Please refer to the @code{readdir_test} sample extension for an
example.

@node Output Wrappers
@subsubsection Customized Output Wrappers
@cindex customized output wrapper

@cindex output wrapper
An @dfn{output wrapper} is the mirror image of an input parser.
It allows an extension to take over the output to a file opened
with the @samp{>} or @samp{>>} I/O redirection operators (@pxref{Redirection}).

The output wrapper is very similar to the input parser structure:

@example
typedef struct awk_output_wrapper @{
    const char *name;   /* name of the wrapper */
    awk_bool_t (*can_take_file)(const awk_output_buf_t *outbuf);
    awk_bool_t (*take_control_of)(awk_output_buf_t *outbuf);
    awk_const struct awk_output_wrapper *awk_const next;  /* for gawk */
@} awk_output_wrapper_t;
@end example

The members are as follows:

@table @code
@item const char *name;
This is the name of the output wrapper.

@item awk_bool_t (*can_take_file)(const awk_output_buf_t *outbuf);
This points to a function that examines the information in
the @code{awk_output_buf_t} structure pointed to by @code{outbuf}.
It should return true if the output wrapper wants to take over the
file, and false otherwise.  It should not change any state (variable
values, etc.) within @command{gawk}.

@item awk_bool_t (*take_control_of)(awk_output_buf_t *outbuf);
The function pointed to by this field is called when @command{gawk}
decides to let the output wrapper take control of the file. It should
fill in appropriate members of the @code{awk_output_buf_t} structure,
as described next, and return true if successful, false otherwise.

@item awk_const struct output_wrapper *awk_const next;
This is for use by @command{gawk};
therefore it is marked @code{awk_const} so that the extension cannot
modify it.
@end table

The @code{awk_output_buf_t} structure looks like this:

@example
typedef struct awk_output_buf @{
    const char *name;   /* name of output file */
    const char *mode;   /* mode argument to fopen */
    FILE *fp;           /* stdio file pointer */
    awk_bool_t redirected;  /* true if a wrapper is active */
    void *opaque;       /* for use by output wrapper */
    size_t (*gawk_fwrite)(const void *buf, size_t size, size_t count,
                FILE *fp, void *opaque);
    int (*gawk_fflush)(FILE *fp, void *opaque);
    int (*gawk_ferror)(FILE *fp, void *opaque);
    int (*gawk_fclose)(FILE *fp, void *opaque);
@} awk_output_buf_t;
@end example

Here too, your extension will define @code{@var{XXX}_can_take_file()}
and @code{@var{XXX}_take_control_of()} functions that examine and update
data members in the @code{awk_output_buf_t}.
The data members are as follows:

@table @code
@item const char *name;
The name of the output file.

@item const char *mode;
The mode string (as would be used in the second argument to @code{fopen()})
with which the file was opened.

@item FILE *fp;
The @code{FILE} pointer from @code{<stdio.h>}. @command{gawk} opens the file
before attempting to find an output wrapper.

@item awk_bool_t redirected;
This field must be set to true by the @code{@var{XXX}_take_control_of()} function.

@item void *opaque;
This pointer is opaque to @command{gawk}. The extension should use it to store
a pointer to any private data associated with the file.

@item size_t (*gawk_fwrite)(const void *buf, size_t size, size_t count,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ FILE *fp, void *opaque);
@itemx int (*gawk_fflush)(FILE *fp, void *opaque);
@itemx int (*gawk_ferror)(FILE *fp, void *opaque);
@itemx int (*gawk_fclose)(FILE *fp, void *opaque);
These pointers should be set to point to functions that perform
the equivalent function as the @code{<stdio.h>} functions do, if appropriate.
@command{gawk} uses these function pointers for all output.
@command{gawk} initializes the pointers to point to internal ``pass-through''
functions that just call the regular @code{<stdio.h>} functions, so an
extension only needs to redefine those functions that are appropriate for
what it does.
@end table

The @code{@var{XXX}_can_take_file()} function should make a decision based
upon the @code{name} and @code{mode} fields, and any additional state
(such as @command{awk} variable values) that is appropriate.

When @command{gawk} calls @code{@var{XXX}_take_control_of()}, that function should fill
in the other fields as appropriate, except for @code{fp}, which it should just
use normally.

You register your output wrapper with the following function:

@table @code
@item void register_output_wrapper(awk_output_wrapper_t *output_wrapper);
Register the output wrapper pointed to by @code{output_wrapper} with
@command{gawk}.
@end table

@node Two-way processors
@subsubsection Customized Two-way Processors
@cindex customized two-way processor

A @dfn{two-way processor} combines an input parser and an output wrapper for
two-way I/O with the @samp{|&} operator (@pxref{Redirection}).  It makes identical
use of the @code{awk_input_parser_t} and @code{awk_output_buf_t} structures
as described earlier.

A two-way processor is represented by the following structure:

@example
typedef struct awk_two_way_processor @{
    const char *name;   /* name of the two-way processor */
    awk_bool_t (*can_take_two_way)(const char *name);
    awk_bool_t (*take_control_of)(const char *name,
                                  awk_input_buf_t *inbuf,
                                  awk_output_buf_t *outbuf);
    awk_const struct awk_two_way_processor *awk_const next;  /* for gawk */
@} awk_two_way_processor_t;
@end example

The fields are as follows:

@table @code
@item const char *name;
The name of the two-way processor.

@item awk_bool_t (*can_take_two_way)(const char *name);
The function pointed to by this field should return true if it wants to take over two-way I/O for this @value{FN}.
It should not change any state (variable
values, etc.) within @command{gawk}.

@item awk_bool_t (*take_control_of)(const char *name,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_input_buf_t *inbuf,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_output_buf_t *outbuf);
The function pointed to by this field should fill in the @code{awk_input_buf_t} and
@code{awk_output_buf_t} structures pointed to by @code{inbuf} and
@code{outbuf}, respectively.  These structures were described earlier.

@item awk_const struct two_way_processor *awk_const next;
This is for use by @command{gawk};
therefore it is marked @code{awk_const} so that the extension cannot
modify it.
@end table

As with the input parser and output processor, you provide
``yes I can take this'' and ``take over for this'' functions,
@code{@var{XXX}_can_take_two_way()} and @code{@var{XXX}_take_control_of()}.

You register your two-way processor with the following function:

@table @code
@item void register_two_way_processor(awk_two_way_processor_t *two_way_processor);
Register the two-way processor pointed to by @code{two_way_processor} with
@command{gawk}.
@end table

@node Printing Messages
@subsection Printing Messages
@cindex printing messages from extensions
@cindex messages from extensions

You can print different kinds of warning messages from your
extension, as described here.  Note that for these functions,
you must pass in the extension ID received from @command{gawk}
when the extension was loaded:@footnote{Because the API uses only ISO C 90
features, it cannot make use of the ISO C 99 variadic macro feature to hide
that parameter. More's the pity.}

@table @code
@item void fatal(awk_ext_id_t id, const char *format, ...);
Print a message and then cause @command{gawk} to exit immediately.

@item void nonfatal(awk_ext_id_t id, const char *format, ...);
Print a nonfatal error message.

@item void warning(awk_ext_id_t id, const char *format, ...);
Print a warning message.

@item void lintwarn(awk_ext_id_t id, const char *format, ...);
Print a ``lint warning.''  Normally this is the same as printing a
warning message, but if @command{gawk} was invoked with @samp{--lint=fatal},
then lint warnings become fatal error messages.
@end table

All of these functions are otherwise like the C @code{printf()}
family of functions, where the @code{format} parameter is a string
with literal characters and formatting codes intermixed.

@node Updating @code{ERRNO}
@subsection Updating @code{ERRNO}

The following functions allow you to update the @code{ERRNO}
variable:

@table @code
@item void update_ERRNO_int(int errno_val);
Set @code{ERRNO} to the string equivalent of the error code
in @code{errno_val}. The value should be one of the defined
error codes in @code{<errno.h>}, and @command{gawk} turns it
into a (possibly translated) string using the C @code{strerror()} function.

@item void update_ERRNO_string(const char *string);
Set @code{ERRNO} directly to the string value of @code{ERRNO}.
@command{gawk} makes a copy of the value of @code{string}.

@item void unset_ERRNO(void);
Unset @code{ERRNO}.
@end table

@node Requesting Values
@subsection Requesting Values

All of the functions that return values from @command{gawk}
work in the same way. You pass in an @code{awk_valtype_t} value
to indicate what kind of value you expect.  If the actual value
matches what you requested, the function returns true and fills
in the @code{awk_value_t} result.
Otherwise, the function returns false, and the @code{val_type}
member indicates the type of the actual value.  You may then
print an error message or reissue the request for the actual
value type, as appropriate.  This behavior is summarized in
@ref{table-value-types-returned}.

@float Table,table-value-types-returned
@caption{API value types returned}
@docbook
<informaltable>
<tgroup cols="8">
  <colspec colname="c1"/>
  <colspec colname="c2"/>
  <colspec colname="c3"/>
  <colspec colname="c4"/>
  <colspec colname="c5"/>
  <colspec colname="c6"/>
  <colspec colname="c7"/>
  <colspec colname="c8"/>
  <spanspec spanname="hspan" namest="c3" nameend="c8" align="center"/>
  <thead>
    <row><entry></entry><entry spanname="hspan"><para>Type of Actual Value</para></entry></row>
    <row>
      <entry></entry>
      <entry></entry>
      <entry><para>String</para></entry>
      <entry><para>Strnum</para></entry>
      <entry><para>Number</para></entry>
      <entry><para>Regex</para></entry>
      <entry><para>Array</para></entry>
      <entry><para>Undefined</para></entry>
    </row>
  </thead>
  <tbody>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">String</emphasis></para></entry>
      <entry><para>String</para></entry>
      <entry><para>String</para></entry>
      <entry><para>String</para></entry>
      <entry><para>String</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">Strnum</emphasis></para></entry>
      <entry><para>false</para></entry>
      <entry><para>Strnum</para></entry>
      <entry><para>Strnum</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">Number</emphasis></para></entry>
      <entry><para>Number</para></entry>
      <entry><para>Number</para></entry>
      <entry><para>Number</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry><para><emphasis role="bold">Type</emphasis></para></entry>
      <entry><para><emphasis role="bold">Regex</emphasis></para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>Regex</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry><para><emphasis role="bold">Requested</emphasis></para></entry>
      <entry><para><emphasis role="bold">Array</emphasis></para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>Array</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">Scalar</emphasis></para></entry>
      <entry><para>Scalar</para></entry>
      <entry><para>Scalar</para></entry>
      <entry><para>Scalar</para></entry>
      <entry><para>Scalar</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">Undefined</emphasis></para></entry>
      <entry><para>String</para></entry>
      <entry><para>Strnum</para></entry>
      <entry><para>Number</para></entry>
      <entry><para>Regex</para></entry>
      <entry><para>Array</para></entry>
      <entry><para>Undefined</para></entry>
    </row>
    <row>
      <entry></entry>
      <entry><para><emphasis role="bold">Value cookie</emphasis></para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
      <entry><para>false</para></entry>
    </row>
  </tbody>
</tgroup>
</informaltable>
@end docbook

@ifnotplaintext
@ifnotdocbook
@multitable @columnfractions .50 .50
@headitem @tab Type of Actual Value
@end multitable
@c 10/2014: Thanks to Karl Berry for this bit to reduce the space:
@tex
\vglue-1.1\baselineskip
@end tex
@c @multitable @columnfractions .166 .166 .198 .15 .15 .166
@multitable {Requested} {Undefined} {Number} {Number} {Scalar} {Regex} {Array} {Undefined}
@headitem @tab @tab String @tab Strnum @tab Number @tab Regex @tab Array @tab Undefined
@item @tab @b{String} @tab String @tab String @tab String @tab String @tab false @tab false
@item @tab @b{Strnum} @tab false @tab Strnum @tab Strnum @tab false @tab false @tab false
@item @tab @b{Number} @tab Number @tab Number @tab Number @tab false @tab false @tab false
@item @b{Type} @tab @b{Regex} @tab false @tab false @tab false @tab Regex @tab false @tab false
@item @b{Requested} @tab @b{Array} @tab false @tab false @tab false @tab false @tab Array @tab false
@item @tab @b{Scalar} @tab Scalar @tab Scalar @tab Scalar @tab Scalar @tab false @tab false
@item @tab @b{Undefined} @tab String @tab Strnum @tab Number @tab Regex @tab Array @tab Undefined
@item @tab @b{Value cookie} @tab false @tab false @tab false @tab false @tab false @tab false
@end multitable
@end ifnotdocbook
@end ifnotplaintext
@ifplaintext
@verbatim
                        +-------------------------------------------------------+
                        |                   Type of Actual Value:               |
                        +--------+--------+--------+--------+-------+-----------+
                        | String | Strnum | Number | Regex  | Array | Undefined |
+-----------+-----------+--------+--------+--------+--------+-------+-----------+
|           | String    | String | String | String | String | false | false     |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Strnum    | false  | Strnum | Strnum | false  | false | false     |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Number    | Number | Number | Number | false  | false | false     |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Regex     | false  | false  | false  | Regex  | false | false     |
|   Type    +-----------+--------+--------+--------+--------+-------+-----------+
| Requested | Array     | false  | false  | false  | false  | Array | false     |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Scalar    | Scalar | Scalar | Scalar | Scalar | false | false     |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Undefined | String | Strnum | Number | Regex  | Array | Undefined |
|           +-----------+--------+--------+--------+--------+-------+-----------+
|           | Value     | false  | false  | false  | false  | false | false     |
|           | Cookie    |        |        |        |        |       |           |
+-----------+-----------+--------+--------+--------+--------+-------+-----------+
@end verbatim
@end ifplaintext
@end float

@node Accessing Parameters
@subsection Accessing and Updating Parameters

Two functions give you access to the arguments (parameters)
passed to your extension function. They are:

@table @code
@item awk_bool_t get_argument(size_t count,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t wanted,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result);
Fill in the @code{awk_value_t} structure pointed to by @code{result}
with the @code{count}th argument.  Return true if the actual
type matches @code{wanted}, and false otherwise.  In the latter
case, @code{result@w{->}val_type} indicates the actual type
(@pxref{table-value-types-returned}).  Counts are zero-based---the first
argument is numbered zero, the second one, and so on. @code{wanted}
indicates the type of value expected.

@item awk_bool_t set_argument(size_t count, awk_array_t array);
Convert a parameter that was undefined into an array; this provides
call by reference for arrays.  Return false if @code{count} is too big,
or if the argument's type is not undefined.  @xref{Array Manipulation}
for more information on creating arrays.
@end table

@node Symbol Table Access
@subsection Symbol Table Access
@cindex accessing global variables from extensions

Two sets of routines provide access to global variables, and one set
allows you to create and release cached values.

@menu
* Symbol table by name::        Accessing variables by name.
* Symbol table by cookie::      Accessing variables by ``cookie''.
* Cached values::               Creating and using cached values.
@end menu

@node Symbol table by name
@subsubsection Variable Access and Update by Name

The following routines provide the ability to access and update
global @command{awk}-level variables by name.  In compiler terminology,
identifiers of different kinds are termed @dfn{symbols}, thus the ``sym''
in the routines' names.  The data structure that stores information
about symbols is termed a @dfn{symbol table}.
The functions are as follows:

@table @code
@item awk_bool_t sym_lookup(const char *name,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t wanted,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result);
Fill in the @code{awk_value_t} structure pointed to by @code{result}
with the value of the variable named by the string @code{name}, which is
a regular C string.  @code{wanted} indicates the type of value expected.
Return true if the actual type matches @code{wanted}, and false otherwise.
In the latter case, @code{result->val_type} indicates the actual type
(@pxref{table-value-types-returned}).

@item awk_bool_t sym_update(const char *name, awk_value_t *value);
Update the variable named by the string @code{name}, which is a regular
C string.  The variable is added to @command{gawk}'s symbol table
if it is not there.  Return true if everything worked, and false otherwise.

Changing types (scalar to array or vice versa) of an existing variable
is @emph{not} allowed, nor may this routine be used to update an array.
This routine cannot be used to update any of the predefined
variables (such as @code{ARGC} or @code{NF}).
@end table

An extension can look up the value of @command{gawk}'s special variables.
However, with the exception of the @code{PROCINFO} array, an extension
cannot change any of those variables.

@node Symbol table by cookie
@subsubsection Variable Access and Update by Cookie

A @dfn{scalar cookie} is an opaque handle that provides access
to a global variable or array. It is an optimization that
avoids looking up variables in @command{gawk}'s symbol table every time
access is needed. This was discussed earlier, in @ref{General Data Types}.

@need 1500
The following functions let you work with scalar cookies:

@table @code
@item awk_bool_t sym_lookup_scalar(awk_scalar_t cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t wanted,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result);
Retrieve the current value of a scalar cookie.
Once you have obtained a scalar cookie using @code{sym_lookup()}, you can
use this function to get its value more efficiently.
Return false if the value cannot be retrieved.

@item awk_bool_t sym_update_scalar(awk_scalar_t cookie, awk_value_t *value);
Update the value associated with a scalar cookie.  Return false if
the new value is not of type @code{AWK_STRING}, @code{AWK_STRNUM}, @code{AWK_REGEX}, or @code{AWK_NUMBER}.
Here too, the predefined variables may not be updated.
@end table

It is not obvious at first glance how to work with scalar cookies or
what their @i{raison d'@^etre} really is.  In theory, the @code{sym_lookup()}
and @code{sym_update()} routines are all you really need to work with
variables.  For example, you might have code that looks up the value of
a variable, evaluates a condition, and then possibly changes the value
of the variable based on the result of that evaluation, like so:

@example
/*  do_magic --- do something really great */

static awk_value_t *
do_magic(int nargs, awk_value_t *result)
@{
    awk_value_t value;

    if (   sym_lookup("MAGIC_VAR", AWK_NUMBER, & value)
        && some_condition(value.num_value)) @{
            value.num_value += 42;
            sym_update("MAGIC_VAR", & value);
    @}

    return make_number(0.0, result);
@}
@end example

@noindent
This code looks (and is) simple and straightforward. So what's the problem?

Well, consider what happens if @command{awk}-level code associated
with your extension calls the @code{magic()} function (implemented in
C by @code{do_magic()}), once per record, while processing hundreds
of thousands or millions of records.  The @code{MAGIC_VAR} variable is
looked up in the symbol table once or twice per function call!

The symbol table lookup is really pure overhead; it is considerably
more efficient to get a cookie that represents the variable, and use
that to get the variable's value and update it as needed.@footnote{The
difference is measurable and quite real. Trust us.}

Thus, the way to use cookies is as follows.  First, install
your extension's variable in @command{gawk}'s symbol table using
@code{sym_update()}, as usual. Then get a scalar cookie for the variable
using @code{sym_lookup()}:

@example
@group
static awk_scalar_t magic_var_cookie;    /* cookie for MAGIC_VAR */

static void
my_extension_init()
@{
    awk_value_t value;
@end group

    /* install initial value */
    sym_update("MAGIC_VAR", make_number(42.0, & value));

    /* get the cookie */
    sym_lookup("MAGIC_VAR", AWK_SCALAR, & value);

    /* save the cookie */
    magic_var_cookie = value.scalar_cookie;
    @dots{}
@}
@end example

Next, use the routines in this @value{SECTION} for retrieving and updating
the value through the cookie.  Thus, @code{do_magic()} now becomes
something like this:

@example
/*  do_magic --- do something really great */

static awk_value_t *
do_magic(int nargs, awk_value_t *result)
@{
    awk_value_t value;

    if (   sym_lookup_scalar(magic_var_cookie, AWK_NUMBER, & value)
        && some_condition(value.num_value)) @{
            value.num_value += 42;
            sym_update_scalar(magic_var_cookie, & value);
    @}
    @dots{}

    return make_number(0.0, result);
@}
@end example

@quotation NOTE
The previous code omitted error checking for
presentation purposes.  Your extension code should be more robust
and carefully check the return values from the API functions.
@end quotation

@node Cached values
@subsubsection Creating and Using Cached Values

The routines in this @value{SECTION} allow you to create and release
cached values.  Like scalar cookies, in theory, cached values
are not necessary. You can create numbers and strings using
the functions in @ref{Constructor Functions}. You can then
assign those values to variables using @code{sym_update()}
or @code{sym_update_scalar()}, as you like.

However, you can understand the point of cached values if you remember that
@emph{every} string value's storage @emph{must} come from @code{gawk_malloc()},
@code{gawk_calloc()}, or @code{gawk_realloc()}.
If you have 20 variables, all of which have the same string value, you
must create 20 identical copies of the string.@footnote{Numeric values
are clearly less problematic, requiring only a C @code{double} to store.
But of course, GMP and MPFR values @emph{do} take up more memory.}

It is clearly more efficient, if possible, to create a value once, and
then tell @command{gawk} to reuse the value for multiple variables. That
is what the routines in this @value{SECTION} let you do.  The functions are as follows:

@table @code
@item awk_bool_t create_value(awk_value_t *value, awk_value_cookie_t *result);
Create a cached string or numeric value from @code{value} for
efficient later assignment.  Only values of type @code{AWK_NUMBER}, @code{AWK_REGEX}, @code{AWK_STRNUM},
and @code{AWK_STRING} are allowed.  Any other type is rejected.
@code{AWK_UNDEFINED} could be allowed, but doing so would result in
inferior performance.

@item awk_bool_t release_value(awk_value_cookie_t vc);
Release the memory associated with a value cookie obtained
from @code{create_value()}.
@end table

You use value cookies in a fashion similar to the way you use scalar cookies.
In the extension initialization routine, you create the value cookie:

@example
static awk_value_cookie_t answer_cookie;  /* static value cookie */

static void
my_extension_init()
@{
    awk_value_t value;
    char *long_string;
    size_t long_string_len;

    /* code from earlier */
    @dots{}
    /* @dots{} fill in long_string and long_string_len @dots{} */
    make_malloced_string(long_string, long_string_len, & value);
    create_value(& value, & answer_cookie);    /* create cookie */
    @dots{}
@}
@end example

Once the value is created, you can use it as the value of any number
of variables:

@example
static awk_value_t *
do_magic(int nargs, awk_value_t *result)
@{
    awk_value_t new_value;

    @dots{}    /* as earlier */

    value.val_type = AWK_VALUE_COOKIE;
    value.value_cookie = answer_cookie;
    sym_update("VAR1", & value);
    sym_update("VAR2", & value);
    @dots{}
    sym_update("VAR100", & value);
    @dots{}
@}
@end example

@noindent
Using value cookies in this way saves considerable storage, as all of
@code{VAR1} through @code{VAR100} share the same value.

You might be wondering, ``Is this sharing problematic?
What happens if @command{awk} code assigns a new value to @code{VAR1};
are all the others changed too?''

That's a great question. The answer is that no, it's not a problem.
Internally, @command{gawk} uses @dfn{reference-counted strings}. This means
that many variables can share the same string value, and @command{gawk}
keeps track of the usage.  When a variable's value changes, @command{gawk}
simply decrements the reference count on the old value and updates
the variable to use the new value.

Finally, as part of your cleanup action (@pxref{Exit Callback Functions})
you should release any cached values that you created, using
@code{release_value()}.

@node Array Manipulation
@subsection Array Manipulation
@cindex array manipulation in extensions

The primary data structure@footnote{OK, the only data structure.} in @command{awk}
is the associative array (@pxref{Arrays}).
Extensions need to be able to manipulate @command{awk} arrays.
The API provides a number of data structures for working with arrays,
functions for working with individual elements, and functions for
working with arrays as a whole. This includes the ability to
``flatten'' an array so that it is easy for C code to traverse
every element in an array.  The array data structures integrate
nicely with the data structures for values to make it easy to
both work with and create true arrays of arrays (@pxref{General Data Types}).

@menu
* Array Data Types::            Data types for working with arrays.
* Array Functions::             Functions for working with arrays.
* Flattening Arrays::           How to flatten arrays.
* Creating Arrays::             How to create and populate arrays.
@end menu

@node Array Data Types
@subsubsection Array Data Types

The data types associated with arrays are as follows:

@table @code
@item typedef void *awk_array_t;
If you request the value of an array variable, you get back an
@code{awk_array_t} value. This value is opaque@footnote{It is also
a ``cookie,'' but the @command{gawk} developers did not wish to overuse this
term.} to the extension; it uniquely identifies the array but can
only be used by passing it into API functions or receiving it from API
functions. This is very similar to way @samp{FILE *} values are used
with the @code{<stdio.h>} library routines.

@item typedef struct awk_element @{
@itemx @ @ @ @ /* convenience linked list pointer, not used by gawk */
@itemx @ @ @ @ struct awk_element *next;
@itemx @ @ @ @ enum @{
@itemx @ @ @ @ @ @ @ @ AWK_ELEMENT_DEFAULT = 0,@ @ /* set by gawk */
@itemx @ @ @ @ @ @ @ @ AWK_ELEMENT_DELETE = 1@ @ @ @ /* set by extension */
@itemx @ @ @ @ @} flags;
@itemx @ @ @ @ awk_value_t    index;
@itemx @ @ @ @ awk_value_t    value;
@itemx @} awk_element_t;
The @code{awk_element_t} is a ``flattened''
array element. @command{awk} produces an array of these
inside the @code{awk_flat_array_t} (see the next item).
Individual elements may be marked for deletion. New elements must be added
individually, one at a time, using the separate API for that purpose.
The fields are as follows:

@c nested table
@table @code
@item struct awk_element *next;
This pointer is for the convenience of extension writers.  It allows
an extension to create a linked list of new elements that can then be
added to an array in a loop that traverses the list.

@item enum @{ @dots{} @} flags;
A set of flag values that convey information between the extension
and @command{gawk}.  Currently there is only one: @code{AWK_ELEMENT_DELETE}.
Setting it causes @command{gawk} to delete the
element from the original array upon release of the flattened array.

@item index
@itemx value
The index and value of the element, respectively.
@emph{All} memory pointed to by @code{index} and @code{value} belongs to @command{gawk}.
@end table

@item typedef struct awk_flat_array @{
@itemx @ @ @ @ awk_const void *awk_const opaque1;@ @ @ @ /* for use by gawk */
@itemx @ @ @ @ awk_const void *awk_const opaque2;@ @ @ @ /* for use by gawk */
@itemx @ @ @ @ awk_const size_t count;@ @ @ @ @ /* how many elements */
@itemx @ @ @ @ awk_element_t elements[1];@ @ /* will be extended */
@itemx @} awk_flat_array_t;
This is a flattened array. When an extension gets one of these
from @command{gawk}, the @code{elements} array is of actual
size @code{count}.
The @code{opaque1} and @code{opaque2} pointers are for use by @command{gawk};
therefore they are marked @code{awk_const} so that the extension cannot
modify them.
@end table

@node Array Functions
@subsubsection Array Functions

The following functions relate to individual array elements:

@table @code
@item awk_bool_t get_element_count(awk_array_t a_cookie, size_t *count);
For the array represented by @code{a_cookie}, place in @code{*count}
the number of elements it contains. A subarray counts as a single element.
Return false if there is an error.

@item awk_bool_t get_array_element(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const awk_value_t *const index,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t wanted,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_value_t *result);
For the array represented by @code{a_cookie}, return in @code{*result}
the value of the element whose index is @code{index}.
@code{wanted} specifies the type of value you wish to retrieve.
Return false if @code{wanted} does not match the actual type or if
@code{index} is not in the array (@pxref{table-value-types-returned}).

The value for @code{index} can be numeric, in which case @command{gawk}
converts it to a string. Using nonintegral values is possible, but
requires that you understand how such values are converted to strings
(@pxref{Conversion}); thus, using integral values is safest.

As with @emph{all} strings passed into @command{gawk} from an extension,
the string value of @code{index} must come from @code{gawk_malloc()},
@code{gawk_calloc()}, or @code{gawk_realloc()}, and
@command{gawk} releases the storage.

@item awk_bool_t set_array_element(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const@ awk_value_t *const index,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const@ awk_value_t *const value);
In the array represented by @code{a_cookie}, create or modify
the element whose index is given by @code{index}.
The @code{ARGV} and @code{ENVIRON} arrays may not be changed,
although the @code{PROCINFO} array can be.

@item awk_bool_t set_array_element_by_elem(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_element_t element);
Like @code{set_array_element()}, but take the @code{index} and @code{value}
from @code{element}. This is a convenience macro.

@item awk_bool_t del_array_element(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const awk_value_t* const index);
Remove the element with the given index from the array
represented by @code{a_cookie}.
Return true if the element was removed, or false if the element did
not exist in the array.
@end table

The following functions relate to arrays as a whole:

@table @code
@item awk_array_t create_array(void);
Create a new array to which elements may be added.
@xref{Creating Arrays} for a discussion of how to
create a new array and add elements to it.

@item awk_bool_t clear_array(awk_array_t a_cookie);
Clear the array represented by @code{a_cookie}.
Return false if there was some kind of problem, true otherwise.
The array remains an array, but after calling this function, it
has no elements. This is equivalent to using the @code{delete}
statement (@pxref{Delete}).

@item awk_bool_t flatten_array_typed(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_flat_array_t **data,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t index_type,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_valtype_t value_type);
For the array represented by @code{a_cookie}, create an @code{awk_flat_array_t}
structure and fill it in with indices and values of the requested types.
Set the pointer whose address is passed as @code{data}
to point to this structure.
Return true upon success, or false otherwise.
@ifset FOR_PRINT
See the next @value{SECTION}
@end ifset
@ifclear FOR_PRINT
@xref{Flattening Arrays},
@end ifclear
for a discussion of how to
flatten an array and work with it.

@item awk_bool_t flatten_array(awk_array_t a_cookie, awk_flat_array_t **data);
For the array represented by @code{a_cookie}, create an @code{awk_flat_array_t}
structure and fill it in with @code{AWK_STRING} indices and
@code{AWK_UNDEFINED} values.
This is superseded by @code{flatten_array_typed()}.
It is provided as a macro, and remains for convenience and for source code
compatibility with the previous version of the API.

@item awk_bool_t release_flattened_array(awk_array_t a_cookie,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ awk_flat_array_t *data);
When done with a flattened array, release the storage using this function.
You must pass in both the original array cookie and the address of
the created @code{awk_flat_array_t} structure.
The function returns true upon success, false otherwise.
@end table

@node Flattening Arrays
@subsubsection Working With All The Elements of an Array

To @dfn{flatten} an array is to create a structure that
represents the full array in a fashion that makes it easy
for C code to traverse the entire array.  Some of the code
in @file{extension/testext.c} does this, and also serves
as a nice example showing how to use the APIs.

We walk through that part of the code one step at a time.
First, the @command{gawk} script that drives the test extension:

@example
@@load "testext"
BEGIN @{
    n = split("blacky rusty sophie raincloud lucky", pets)
    printf("pets has %d elements\n", length(pets))
    ret = dump_array_and_delete("pets", "3")
    printf("dump_array_and_delete(pets) returned %d\n", ret)
    if ("3" in pets)
        printf("dump_array_and_delete() did NOT remove index \"3\"!\n")
    else
        printf("dump_array_and_delete() did remove index \"3\"!\n")
    print ""
@}
@end example

@noindent
This code creates an array with @code{split()} (@pxref{String Functions})
and then calls @code{dump_array_and_delete()}. That function looks up
the array whose name is passed as the first argument, and
deletes the element at the index passed in the second argument.
The @command{awk} code then prints the return value and checks if the element
was indeed deleted.  Here is the C code that implements
@code{dump_array_and_delete()}. It has been edited slightly for
presentation.

The first part declares variables, sets up the default
return value in @code{result}, and checks that the function
was called with the correct number of arguments:

@example
static awk_value_t *
dump_array_and_delete(int nargs, awk_value_t *result)
@{
    awk_value_t value, value2, value3;
    awk_flat_array_t *flat_array;
    size_t count;
    char *name;
    int i;

    assert(result != NULL);
    make_number(0.0, result);

    if (nargs != 2) @{
        printf("dump_array_and_delete: nargs not right "
               "(%d should be 2)\n", nargs);
        goto out;
    @}
@end example

The function then proceeds in steps, as follows. First, retrieve
the name of the array, passed as the first argument, followed by
the array itself. If either operation fails, print an
error message and return:

@example
    /* get argument named array as flat array and print it */
    if (get_argument(0, AWK_STRING, & value)) @{
        name = value.str_value.str;
        if (sym_lookup(name, AWK_ARRAY, & value2))
            printf("dump_array_and_delete: sym_lookup of %s passed\n",
                   name);
        else @{
            printf("dump_array_and_delete: sym_lookup of %s failed\n",
                   name);
            goto out;
        @}
    @} else @{
        printf("dump_array_and_delete: get_argument(0) failed\n");
        goto out;
    @}
@end example

For testing purposes and to make sure that the C code sees
the same number of elements as the @command{awk} code,
the second step is to get the count of elements in the array
and print it:

@example
    if (! get_element_count(value2.array_cookie, & count)) @{
        printf("dump_array_and_delete: get_element_count failed\n");
        goto out;
    @}

    printf("dump_array_and_delete: incoming size is %lu\n",
           (unsigned long) count);
@end example

The third step is to actually flatten the array, and then
to double-check that the count in the @code{awk_flat_array_t}
is the same as the count just retrieved:

@example
    if (! flatten_array_typed(value2.array_cookie, & flat_array,
                              AWK_STRING, AWK_UNDEFINED)) @{
        printf("dump_array_and_delete: could not flatten array\n");
        goto out;
    @}

    if (flat_array->count != count) @{
        printf("dump_array_and_delete: flat_array->count (%lu)"
               " != count (%lu)\n",
                (unsigned long) flat_array->count,
                (unsigned long) count);
        goto out;
    @}
@end example

The fourth step is to retrieve the index of the element
to be deleted, which was passed as the second argument.
Remember that argument counts passed to @code{get_argument()}
are zero-based, and thus the second argument is numbered one:

@example
    if (! get_argument(1, AWK_STRING, & value3)) @{
        printf("dump_array_and_delete: get_argument(1) failed\n");
        goto out;
    @}
@end example

The fifth step is where the ``real work'' is done. The function
loops over every element in the array, printing the index and
element values. In addition, upon finding the element with the
index that is supposed to be deleted, the function sets the
@code{AWK_ELEMENT_DELETE} bit in the @code{flags} field
of the element.  When the array is released, @command{gawk}
traverses the flattened array, and deletes any elements that
have this flag bit set:

@example
    for (i = 0; i < flat_array->count; i++) @{
        printf("\t%s[\"%.*s\"] = %s\n",
            name,
            (int) flat_array->elements[i].index.str_value.len,
            flat_array->elements[i].index.str_value.str,
            valrep2str(& flat_array->elements[i].value));

        if (strcmp(value3.str_value.str,
                   flat_array->elements[i].index.str_value.str) == 0) @{
            flat_array->elements[i].flags |= AWK_ELEMENT_DELETE;
            printf("dump_array_and_delete: marking element \"%s\" "
                   "for deletion\n",
                flat_array->elements[i].index.str_value.str);
        @}
    @}
@end example

The sixth step is to release the flattened array. This tells
@command{gawk} that the extension is no longer using the array,
and that it should delete any elements marked for deletion.
@command{gawk} also frees any storage that was allocated,
so you should not use the pointer (@code{flat_array} in this
code) once you have called @code{release_flattened_array()}:

@example
    if (! release_flattened_array(value2.array_cookie, flat_array)) @{
        printf("dump_array_and_delete: could not release flattened array\n");
        goto out;
    @}
@end example

Finally, because everything was successful, the function sets the
return value to success, and returns:

@example
@group
    make_number(1.0, result);
out:
    return result;
@}
@end group
@end example

Here is the output from running this part of the test:

@example
pets has 5 elements
dump_array_and_delete: sym_lookup of pets passed
dump_array_and_delete: incoming size is 5
        pets["1"] = "blacky"
        pets["2"] = "rusty"
        pets["3"] = "sophie"
dump_array_and_delete: marking element "3" for deletion
        pets["4"] = "raincloud"
        pets["5"] = "lucky"
dump_array_and_delete(pets) returned 1
dump_array_and_delete() did remove index "3"!
@end example

@node Creating Arrays
@subsubsection How To Create and Populate Arrays

Besides working with arrays created by @command{awk} code, you can
create arrays and populate them as you see fit, and then @command{awk}
code can access them and manipulate them.

There are two important points about creating arrays from extension code:

@itemize @value{BULLET}
@item
You must install a new array into @command{gawk}'s symbol
table immediately upon creating it.  Once you have done so,
you can then populate the array.

@ignore
Strictly speaking, this is required only
for arrays that will have subarrays as elements; however it is
a good idea to always do this.  This restriction may be relaxed
in a subsequent revision of the API.
@end ignore

Similarly, if installing a new array as a subarray of an existing array,
you must add the new array to its parent before adding any elements to it.

Thus, the correct way to build an array is to work ``top down.''  Create
the array, and immediately install it in @command{gawk}'s symbol table
using @code{sym_update()}, or install it as an element in a previously
existing array using @code{set_array_element()}.  We show example code shortly.

@item
Due to @command{gawk} internals, after using @code{sym_update()} to install an array
into @command{gawk}, you have to retrieve the array cookie from the value
passed in to @command{sym_update()} before doing anything else with it, like so:

@example
awk_value_t val;
awk_array_t new_array;

new_array = create_array();
val.val_type = AWK_ARRAY;
val.array_cookie = new_array;

/* install array in the symbol table */
sym_update("array", & val);

new_array = val.array_cookie;    /* YOU MUST DO THIS */
@end example

If installing an array as a subarray, you must also retrieve the value
of the array cookie after the call to @code{set_element()}.
@end itemize

The following C code is a simple test extension to create an array
with two regular elements and with a subarray. The leading @code{#include}
directives and boilerplate variable declarations
(@pxref{Extension API Boilerplate})
are omitted for brevity.
The first step is to create a new array and then install it
in the symbol table:

@example
@ignore
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <stdio.h>
#include <assert.h>
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

#include <sys/types.h>
#include <sys/stat.h>

#include "gawkapi.h"

static const gawk_api_t *api;   /* for convenience macros to work */
static awk_ext_id_t ext_id;
static const char *ext_version = "testarray extension: version 1.0";

int plugin_is_GPL_compatible;

@end ignore
/* create_new_array --- create a named array */

static void
create_new_array()
@{
    awk_array_t a_cookie;
    awk_array_t subarray;
    awk_value_t index, value;

    a_cookie = create_array();
    value.val_type = AWK_ARRAY;
    value.array_cookie = a_cookie;

    if (! sym_update("new_array", & value))
        printf("create_new_array: sym_update(\"new_array\") failed!\n");
    a_cookie = value.array_cookie;
@end example

@noindent
Note how @code{a_cookie} is reset from the @code{array_cookie} field in
the @code{value} structure.

The second step is to install two regular values into @code{new_array}:

@example
    (void) make_const_string("hello", 5, & index);
    (void) make_const_string("world", 5, & value);
    if (! set_array_element(a_cookie, & index, & value)) @{
        printf("fill_in_array: set_array_element failed\n");
        return;
    @}

    (void) make_const_string("answer", 6, & index);
    (void) make_number(42.0, & value);
    if (! set_array_element(a_cookie, & index, & value)) @{
        printf("fill_in_array: set_array_element failed\n");
        return;
    @}
@end example

The third step is to create the subarray and install it:

@example
    (void) make_const_string("subarray", 8, & index);
    subarray = create_array();
    value.val_type = AWK_ARRAY;
    value.array_cookie = subarray;
    if (! set_array_element(a_cookie, & index, & value)) @{
        printf("fill_in_array: set_array_element failed\n");
        return;
    @}
    subarray = value.array_cookie;
@end example

The final step is to populate the subarray with its own element:

@example
    (void) make_const_string("foo", 3, & index);
    (void) make_const_string("bar", 3, & value);
    if (! set_array_element(subarray, & index, & value)) @{
        printf("fill_in_array: set_array_element failed\n");
        return;
    @}
@}
@ignore
static awk_ext_func_t func_table[] = @{
    @{ NULL, NULL, 0 @}
@};

/* init_testarray --- additional initialization function */

static awk_bool_t init_testarray(void)
@{
    create_new_array();

    return awk_true;
@}

static awk_bool_t (*init_func)(void) = init_testarray;

dl_load_func(func_table, testarray, "")
@end ignore
@end example

Here is a sample script that loads the extension
and then dumps the array:

@example
@@load "subarray"

function dumparray(name, array,     i)
@{
    for (i in array)
        if (isarray(array[i]))
            dumparray(name "[\"" i "\"]", array[i])
        else
            printf("%s[\"%s\"] = %s\n", name, i, array[i])
@}

BEGIN @{
    dumparray("new_array", new_array);
@}
@end example

Here is the result of running the script:

@example
$ @kbd{AWKLIBPATH=$PWD gawk -f subarray.awk}
@print{} new_array["subarray"]["foo"] = bar
@print{} new_array["hello"] = world
@print{} new_array["answer"] = 42
@end example

@noindent
(@xref{Finding Extensions} for more information on the
@env{AWKLIBPATH} environment variable.)

@node Redirection API
@subsection Accessing and Manipulating Redirections

The following function allows extensions to access and manipulate redirections.

@table @code
@item awk_bool_t get_file(const char *name,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ size_t name_len,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const char *filetype,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ int fd,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const awk_input_buf_t **ibufp,
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ const awk_output_buf_t **obufp);
Look up file @code{name} in @command{gawk}'s internal redirection table.
If @code{name} is @code{NULL} or @code{name_len} is zero, return
data for the currently open input file corresponding to @code{FILENAME}.
(This does not access the @code{filetype} argument, so that may be undefined).  
If the file is not already open, attempt to open it.
The @code{filetype} argument must be zero-terminated and should be one of:

@table @code
@item ">"
A file opened for output.

@item ">>"
A file opened for append.

@item "<"
A file opened for input.

@item "|>"
A pipe opened for output.

@item "|<"
A pipe opened for input.

@item "|&"
A two-way coprocess.
@end table

On error, return @code{awk_false}.  Otherwise, return
@code{awk_true}, and return additional information about the redirection
in the @code{ibufp} and @code{obufp} pointers.

For input redirections, the @code{*ibufp} value should be non-@code{NULL},
and @code{*obufp} should be @code{NULL}.  For output redirections,
the @code{*obufp} value should be non-@code{NULL}, and @code{*ibufp}
should be @code{NULL}.  For two-way coprocesses, both values should
be non-@code{NULL}.

In the usual case, the extension is interested in @code{(*ibufp)->fd}
and/or @code{fileno((*obufp)->fp)}.  If the file is not already
open, and the @code{fd} argument is nonnegative, @command{gawk}
will use that file descriptor instead of opening the file in the
usual way.  If @code{fd} is nonnegative, but the file exists already,
@command{gawk} ignores @code{fd} and returns the existing file.  It is
the caller's responsibility to notice that neither the @code{fd} in
the returned @code{awk_input_buf_t} nor the @code{fd} in the returned
@code{awk_output_buf_t} matches the requested value.

Note that supplying a file descriptor is currently @emph{not} supported
for pipes.  However, supplying a file descriptor should work for input,
output, append, and two-way (coprocess) sockets.  If @code{filetype}
is two-way, @command{gawk} assumes that it is a socket!  Note that in
the two-way case, the input and output file descriptors may differ.
To check for success, you must check whether either matches.
@end table

It is anticipated that this API function will be used to implement I/O
multiplexing and a socket library.

@node Extension API Variables
@subsection API Variables

The API provides two sets of variables.  The first provides information
about the version of the API (both with which the extension was compiled,
and with which @command{gawk} was compiled).  The second provides
information about how @command{gawk} was invoked.

@menu
* Extension Versioning::          API Version information.
* Extension GMP/MPFR Versioning:: Version information about GMP and MPFR.
* Extension API Informational Variables:: Variables providing information about
                                  @command{gawk}'s invocation.
@end menu

@node Extension Versioning
@subsubsection API Version Constants and Variables
@cindex API version
@cindex extension API version

The API provides both a ``major'' and a ``minor'' version number.
The API versions are available at compile time as C preprocessor defines
to support conditional compilation, and as enum constants to facilitate
debugging:

@float Table,gawk-api-version
@caption{gawk API version constants}
@multitable {@b{API Version}} {@code{gawk_api_major_version}} {@code{GAWK_API_MAJOR_VERSION}}
@headitem API Version @tab C Preprocessor Define @tab enum constant
@item Major @tab @code{gawk_api_major_version} @tab @code{GAWK_API_MAJOR_VERSION}
@item Minor @tab @code{gawk_api_minor_version} @tab @code{GAWK_API_MINOR_VERSION}
@end multitable
@end float

The minor version increases when new functions are added to the API. Such
new functions are always added to the end of the API @code{struct}.

The major version increases (and the minor version is reset to zero) if any
of the data types change size or member order, or if any of the existing
functions change signature.

It could happen that an extension may be compiled against one version
of the API but loaded by a version of @command{gawk} using a different
version. For this reason, the major and minor API versions of the
running @command{gawk} are included in the API @code{struct} as read-only
constant integers:

@table @code
@item api->major_version
The major version of the running @command{gawk}.

@item api->minor_version
The minor version of the running @command{gawk}.
@end table

It is up to the extension to decide if there are API incompatibilities.
Typically, a check like this is enough:

@example
if (   api->major_version != GAWK_API_MAJOR_VERSION
    || api->minor_version < GAWK_API_MINOR_VERSION) @{
        fprintf(stderr, "foo_extension: version mismatch with gawk!\n");
        fprintf(stderr, "\tmy version (%d, %d), gawk version (%d, %d)\n",
                GAWK_API_MAJOR_VERSION, GAWK_API_MINOR_VERSION,
                api->major_version, api->minor_version);
        exit(1);
@}
@end example

Such code is included in the boilerplate @code{dl_load_func()} macro
provided in @file{gawkapi.h} (discussed in
@ref{Extension API Boilerplate}).

@node Extension GMP/MPFR Versioning
@subsubsection GMP and MPFR Version Information

The API also includes information about the versions of GMP and MPFR
with which the running @command{gawk} was compiled (if any).
They are included in the API @code{struct} as read-only
constant integers:

@table @code
@item api->gmp_major_version
The major version of the GMP library used to compile @command{gawk}.

@item api->gmp_minor_version
The minor version of the GMP library used to compile @command{gawk}.

@item api->mpfr_major_version
The major version of the MPFR library used to compile @command{gawk}.

@item api->mpfr_minor_version
The minor version of the MPFR library used to compile @command{gawk}.
@end table

These fields are set to zero if @command{gawk} was compiled without
MPFR support.

You can check if the versions of MPFR and GMP that you are using match those
of @command{gawk} with the following macro:

@table @code
@item check_mpfr_version(extension)
The @code{extension} is the extension id passed to all the other macros
and functions defined in @file{gawkapi.h}.  If you have not included
the @code{<mpfr.h>} header file, then this macro will be defined to do nothing.

If you have included that file, then this macro compares the MPFR
and GMP major and minor versions against those of the library you are
compiling against.  If your libraries are newer than @command{gawk}'s, it
produces a fatal error message.

The @code{dl_load_func()} macro (@pxref{Extension API Boilerplate})
calls @code{check_mpfr_version()}.
@end table

@node Extension API Informational Variables
@subsubsection Informational Variables
@cindex API informational variables
@cindex extension API informational variables

The API provides access to several variables that describe
whether the corresponding command-line options were enabled when
@command{gawk} was invoked.  The variables are:

@table @code
@item do_debug
This variable is true if @command{gawk} was invoked with @option{--debug} option.

@item do_lint
This variable is true if @command{gawk} was invoked with @option{--lint} option.

@item do_mpfr
This variable is true if @command{gawk} was invoked with @option{--bignum} option.

@item do_profile
This variable is true if @command{gawk} was invoked with @option{--profile} option.

@item do_sandbox
This variable is true if @command{gawk} was invoked with @option{--sandbox} option.

@item do_traditional
This variable is true if @command{gawk} was invoked with @option{--traditional} option.
@end table

The value of @code{do_lint} can change if @command{awk} code
modifies the @code{LINT} predefined variable (@pxref{Built-in Variables}).
The others should not change during execution.

@node Extension API Boilerplate
@subsection Boilerplate Code

As mentioned earlier (@pxref{Extension Mechanism Outline}), the function
definitions as presented are really macros. To use these macros, your
extension must provide a small amount of boilerplate code (variables and
functions) toward the top of your source file, using predefined names
as described here.  The boilerplate needed is also provided in comments
in the @file{gawkapi.h} header file:

@example
@group
/* Boilerplate code: */
int plugin_is_GPL_compatible;

static gawk_api_t *const api;
@end group
static awk_ext_id_t ext_id;
static const char *ext_version = NULL; /* or @dots{} = "some string" */

static awk_ext_func_t func_table[] = @{
    @{ "name", do_name, 1, 0, awk_false, NULL @},
    /* @dots{} */
@};

/* EITHER: */

static awk_bool_t (*init_func)(void) = NULL;

/* OR: */

static awk_bool_t
init_my_extension(void)
@{
    @dots{}
@}

static awk_bool_t (*init_func)(void) = init_my_extension;

dl_load_func(func_table, some_name, "name_space_in_quotes")
@end example

These variables and functions are as follows:

@table @code
@item int plugin_is_GPL_compatible;
This asserts that the extension is compatible with
@ifclear FOR_PRINT
the GNU GPL (@pxref{Copying}).
@end ifclear
@ifset FOR_PRINT
the GNU GPL.
@end ifset
If your extension does not have this, @command{gawk}
will not load it (@pxref{Plugin License}).

@item static gawk_api_t *const api;
This global @code{static} variable should be set to point to
the @code{gawk_api_t} pointer that @command{gawk} passes to your
@code{dl_load()} function.  This variable is used by all of the macros.

@item static awk_ext_id_t ext_id;
This global static variable should be set to the @code{awk_ext_id_t}
value that @command{gawk} passes to your @code{dl_load()} function.
This variable is used by all of the macros.

@item static const char *ext_version = NULL; /* or @dots{} = "some string" */
This global @code{static} variable should be set either
to @code{NULL}, or to point to a string giving the name and version of
your extension.

@item static awk_ext_func_t func_table[] = @{ @dots{} @};
This is an array of one or more @code{awk_ext_func_t} structures,
as described earlier (@pxref{Extension Functions}).
It can then be looped over for multiple calls to
@code{add_ext_func()}.

@c Use @var{OR} for docbook
@item static awk_bool_t (*init_func)(void) = NULL;
@itemx @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @var{OR}
@itemx static awk_bool_t init_my_extension(void) @{ @dots{} @}
@itemx static awk_bool_t (*init_func)(void) = init_my_extension;
If you need to do some initialization work, you should define a
function that does it (creates variables, opens files, etc.)
and then define the @code{init_func} pointer to point to your
function.
The function should return @code{awk_false} upon failure, or @code{awk_true}
if everything goes well.

If you don't need to do any initialization, define the pointer and
initialize it to @code{NULL}.

@item dl_load_func(func_table, some_name, "name_space_in_quotes")
This macro expands to a @code{dl_load()} function that performs
all the necessary initializations.
@end table

The point of all the variables and arrays is to let the
@code{dl_load()} function (from the @code{dl_load_func()}
macro) do all the standard work. It does the following:

@enumerate 1
@item
Check the API versions. If the extension major version does not match
@command{gawk}'s, or if the extension minor version is greater than
@command{gawk}'s, it prints a fatal error message and exits.

@item
Check the MPFR and GMP versions. If there is a mismatch, it prints
a fatal error message and exits.

@item
Load the functions defined in @code{func_table}.
If any of them fails to load, it prints a warning message but
continues on.

@item
If the @code{init_func} pointer is not @code{NULL}, call the
function it points to. If it returns @code{awk_false}, print a
warning message.

@item
If @code{ext_version} is not @code{NULL}, register
the version string with @command{gawk}.
@end enumerate


@node Changes from API V1
@subsection Changes From Version 1 of the API

The current API is @emph{not} binary compatible with version 1 of the API.
You will have to recompile your extensions in order to use them with
the current version of @command{gawk}.

Fortunately, at the possible expense of some compile-time warnings, the API remains
source-code--compatible with the previous API. The major differences are
the additional members in the @code{awk_ext_func_t} structure, and the
addition of the third argument to the C implementation function
(@pxref{Extension Functions}).

Here is a list of individual features that changed from version 1 to
version 2 of the API:

@itemize @bullet

@item
Numeric values can now have MPFR/MPZ variants
(@pxref{General Data Types}).

@item
There are new string types: @code{AWK_REGEX} and @code{AWK_STRNUM}
(@pxref{General Data Types}).

@item
The @code{ezalloc()} macro is new
(@pxref{Memory Allocation Functions}).

@item
The @code{awk_ext_func_t} structure changed. Instead of
@code{num_expected_args}, it now has @code{max_expected} and
@code{min_required}
(@pxref{Extension Functions}).

@item
For @code{get_record()}, an input parser can now specify field widths
(@pxref{Input Parsers}).

@item
Extensions can now produce nonfatal error messages
(@pxref{Printing Messages}).

@item
When flattening an array, you can now specify the index and value types
(@pxref{Array Functions}).

@item
The @code{get_file()} API is new
(@pxref{Redirection API}).
@end itemize

@node Finding Extensions
@section How @command{gawk} Finds Extensions
@cindex extension search path
@cindex finding extensions

Compiled extensions have to be installed in a directory where
@command{gawk} can find them.  If @command{gawk} is configured and
built in the default fashion, the directory in which to find
extensions is @file{/usr/local/lib/gawk}.  You can also specify a search
path with a list of directories to search for compiled extensions.
@xref{AWKLIBPATH Variable} for more information.

@node Extension Example
@section Example: Some File Functions
@cindex extension example

@quotation
@i{No matter where you go, there you are.}
@author Buckaroo Banzai
@end quotation

@c It's enough to show chdir and stat, no need for fts

Two useful functions that are not in @command{awk} are @code{chdir()} (so
that an @command{awk} program can change its directory) and @code{stat()}
(so that an @command{awk} program can gather information about a file).
In order to illustrate the API in action, this @value{SECTION} implements
these functions for @command{gawk} in an extension.

@menu
* Internal File Description::   What the new functions will do.
* Internal File Ops::           The code for internal file operations.
* Using Internal File Ops::     How to use an external extension.
@end menu

@node Internal File Description
@subsection Using @code{chdir()} and @code{stat()}

This @value{SECTION} shows how to use the new functions at
the @command{awk} level once they've been integrated into the
running @command{gawk} interpreter.  Using @code{chdir()} is very
straightforward. It takes one argument, the new directory to change to:

@example
@@load "filefuncs"
@dots{}
newdir = "/home/arnold/funstuff"
ret = chdir(newdir)
if (ret < 0) @{
    printf("could not change to %s: %s\n", newdir, ERRNO) > "/dev/stderr"
    exit 1
@}
@dots{}
@end example

The return value is negative if the @code{chdir()} failed, and
@code{ERRNO} (@pxref{Built-in Variables}) is set to a string indicating
the error.

Using @code{stat()} is a bit more complicated.  The C @code{stat()}
function fills in a structure that has a fair amount of information.
The right way to model this in @command{awk} is to fill in an associative
array with the appropriate information:

@c broke printf for page breaking
@example
file = "/home/arnold/.profile"
ret = stat(file, fdata)
if (ret < 0) @{
    printf("could not stat %s: %s\n",
             file, ERRNO) > "/dev/stderr"
    exit 1
@}
printf("size of %s is %d bytes\n", file, fdata["size"])
@end example

The @code{stat()} function always clears the data array, even if
the @code{stat()} fails.  It fills in the following elements:

@table @code
@item "name"
The name of the file that was @code{stat()}ed.

@item "dev"
@itemx "ino"
The file's device and inode numbers, respectively.

@item "mode"
The file's mode, as a numeric value. This includes both the file's
type and its permissions.

@item "nlink"
The number of hard links (directory entries) the file has.

@item "uid"
@itemx "gid"
The numeric user and group ID numbers of the file's owner.

@item "size"
The size in bytes of the file.

@item "blocks"
The number of disk blocks the file actually occupies. This may not
be a function of the file's size if the file has holes.

@item "atime"
@itemx "mtime"
@itemx "ctime"
The file's last access, modification, and inode update times,
respectively.  These are numeric timestamps, suitable for formatting
with @code{strftime()}
(@pxref{Time Functions}).

@item "pmode"
The file's ``printable mode.''  This is a string representation of
the file's type and permissions, such as is produced by
@samp{ls -l}---for example, @code{"drwxr-xr-x"}.

@item "type"
A printable string representation of the file's type.  The value
is one of the following:

@table @code
@item "blockdev"
@itemx "chardev"
The file is a block or character device (``special file'').

@ignore
@item "door"
The file is a Solaris ``door'' (special file used for
interprocess communications).
@end ignore

@item "directory"
The file is a directory.

@item "fifo"
The file is a named pipe (also known as a FIFO).

@item "file"
The file is just a regular file.

@item "socket"
The file is an @code{AF_UNIX} (``Unix domain'') socket in the
filesystem.

@item "symlink"
The file is a symbolic link.
@end table

@c 5/2013: Thanks to Corinna Vinschen for this information.
@item "devbsize"
The size of a block for the element indexed by @code{"blocks"}.
This information is derived from either the @code{DEV_BSIZE}
constant defined in @code{<sys/param.h>} on most systems,
or the @code{S_BLKSIZE} constant in @code{<sys/stat.h>} on BSD systems.
For some other systems, @dfn{a priori} knowledge is used to provide
a value. Where no value can be determined, it defaults to 512.
@end table

Several additional elements may be present, depending upon the operating
system and the type of the file.  You can test for them in your @command{awk}
program by using the @code{in} operator
(@pxref{Reference to Elements}):

@table @code
@item "blksize"
The preferred block size for I/O to the file. This field is not
present on all POSIX-like systems in the C @code{stat} structure.

@item "linkval"
If the file is a symbolic link, this element is the name of the
file the link points to (i.e., the value of the link).

@item "rdev"
@itemx "major"
@itemx "minor"
If the file is a block or character device file, then these values
represent the numeric device number and the major and minor components
of that number, respectively.
@end table

@node Internal File Ops
@subsection C Code for @code{chdir()} and @code{stat()}

Here is the C code for these extensions.@footnote{This version is
edited slightly for presentation.  See @file{extension/filefuncs.c}
in the @command{gawk} distribution for the complete version.}

The file includes a number of standard header files, and then includes
the @file{gawkapi.h} header file, which provides the API definitions.
Those are followed by the necessary variable declarations
to make use of the API macros and boilerplate code
(@pxref{Extension API Boilerplate}):

@example
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <stdio.h>
#include <assert.h>
#include <errno.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>

#include <sys/types.h>
#include <sys/stat.h>

#include "gawkapi.h"

#include "gettext.h"
#define _(msgid)  gettext(msgid)
#define N_(msgid) msgid

#include "gawkfts.h"
#include "stack.h"

static const gawk_api_t *api;    /* for convenience macros to work */
static awk_ext_id_t ext_id;
static awk_bool_t init_filefuncs(void);
static awk_bool_t (*init_func)(void) = init_filefuncs;
static const char *ext_version = "filefuncs extension: version 1.0";

int plugin_is_GPL_compatible;
@end example

@cindex programming conventions, @command{gawk} extensions
By convention, for an @command{awk} function @code{foo()}, the C function
that implements it is called @code{do_foo()}.  The function should have
two arguments. The first is an @code{int}, usually called @code{nargs},
that represents the number of actual arguments for the function.
The second is a pointer to an @code{awk_value_t} structure, usually named
@code{result}:

@example
@group
/*  do_chdir --- provide dynamically loaded chdir() function for gawk */

static awk_value_t *
do_chdir(int nargs, awk_value_t *result, struct awk_ext_func *unused)
@end group
@{
    awk_value_t newdir;
    int ret = -1;

    assert(result != NULL);
@end example

The @code{newdir}
variable represents the new directory to change to, which is retrieved
with @code{get_argument()}.  Note that the first argument is
numbered zero.

If the argument is retrieved successfully, the function calls the
@code{chdir()} system call. Otherwise, if the @code{chdir()} fails,
it updates @code{ERRNO}:

@example
    if (get_argument(0, AWK_STRING, & newdir)) @{
        ret = chdir(newdir.str_value.str);
        if (ret < 0)
            update_ERRNO_int(errno);
    @}
@end example

Finally, the function returns the return value to the @command{awk} level:

@example
    return make_number(ret, result);
@}
@end example

The @code{stat()} extension is more involved.  First comes a function
that turns a numeric mode into a printable representation
(e.g., octal @code{0644} becomes @samp{-rw-r--r--}). This is omitted here for brevity:

@example
/* format_mode --- turn a stat mode field into something readable */

static char *
format_mode(unsigned long fmode)
@{
    @dots{}
@}
@end example

Next comes a function for reading symbolic links, which is also
omitted here for brevity:

@example
/* read_symlink --- read a symbolic link into an allocated buffer.
   @dots{} */

static char *
read_symlink(const char *fname, size_t bufsize, ssize_t *linksize)
@{
    @dots{}
@}
@end example

Two helper functions simplify entering values in the
array that will contain the result of the @code{stat()}:

@example
/* array_set --- set an array element */

static void
array_set(awk_array_t array, const char *sub, awk_value_t *value)
@{
    awk_value_t index;

    set_array_element(array,
                      make_const_string(sub, strlen(sub), & index),
                      value);

@}

/* array_set_numeric --- set an array element with a number */

static void
array_set_numeric(awk_array_t array, const char *sub, double num)
@{
    awk_value_t tmp;

    array_set(array, sub, make_number(num, & tmp));
@}
@end example

The following function does most of the work to fill in
the @code{awk_array_t} result array with values obtained
from a valid @code{struct stat}. This work is done in a separate function
to support the @code{stat()} function for @command{gawk} and also
to support the @code{fts()} extension, which is included in
the same file but whose code is not shown here
(@pxref{Extension Sample File Functions}).

The first part of the function is variable declarations,
including a table to map file types to strings:

@example
/* fill_stat_array --- do the work to fill an array with stat info */

static int
fill_stat_array(const char *name, awk_array_t array, struct stat *sbuf)
@{
    char *pmode;    /* printable mode */
    const char *type = "unknown";
    awk_value_t tmp;
    static struct ftype_map @{
        unsigned int mask;
        const char *type;
    @} ftype_map[] = @{
        @{ S_IFREG, "file" @},
        @{ S_IFBLK, "blockdev" @},
        @{ S_IFCHR, "chardev" @},
        @{ S_IFDIR, "directory" @},
#ifdef S_IFSOCK
        @{ S_IFSOCK, "socket" @},
#endif
#ifdef S_IFIFO
        @{ S_IFIFO, "fifo" @},
#endif
#ifdef S_IFLNK
        @{ S_IFLNK, "symlink" @},
#endif
#ifdef S_IFDOOR /* Solaris weirdness */
        @{ S_IFDOOR, "door" @},
#endif
    @};
    int j, k;
@end example

The destination array is cleared, and then code fills in
various elements based on values in the @code{struct stat}:

@example
    /* empty out the array */
    clear_array(array);

    /* fill in the array */
    array_set(array, "name", make_const_string(name, strlen(name),
                                               & tmp));
    array_set_numeric(array, "dev", sbuf->st_dev);
    array_set_numeric(array, "ino", sbuf->st_ino);
    array_set_numeric(array, "mode", sbuf->st_mode);
    array_set_numeric(array, "nlink", sbuf->st_nlink);
    array_set_numeric(array, "uid", sbuf->st_uid);
    array_set_numeric(array, "gid", sbuf->st_gid);
    array_set_numeric(array, "size", sbuf->st_size);
    array_set_numeric(array, "blocks", sbuf->st_blocks);
    array_set_numeric(array, "atime", sbuf->st_atime);
    array_set_numeric(array, "mtime", sbuf->st_mtime);
    array_set_numeric(array, "ctime", sbuf->st_ctime);

    /* for block and character devices, add rdev,
       major and minor numbers */
    if (S_ISBLK(sbuf->st_mode) || S_ISCHR(sbuf->st_mode)) @{
        array_set_numeric(array, "rdev", sbuf->st_rdev);
        array_set_numeric(array, "major", major(sbuf->st_rdev));
        array_set_numeric(array, "minor", minor(sbuf->st_rdev));
    @}
@end example

@noindent
The latter part of the function makes selective additions
to the destination array, depending upon the availability of
certain members and/or the type of the file. It then returns zero,
for success:

@example
@group
#ifdef HAVE_STRUCT_STAT_ST_BLKSIZE
    array_set_numeric(array, "blksize", sbuf->st_blksize);
#endif
@end group

    pmode = format_mode(sbuf->st_mode);
    array_set(array, "pmode", make_const_string(pmode, strlen(pmode),
                                                & tmp));

    /* for symbolic links, add a linkval field */
    if (S_ISLNK(sbuf->st_mode)) @{
        char *buf;
        ssize_t linksize;

        if ((buf = read_symlink(name, sbuf->st_size,
                    & linksize)) != NULL)
            array_set(array, "linkval",
                      make_malloced_string(buf, linksize, & tmp));
        else
            warning(ext_id, _("stat: unable to read symbolic link `%s'"),
                    name);
    @}

    /* add a type field */
    type = "unknown";   /* shouldn't happen */
    for (j = 0, k = sizeof(ftype_map)/sizeof(ftype_map[0]); j < k; j++) @{
        if ((sbuf->st_mode & S_IFMT) == ftype_map[j].mask) @{
            type = ftype_map[j].type;
            break;
        @}
    @}

    array_set(array, "type", make_const_string(type, strlen(type), & tmp));

    return 0;
@}
@end example

The third argument to @code{stat()} was not discussed previously. This
argument is optional. If present, it causes @code{do_stat()} to use
the @code{stat()} system call instead of the @code{lstat()} system
call.  This is done by using a function pointer: @code{statfunc}.
@code{statfunc} is initialized to point to @code{lstat()} (instead
of @code{stat()}) to get the file information, in case the file is a
symbolic link. However, if the third argument is included, @code{statfunc}
is set to point to @code{stat()}, instead.

Here is the @code{do_stat()} function, which starts with
variable declarations and argument checking:

@example
/* do_stat --- provide a stat() function for gawk */

static awk_value_t *
do_stat(int nargs, awk_value_t *result, struct awk_ext_func *unused)
@{
    awk_value_t file_param, array_param;
    char *name;
    awk_array_t array;
    int ret;
    struct stat sbuf;
    /* default is lstat() */
    int (*statfunc)(const char *path, struct stat *sbuf) = lstat;

    assert(result != NULL);
@end example

Then comes the actual work. First, the function gets the arguments.
Next, it gets the information for the file.  If the called function
(@code{lstat()} or @code{stat()}) returns an error, the code sets
@code{ERRNO} and returns:

@example
    /* file is first arg, array to hold results is second */
    if (   ! get_argument(0, AWK_STRING, & file_param)
        || ! get_argument(1, AWK_ARRAY, & array_param)) @{
        warning(ext_id, _("stat: bad parameters"));
        return make_number(-1, result);
    @}

    if (nargs == 3) @{
        statfunc = stat;
    @}

    name = file_param.str_value.str;
    array = array_param.array_cookie;

    /* always empty out the array */
    clear_array(array);

    /* stat the file; if error, set ERRNO and return */
    ret = statfunc(name, & sbuf);
@group
    if (ret < 0) @{
        update_ERRNO_int(errno);
        return make_number(ret, result);
    @}
@end group
@end example

The tedious work is done by @code{fill_stat_array()}, shown
earlier.  When done, the function returns the result from @code{fill_stat_array()}:

@example
@group
    ret = fill_stat_array(name, array, & sbuf);

    return make_number(ret, result);
@}
@end group
@end example

Finally, it's necessary to provide the ``glue'' that loads the
new function(s) into @command{gawk}.

The @code{filefuncs} extension also provides an @code{fts()}
function, which we omit here
(@pxref{Extension Sample File Functions}).
For its sake, there is an initialization
function:

@example
/* init_filefuncs --- initialization routine */

static awk_bool_t
init_filefuncs(void)
@{
    @dots{}
@}
@end example

We are almost done. We need an array of @code{awk_ext_func_t}
structures for loading each function into @command{gawk}:

@example
static awk_ext_func_t func_table[] = @{
    @{ "chdir", do_chdir, 1, 1, awk_false, NULL @},
    @{ "stat",  do_stat, 3, 2, awk_false, NULL @},
    @dots{}
@};
@end example

Each extension must have a routine named @code{dl_load()} to load
everything that needs to be loaded.  It is simplest to use the
@code{dl_load_func()} macro in @code{gawkapi.h}:

@example
/* define the dl_load() function using the boilerplate macro */

dl_load_func(func_table, filefuncs, "")
@end example

And that's it!

@node Using Internal File Ops
@subsection Integrating the Extensions

@cindex @command{gawk}, interpreter@comma{} adding code to
Now that the code is written, it must be possible to add it at
runtime to the running @command{gawk} interpreter.  First, the
code must be compiled.  Assuming that the functions are in
a file named @file{filefuncs.c}, and @var{idir} is the location
of the @file{gawkapi.h} header file,
the following steps@footnote{In practice, you would probably want to
use the GNU Autotools (Automake, Autoconf, Libtool, and @command{gettext}) to
configure and build your libraries. Instructions for doing so are beyond
the scope of this @value{DOCUMENT}. @xref{gawkextlib} for Internet links to
the tools.} create a GNU/Linux shared library:

@example
$ @kbd{gcc -fPIC -shared -DHAVE_CONFIG_H -c -O -g -I@var{idir} filefuncs.c}
$ @kbd{gcc -o filefuncs.so -shared filefuncs.o}
@end example

Once the library exists, it is loaded by using the @code{@@load} keyword:

@example
# file testff.awk
@@load "filefuncs"

BEGIN @{
    "pwd" | getline curdir  # save current directory
    close("pwd")

    chdir("/tmp")
    system("pwd")   # test it
    chdir(curdir)   # go back

    print "Info for testff.awk"
    ret = stat("testff.awk", data)
    print "ret =", ret
    for (i in data)
        printf "data[\"%s\"] = %s\n", i, data[i]
    print "testff.awk modified:",
        strftime("%m %d %Y %H:%M:%S", data["mtime"])

    print "\nInfo for JUNK"
    ret = stat("JUNK", data)
    print "ret =", ret
    for (i in data)
        printf "data[\"%s\"] = %s\n", i, data[i]
    print "JUNK modified:", strftime("%m %d %Y %H:%M:%S", data["mtime"])
@}
@end example

The @env{AWKLIBPATH} environment variable tells
@command{gawk} where to find extensions (@pxref{Finding Extensions}).
We set it to the current directory and run the program:

@example
$ @kbd{AWKLIBPATH=$PWD gawk -f testff.awk}
@print{} /tmp
@print{} Info for testff.awk
@print{} ret = 0
@print{} data["blksize"] = 4096
@print{} data["devbsize"] = 512
@print{} data["mtime"] = 1412004710
@print{} data["mode"] = 33204
@print{} data["type"] = file
@print{} data["dev"] = 2053
@print{} data["gid"] = 1000
@print{} data["ino"] = 10358899
@print{} data["ctime"] = 1412004710
@print{} data["blocks"] = 8
@print{} data["nlink"] = 1
@print{} data["name"] = testff.awk
@print{} data["atime"] = 1412004716
@print{} data["pmode"] = -rw-rw-r--
@print{} data["size"] = 666
@print{} data["uid"] = 1000
@print{} testff.awk modified: 09 29 2014 18:31:50
@print{}
@print{} Info for JUNK
@print{} ret = -1
@print{} JUNK modified: 01 01 1970 02:00:00
@end example

@node Extension Samples
@section The Sample Extensions in the @command{gawk} Distribution
@cindex extensions distributed with @command{gawk}

This @value{SECTION} provides a brief overview of the sample extensions
that come in the @command{gawk} distribution. Some of them are intended
for production use (e.g., the @code{filefuncs}, @code{readdir}, and
@code{inplace} extensions).  Others mainly provide example code that
shows how to use the extension API.

@menu
* Extension Sample File Functions::   The file functions sample.
* Extension Sample Fnmatch::          An interface to @code{fnmatch()}.
* Extension Sample Fork::             An interface to @code{fork()} and other
                                      process functions.
* Extension Sample Inplace::          Enabling in-place file editing.
* Extension Sample Ord::              Character to value to character
                                      conversions.
* Extension Sample Readdir::          An interface to @code{readdir()}.
* Extension Sample Revout::           Reversing output sample output wrapper.
* Extension Sample Rev2way::          Reversing data sample two-way processor.
* Extension Sample Read write array:: Serializing an array to a file.
* Extension Sample Readfile::         Reading an entire file into a string.
* Extension Sample Time::             An interface to @code{gettimeofday()}
                                      and @code{sleep()}.
* Extension Sample API Tests::        Tests for the API.
@end menu

@node Extension Sample File Functions
@subsection File-Related Functions

The @code{filefuncs} extension provides three different functions, as follows.
The usage is:

@table @asis
@item @code{@@load "filefuncs"}
This is how you load the extension.

@cindex @code{chdir()} extension function
@item @code{result = chdir("/some/directory")}
The @code{chdir()} function is a direct hook to the @code{chdir()}
system call to change the current directory.  It returns zero
upon success or a value less than zero upon error.
In the latter case, it updates @code{ERRNO}.

@cindex @code{stat()} extension function
@item @code{result = stat("/some/path", statdata} [@code{, follow}]@code{)}
The @code{stat()} function provides a hook into the
@code{stat()} system call.
It returns zero upon success or a value less than zero upon error.
In the latter case, it updates @code{ERRNO}.

By default, it uses the @code{lstat()} system call.  However, if passed
a third argument, it uses @code{stat()} instead.

In all cases, it clears the @code{statdata} array.
When the call is successful, @code{stat()} fills the @code{statdata}
array with information retrieved from the filesystem, as follows:

@multitable @columnfractions .15 .50 .20
@headitem Subscript @tab Field in @code{struct stat} @tab File type
@item @code{"name"} @tab The @value{FN} @tab All
@item @code{"dev"} @tab @code{st_dev} @tab All
@item @code{"ino"} @tab @code{st_ino} @tab All
@item @code{"mode"} @tab @code{st_mode} @tab All
@item @code{"nlink"} @tab @code{st_nlink} @tab All
@item @code{"uid"} @tab @code{st_uid} @tab All
@item @code{"gid"} @tab @code{st_gid} @tab All
@item @code{"size"} @tab @code{st_size} @tab All
@item @code{"atime"} @tab @code{st_atime} @tab All
@item @code{"mtime"} @tab @code{st_mtime} @tab All
@item @code{"ctime"} @tab @code{st_ctime} @tab All
@item @code{"rdev"} @tab @code{st_rdev} @tab Device files
@item @code{"major"} @tab @code{st_major} @tab Device files
@item @code{"minor"} @tab @code{st_minor} @tab Device files
@item @code{"blksize"} @tab @code{st_blksize} @tab All
@item @code{"pmode"} @tab A human-readable version of the mode value, like that printed by
@command{ls} (for example, @code{"-rwxr-xr-x"}) @tab All
@item @code{"linkval"} @tab The value of the symbolic link @tab Symbolic links
@item @code{"type"} @tab The type of the file as a string---one of
@code{"file"},
@code{"blockdev"},
@code{"chardev"},
@code{"directory"},
@code{"socket"},
@code{"fifo"},
@code{"symlink"},
@code{"door"},
or
@code{"unknown"}
(not all systems support all file types) @tab All
@end multitable

@cindex @code{fts()} extension function
@item @code{flags = or(FTS_PHYSICAL, ...)}
@itemx @code{result = fts(pathlist, flags, filedata)}
Walk the file trees provided in @code{pathlist} and fill in the
@code{filedata} array, as described next.  @code{flags} is the bitwise
OR of several predefined values, also described in a moment.
Return zero if there were no errors, otherwise return @minus{}1.
@end table

The @code{fts()} function provides a hook to the C library @code{fts()}
routines for traversing file hierarchies.  Instead of returning data
about one file at a time in a stream, it fills in a multidimensional
array with data about each file and directory encountered in the requested
hierarchies.

The arguments are as follows:

@table @code
@item pathlist
An array of @value{FN}s.  The element values are used; the index values are ignored.

@item flags
This should be the bitwise OR of one or more of the following
predefined constant flag values.  At least one of
@code{FTS_LOGICAL} or @code{FTS_PHYSICAL} must be provided; otherwise
@code{fts()} returns an error value and sets @code{ERRNO}.
The flags are:

@c nested table
@table @code
@item FTS_LOGICAL
Do a ``logical'' file traversal, where the information returned for
a symbolic link refers to the linked-to file, and not to the symbolic
link itself.  This flag is mutually exclusive with @code{FTS_PHYSICAL}.

@item FTS_PHYSICAL
Do a ``physical'' file traversal, where the information returned for a
symbolic link refers to the symbolic link itself.  This flag is mutually
exclusive with @code{FTS_LOGICAL}.

@item FTS_NOCHDIR
As a performance optimization, the C library @code{fts()} routines
change directory as they traverse a file hierarchy.  This flag disables
that optimization.

@item FTS_COMFOLLOW
Immediately follow a symbolic link named in @code{pathlist},
whether or not @code{FTS_LOGICAL} is set.

@item FTS_SEEDOT
By default, the C library @code{fts()} routines do not return entries for
@file{.} (dot) and @file{..} (dot-dot).  This option causes entries for
dot-dot to also be included.  (The extension always includes an entry
for dot; more on this in a moment.)

@item FTS_XDEV
During a traversal, do not cross onto a different mounted filesystem.
@end table

@item filedata
The @code{filedata} array holds the results.
@code{fts()} first clears it.  Then it creates
an element in @code{filedata} for every element in @code{pathlist}.
The index is the name of the directory or file given in @code{pathlist}.
The element for this index is itself an array.  There are two cases:

@c nested table
@table @emph
@item The path is a file
In this case, the array contains two or three elements:

@c doubly nested table
@table @code
@item "path"
The full path to this file, starting from the ``root'' that was given
in the @code{pathlist} array.

@item "stat"
This element is itself an array, containing the same information as provided
by the @code{stat()} function described earlier for its
@code{statdata} argument.  The element may not be present if
the @code{stat()} system call for the file failed.

@item "error"
If some kind of error was encountered, the array will also
contain an element named @code{"error"}, which is a string describing the error.
@end table

@item The path is a directory
In this case, the array contains one element for each entry in the
directory.  If an entry is a file, that element is the same as for files, just
described.  If the entry is a directory, that element is (recursively)
an array describing the subdirectory.  If @code{FTS_SEEDOT} was provided
in the flags, then there will also be an element named @code{".."}.  This
element will be an array containing the data as provided by @code{stat()}.

In addition, there will be an element whose index is @code{"."}.
This element is an array containing the same two or three elements as
for a file: @code{"path"}, @code{"stat"}, and @code{"error"}.
@end table
@end table

The @code{fts()} function returns zero if there were no errors.
Otherwise, it returns @minus{}1.

@quotation NOTE
The @code{fts()} extension does not exactly mimic the
interface of the C library @code{fts()} routines, choosing instead to
provide an interface that is based on associative arrays, which is
more comfortable to use from an @command{awk} program.  This includes the
lack of a comparison function, because @command{gawk} already provides
powerful array sorting facilities.  Although an @code{fts_read()}-like
interface could have been provided, this felt less natural than simply
creating a multidimensional array to represent the file hierarchy and
its information.
@end quotation

See @file{test/fts.awk} in the @command{gawk} distribution for an example
use of the @code{fts()} extension function.

@node Extension Sample Fnmatch
@subsection Interface to @code{fnmatch()}

This extension provides an interface to the C library
@code{fnmatch()} function.  The usage is:

@table @code
@item @@load "fnmatch"
This is how you load the extension.

@cindex @code{fnmatch()} extension function
@item result = fnmatch(pattern, string, flags)
The return value is zero on success, @code{FNM_NOMATCH}
if the string did not match the pattern, or
a different nonzero value if an error occurred.
@end table

In addition to the @code{fnmatch()} function, the @code{fnmatch} extension
adds one constant (@code{FNM_NOMATCH}), and an array of flag values
named @code{FNM}.

The arguments to @code{fnmatch()} are:

@table @code
@item pattern
The @value{FN} wildcard to match

@item string
The @value{FN} string

@item flag
Either zero, or the bitwise OR of one or more of the
flags in the @code{FNM} array
@end table

The flags are as follows:

@multitable @columnfractions .25 .75
@headitem Array element @tab Corresponding flag defined by @code{fnmatch()}
@item @code{FNM["CASEFOLD"]} @tab @code{FNM_CASEFOLD}
@item @code{FNM["FILE_NAME"]} @tab @code{FNM_FILE_NAME}
@item @code{FNM["LEADING_DIR"]} @tab @code{FNM_LEADING_DIR}
@item @code{FNM["NOESCAPE"]} @tab @code{FNM_NOESCAPE}
@item @code{FNM["PATHNAME"]} @tab @code{FNM_PATHNAME}
@item @code{FNM["PERIOD"]} @tab @code{FNM_PERIOD}
@end multitable

Here is an example:

@example
@@load "fnmatch"
@dots{}
flags = or(FNM["PERIOD"], FNM["NOESCAPE"])
if (fnmatch("*.a", "foo.c", flags) == FNM_NOMATCH)
    print "no match"
@end example

@node Extension Sample Fork
@subsection Interface to @code{fork()}, @code{wait()}, and @code{waitpid()}

The @code{fork} extension adds three functions, as follows:

@table @code
@item @@load "fork"
This is how you load the extension.

@cindex @code{fork()} extension function
@item pid = fork()
This function creates a new process. The return value is zero in the
child and the process ID number of the child in the parent, or @minus{}1
upon error. In the latter case, @code{ERRNO} indicates the problem.
In the child, @code{PROCINFO["pid"]} and @code{PROCINFO["ppid"]} are
updated to reflect the correct values.

@cindex @code{waitpid()} extension function
@item ret = waitpid(pid)
This function takes a numeric argument, which is the process ID to
wait for. The return value is that of the
@code{waitpid()} system call.

@cindex @code{wait()} extension function
@item ret = wait()
This function waits for the first child to die.
The return value is that of the
@code{wait()} system call.
@end table

There is no corresponding @code{exec()} function.

Here is an example:

@example
@@load "fork"
@dots{}
if ((pid = fork()) == 0)
    print "hello from the child"
else
    print "hello from the parent"
@end example

@node Extension Sample Inplace
@subsection Enabling In-Place File Editing

@cindex @code{inplace} extension
The @code{inplace} extension emulates GNU @command{sed}'s @option{-i} option,
which performs ``in-place'' editing of each input file.
It uses the bundled @file{inplace.awk} include file to invoke the extension
properly:

@example
@c file eg/lib/inplace.awk
@group
# inplace --- load and invoke the inplace extension.
@c endfile
@ignore
@c file eg/lib/inplace.awk
# 
# Copyright (C) 2013, 2017 the Free Software Foundation, Inc.
# 
# This file is part of GAWK, the GNU implementation of the
# AWK Programming Language.
# 
# GAWK is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
# 
# GAWK is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
# 
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
#
# Andrew J. Schorr, aschorr@@telemetry-investments.com
# January 2013
@c endfile
@end ignore
@c file eg/lib/inplace.awk

@@load "inplace"

# Please set INPLACE_SUFFIX to make a backup copy.  For example, you may
# want to set INPLACE_SUFFIX to .bak on the command line or in a BEGIN rule.

# By default, each filename on the command line will be edited inplace.
# But you can selectively disable this by adding an inplace=0 argument
# prior to files that you do not want to process this way.  You can then
# reenable it later on the commandline by putting inplace=1 before files
# that you wish to be subject to inplace editing.

# N.B. We call inplace_end() in the BEGINFILE and END rules so that any
# actions in an ENDFILE rule will be redirected as expected.
@end group

@group
BEGIN @{
    inplace = 1         # enabled by default
@}
@end group

@group
BEGINFILE @{
    if (_inplace_filename != "")
        inplace_end(_inplace_filename, INPLACE_SUFFIX)
    if (inplace)
        inplace_begin(_inplace_filename = FILENAME, INPLACE_SUFFIX)
    else
        _inplace_filename = ""
@}
@end group

@group
END @{
    if (_inplace_filename != "")
        inplace_end(_inplace_filename, INPLACE_SUFFIX)
@}
@end group
@c endfile
@end example

For each regular file that is processed, the extension redirects
standard output to a temporary file configured to have the same owner
and permissions as the original.  After the file has been processed,
the extension restores standard output to its original destination.
If @code{INPLACE_SUFFIX} is not an empty string, the original file is
linked to a backup @value{FN} created by appending that suffix.  Finally,
the temporary file is renamed to the original @value{FN}.

Note that the use of this feature can be controlled by placing @samp{inplace=0}
on the command-line prior to listing files that should not be processed this
way.  You can reenable inplace editing by adding an @samp{inplace=1} argument
prior to files that should be subject to inplace editing.

The @code{_inplace_filename} variable serves to keep track of the
current filename so as to not invoke @code{inplace_end()} before
processing the first file.

If any error occurs, the extension issues a fatal error to terminate
processing immediately without damaging the original file.

Here are some simple examples:

@example
$ @kbd{gawk -i inplace '@{ gsub(/foo/, "bar") @}; @{ print @}' file1 file2 file3}
@end example

To keep a backup copy of the original files, try this:

@example
$ @kbd{gawk -i inplace -v INPLACE_SUFFIX=.bak '@{ gsub(/foo/, "bar") @}}
> @kbd{@{ print @}' file1 file2 file3}
@end example

Please note that, while the extension does attempt to preserve ownership and permissions, it makes no attempt to copy the ACLs from the original file.

If the program dies prematurely, as might happen if an unhandled signal is received, a temporary file may be left behind.

@node Extension Sample Ord
@subsection Character and Numeric values: @code{ord()} and @code{chr()}

The @code{ordchr} extension adds two functions, named
@code{ord()} and @code{chr()}, as follows:

@table @code
@item @@load "ordchr"
This is how you load the extension.

@cindex @code{ord()} extension function
@item number = ord(string)
Return the numeric value of the first character in @code{string}.

@cindex @code{chr()} extension function
@item char = chr(number)
Return a string whose first character is that represented by @code{number}.
@end table

These functions are inspired by the Pascal language functions
of the same name.  Here is an example:

@example
@@load "ordchr"
@dots{}
printf("The numeric value of 'A' is %d\n", ord("A"))
printf("The string value of 65 is %s\n", chr(65))
@end example

@node Extension Sample Readdir
@subsection Reading Directories

The @code{readdir} extension adds an input parser for directories.
The usage is as follows:

@cindex @code{readdir} extension
@example
@@load "readdir"
@end example

When this extension is in use, instead of skipping directories named
on the command line (or with @code{getline}),
they are read, with each entry returned as a record.

The record consists of three fields. The first two are the inode number and the
@value{FN}, separated by a forward slash character.
On systems where the directory entry contains the file type, the record
has a third field (also separated by a slash), which is a single letter
indicating the type of the file. The letters and their corresponding file
types are shown in @ref{table-readdir-file-types}.

@float Table,table-readdir-file-types
@caption{File types returned by the @code{readdir} extension}
@multitable @columnfractions .1 .9
@headitem Letter @tab File type
@item @code{b} @tab Block device
@item @code{c} @tab Character device
@item @code{d} @tab Directory
@item @code{f} @tab Regular file
@item @code{l} @tab Symbolic link
@item @code{p} @tab Named pipe (FIFO)
@item @code{s} @tab Socket
@item @code{u} @tab Anything else (unknown)
@end multitable
@end float

On systems without the file type information, the third field is always
@samp{u}.

@quotation NOTE
On GNU/Linux systems, there are filesystems that don't support the
@code{d_type} entry (see the @i{readdir}(3) manual page), and so the file
type is always @samp{u}.  You can use the @code{filefuncs} extension to call
@code{stat()} in order to get correct type information.
@end quotation

Here is an example:

@example
@@load "readdir"
@dots{}
BEGIN @{ FS = "/" @}
@{ print "@value{FN} is", $2 @}
@end example

@node Extension Sample Revout
@subsection Reversing Output

The @code{revoutput} extension adds a simple output wrapper that reverses
the characters in each output line.  Its main purpose is to show how to
write an output wrapper, although it may be mildly amusing for the unwary.
Here is an example:

@cindex @code{revoutput} extension
@example
@@load "revoutput"

BEGIN @{
    REVOUT = 1
    print "don't panic" > "/dev/stdout"
@}
@end example

The output from this program is @samp{cinap t'nod}.

@node Extension Sample Rev2way
@subsection Two-Way I/O Example

The @code{revtwoway} extension adds a simple two-way processor that
reverses the characters in each line sent to it for reading back by
the @command{awk} program.  Its main purpose is to show how to write
a two-way processor, although it may also be mildly amusing.
The following example shows how to use it:

@cindex @code{revtwoway} extension
@example
@@load "revtwoway"

BEGIN @{
    cmd = "/magic/mirror"
    print "don't panic" |& cmd
    cmd |& getline result
    print result
    close(cmd)
@}
@end example

The output from this program
@ifnotinfo
also is:
@end ifnotinfo
@ifinfo
is:
@end ifinfo
@samp{cinap t'nod}.

@node Extension Sample Read write array
@subsection Dumping and Restoring an Array

The @code{rwarray} extension adds two functions,
named @code{writea()} and @code{reada()}, as follows:

@table @code
@item @@load "rwarray"
This is how you load the extension.

@cindex @code{writea()} extension function
@item ret = writea(file, array)
This function takes a string argument, which is the name of the file
to which to dump the array, and the array itself as the second argument.
@code{writea()} understands arrays of arrays.  It returns one on
success, or zero upon failure.

@cindex @code{reada()} extension function
@item ret = reada(file, array)
@code{reada()} is the inverse of @code{writea()};
it reads the file named as its first argument, filling in
the array named as the second argument. It clears the array first.
Here too, the return value is one on success, or zero upon failure.
@end table

The array created by @code{reada()} is identical to that written by
@code{writea()} in the sense that the contents are the same. However,
due to implementation issues, the array traversal order of the re-created
array is likely to be different from that of the original array.  As array
traversal order in @command{awk} is by default undefined, this is (technically)
not a problem.  If you need to guarantee a particular traversal
order, use the array sorting features in @command{gawk} to do so
(@pxref{Array Sorting}).

The file contains binary data.  All integral values are written in network
byte order.  However, double-precision floating-point values are written
as native binary data.  Thus, arrays containing only string data can
theoretically be dumped on systems with one byte order and restored on
systems with a different one, but this has not been tried.

Here is an example:

@example
@@load "rwarray"
@dots{}
ret = writea("arraydump.bin", array)
@dots{}
ret = reada("arraydump.bin", array)
@end example

@node Extension Sample Readfile
@subsection Reading an Entire File

The @code{readfile} extension adds a single function
named @code{readfile()}, and an input parser:

@table @code
@item @@load "readfile"
This is how you load the extension.

@cindex @code{readfile()} extension function
@item result = readfile("/some/path")
The argument is the name of the file to read.  The return value is a
string containing the entire contents of the requested file.  Upon error,
the function returns the empty string and sets @code{ERRNO}.

@item BEGIN @{ PROCINFO["readfile"] = 1 @}
In addition, the extension adds an input parser that is activated if
@code{PROCINFO["readfile"]} exists.
When activated, each input file is returned in its entirety as @code{$0}.
@code{RT} is set to the null string.
@end table

Here is an example:

@example
@@load "readfile"
@dots{}
contents = readfile("/path/to/file");
if (contents == "" && ERRNO != "") @{
    print("problem reading file", ERRNO) > "/dev/stderr"
    ...
@}
@end example

@node Extension Sample Time
@subsection Extension Time Functions

The @code{time} extension adds two functions, named @code{gettimeofday()}
and @code{sleep()}, as follows:

@table @code
@item @@load "time"
This is how you load the extension.

@cindex @code{gettimeofday()} extension function
@item the_time = gettimeofday()
Return the time in seconds that has elapsed since 1970-01-01 UTC as a
floating-point value.  If the time is unavailable on this platform, return
@minus{}1 and set @code{ERRNO}.  The returned time should have sub-second
precision, but the actual precision may vary based on the platform.
If the standard C @code{gettimeofday()} system call is available on this
platform, then it simply returns the value.  Otherwise, if on MS-Windows,
it tries to use @code{GetSystemTimeAsFileTime()}.

@cindex @code{sleep()} extension function
@item result = sleep(@var{seconds})
Attempt to sleep for @var{seconds} seconds.  If @var{seconds} is negative,
or the attempt to sleep fails, return @minus{}1 and set @code{ERRNO}.
Otherwise, return zero after sleeping for the indicated amount of time.
Note that @var{seconds} may be a floating-point (nonintegral) value.
Implementation details: depending on platform availability, this function
tries to use @code{nanosleep()} or @code{select()} to implement the delay.
@end table

@node Extension Sample API Tests
@subsection API Tests
@cindex @code{testext} extension

The @code{testext} extension exercises parts of the extension API that
are not tested by the other samples.  The @file{extension/testext.c}
file contains both the C code for the extension and @command{awk}
test code inside C comments that run the tests. The testing framework
extracts the @command{awk} code and runs the tests.  See the source file
for more information.

@node gawkextlib
@section The @code{gawkextlib} Project
@cindex @code{gawkextlib}
@cindex extensions, where to find

@cindex @code{gawkextlib} project
The @uref{https://sourceforge.net/projects/gawkextlib/, @code{gawkextlib}}
project provides a number of @command{gawk} extensions, including one for
processing XML files.  This is the evolution of the original @command{xgawk}
(XML @command{gawk}) project.

There are a number of extensions. Some of the more interesting ones are:

@itemize @value{BULLET}
@item
@code{abort} extension. It allows you to exit immediately from your
@command{awk} program without running the @code{END} rules.

@item
@code{json} extension.
This serializes a multidimensional array into a JSON string, and
can deserialize a JSON string into a @command{gawk} array.
This extension is interesting since it is written in C++ instead of C.

@item
MPFR library extension.
This provides access to a number of MPFR functions that @command{gawk}'s
native MPFR support does not.

@item
Select extension. It provides functionality based on the
@code{select()} system call.

@item
XML parser extension, using the @uref{https://expat.sourceforge.net, Expat}
XML parsing library
@end itemize

@cindex @command{git} utility
You can check out the code for the @code{gawkextlib} project
using the @uref{https://git-scm.com, Git} distributed source
code control system.  The command is as follows:

@example
git clone git://git.code.sf.net/p/gawkextlib/code gawkextlib-code
@end example

@cindex RapidJson JSON parser library
You will need to have the @uref{http://www.rapidjson.org, RapidJson}
JSON parser library installed in order to build and use the @code{json} extension.

@cindex Expat XML parser library
You will need to have the @uref{https://expat.sourceforge.net, Expat}
XML parser library installed in order to build and use the XML extension.

In addition, you must have the GNU Autotools installed
(@uref{https://www.gnu.org/software/autoconf, Autoconf},
@uref{https://www.gnu.org/software/automake, Automake},
@uref{https://www.gnu.org/software/libtool, Libtool},
and
@uref{https://www.gnu.org/software/gettext, GNU @command{gettext}}).

The simple recipe for building and testing @code{gawkextlib} is as follows.
First, build and install @command{gawk}:

@example
cd .../path/to/gawk/code
./configure --prefix=/tmp/newgawk     @ii{Install in /tmp/newgawk for now}
make && make check                    @ii{Build and check that all is OK}
make install                          @ii{Install gawk}
@end example

Next, go to @url{https://sourceforge.net/projects/gawkextlib/files} to
download @code{gawkextlib} and any extensions that you would like to build.
The @file{README} file at that site explains how to build the code.  If you
installed @command{gawk} in a non-standard location, you will need to
specify @code{./configure --with-gawk=@var{/path/to/gawk}} to find it.
You may need to use the @command{sudo} utility
to install both @command{gawk} and @code{gawkextlib}, depending upon
how your system works.

If you write an extension that you wish to share with other
@command{gawk} users, consider doing so through the
@code{gawkextlib} project.
See the project's website for more information.

@node Extension summary
@section Summary

@itemize @value{BULLET}
@item
You can write extensions (sometimes called plug-ins) for @command{gawk}
in C or C++ using the application programming interface (API) defined
by the @command{gawk} developers.

@item
Extensions must have a license compatible with the GNU General Public
License (GPL), and they must assert that fact by declaring a variable
named @code{plugin_is_GPL_compatible}.

@item
Communication between @command{gawk} and an extension is two-way.
@command{gawk} passes a @code{struct} to the extension that contains
various data fields and function pointers.  The extension can then call
into @command{gawk} via the supplied function pointers to accomplish
certain tasks.

@item
One of these tasks is to ``register'' the name and implementation of
new @command{awk}-level functions with @command{gawk}.  The implementation
takes the form of a C function pointer with a defined signature.
By convention, implementation functions are named @code{do_@var{XXXX}()}
for some @command{awk}-level function @code{@var{XXXX}()}.

@item
The API is defined in a header file named @file{gawkapi.h}. You must include
a number of standard header files @emph{before} including it in your source file.

@item
API function pointers are provided for the following kinds of operations:

@itemize @value{BULLET}
@item
Allocating, reallocating, and releasing memory

@item
Registration functions (you may register
extension functions,
exit callbacks,
a version string,
input parsers,
output wrappers,
and two-way processors)

@item
Printing fatal, nonfatal, warning, and ``lint'' warning messages

@item
Updating @code{ERRNO}, or unsetting it

@item
Accessing parameters, including converting an undefined parameter into
an array

@item
Symbol table access (retrieving a global variable, creating one,
or changing one)

@item
Creating and releasing cached values; this provides an
efficient way to use values for multiple variables and
can be a big performance win

@item
Manipulating arrays
(retrieving, adding, deleting, and modifying elements;
getting the count of elements in an array;
creating a new array;
clearing an array;
and
flattening an array for easy C-style looping over all its indices and elements)
@end itemize

@item
The API defines a number of standard data types for representing
@command{awk} values, array elements, and arrays.

@item
The API provides convenience functions for constructing values.
It also provides memory management functions to ensure compatibility
between memory allocated by @command{gawk} and memory allocated by an
extension.

@item
@emph{All} memory passed from @command{gawk} to an extension must be
treated as read-only by the extension.

@item
@emph{All} memory passed from an extension to @command{gawk} must come from
the API's memory allocation functions. @command{gawk} takes responsibility for
the memory and releases it when appropriate.

@item
The API provides information about the running version of @command{gawk} so
that an extension can make sure it is compatible with the @command{gawk}
that loaded it.

@item
It is easiest to start a new extension by copying the boilerplate code
described in this @value{CHAPTER}.  Macros in the @file{gawkapi.h} header
file make this easier to do.

@item
The @command{gawk} distribution includes a number of small but useful
sample extensions. The @code{gawkextlib} project includes several more
(larger) extensions.  If you wish to write an extension and contribute it
to the community of @command{gawk} users, the @code{gawkextlib} project
is the place to do so.

@end itemize

@c EXCLUDE START
@node Extension Exercises
@section Exercises

@enumerate
@item
Add functions to implement system calls such as @code{chown()},
@code{chmod()}, and @code{umask()} to the file operations extension
presented in @ref{Internal File Ops}.

@c Idea from comp.lang.awk, February 2015
@item
Write an input parser that prints a prompt if the input is
a from a ``terminal'' device.  You can use the @code{isatty()}
function to tell if the input file is a terminal. (Hint: this function
is usually expensive to call; try to call it just once.)
The content of the prompt should come from a variable settable
by @command{awk}-level code.
You can write the prompt to standard error. However,
for best results, open a new file descriptor (or file pointer)
on @file{/dev/tty} and print the prompt there, in case standard
error has been redirected.

Why is standard error a better
choice than standard output for writing the prompt?
Which reading mechanism should you replace, the one to get
a record, or the one to read raw bytes?

@item
(Hard.)
How would you provide namespaces in @command{gawk}, so that the
names of functions in different extensions don't conflict with each other?
If you come up with a really good scheme, contact the @command{gawk}
maintainer to tell him about it.

@item
Write a wrapper script that provides an interface similar to
@samp{sed -i} for the ``inplace'' extension presented in
@ref{Extension Sample Inplace}.

@end enumerate
@c EXCLUDE END

@ifnotinfo
@part @value{PART4}Appendices
@end ifnotinfo

@ifdocbook

@ifclear FOR_PRINT
Part IV contains the appendices (including the two licenses that cover
the @command{gawk} source code and this @value{DOCUMENT}, respectively)
and the Glossary:
@end ifclear

@ifset FOR_PRINT
Part IV contains three appendices, the last of which is the license that
covers the @command{gawk} source code:
@end ifset

@itemize @value{BULLET}
@item
@ref{Language History}

@item
@ref{Installation}

@ifclear FOR_PRINT
@item
@ref{Notes}

@item
@ref{Basic Concepts}

@item
@ref{Glossary}
@end ifclear

@item
@ref{Copying}

@ifclear FOR_PRINT
@item
@ref{GNU Free Documentation License}
@end ifclear
@end itemize
@end ifdocbook

@node Language History
@appendix The Evolution of the @command{awk} Language

This @value{DOCUMENT} describes the GNU implementation of @command{awk},
which follows the POSIX specification.  Many longtime @command{awk}
users learned @command{awk} programming with the original @command{awk}
implementation in Version 7 Unix.  (This implementation was the basis for
@command{awk} in Berkeley Unix, through 4.3-Reno.  Subsequent versions
of Berkeley Unix, and, for a while, some systems derived from 4.4BSD-Lite, used various
versions of @command{gawk} for their @command{awk}.)  This @value{CHAPTER}
briefly describes the evolution of the @command{awk} language, with
cross-references to other parts of the @value{DOCUMENT} where you can
find more information.

@ifset FOR_PRINT
To save space, we have omitted
information on the history of features in @command{gawk} from this
edition. You can find it in the
@uref{https://www.gnu.org/software/gawk/manual/html_node/Feature-History.html,
online documentation}.
@end ifset

@menu
* V7/SVR3.1::                   The major changes between V7 and System V
                                Release 3.1.
* SVR4::                        Minor changes between System V Releases 3.1
                                and 4.
* POSIX::                       New features from the POSIX standard.
* BTL::                         New features from Brian Kernighan's version of
                                @command{awk}.
* POSIX/GNU::                   The extensions in @command{gawk} not in POSIX
                                @command{awk}.
* Feature History::             The history of the features in @command{gawk}.
* Common Extensions::           Common Extensions Summary.
* Ranges and Locales::          How locales used to affect regexp ranges.
* Contributors::                The major contributors to @command{gawk}.
* History summary::             History summary.
@end menu

@node V7/SVR3.1
@appendixsec Major Changes Between V7 and SVR3.1
@cindex @command{awk}, versions of
@cindex @command{awk}, versions of, changes between V7 and SVR3.1

The @command{awk} language evolved considerably between the release of
Version 7 Unix (1978) and the new version that was first made generally available in
System V Release 3.1 (1987).  This @value{SECTION} summarizes the changes, with
cross-references to further details:

@itemize @value{BULLET}
@item
The requirement for @samp{;} to separate rules on a line
(@pxref{Statements/Lines})

@item
User-defined functions and the @code{return} statement
(@pxref{User-defined})

@item
The @code{delete} statement (@pxref{Delete})

@item
The @code{do}-@code{while} statement
(@pxref{Do Statement})

@item
The built-in functions @code{atan2()}, @code{cos()}, @code{sin()}, @code{rand()}, and
@code{srand()} (@pxref{Numeric Functions})

@item
The built-in functions @code{gsub()}, @code{sub()}, and @code{match()}
(@pxref{String Functions})

@item
The built-in functions @code{close()} and @code{system()}
(@pxref{I/O Functions})

@item
The @code{ARGC}, @code{ARGV}, @code{FNR}, @code{RLENGTH}, @code{RSTART},
and @code{SUBSEP} predefined variables (@pxref{Built-in Variables})

@item
Assignable @code{$0} (@pxref{Changing Fields})

@item
The conditional expression using the ternary operator @samp{?:}
(@pxref{Conditional Exp})

@item
The expression @samp{@var{indx} in @var{array}} outside of @code{for}
statements (@pxref{Reference to Elements})

@item
The exponentiation operator @samp{^}
(@pxref{Arithmetic Ops}) and its assignment operator
form @samp{^=} (@pxref{Assignment Ops})

@item
C-compatible operator precedence, which breaks some old @command{awk}
programs (@pxref{Precedence})

@item
Regexps as the value of @code{FS}
(@pxref{Field Separators}) and as the
third argument to the @code{split()} function
(@pxref{String Functions}), rather than using only the first character
of @code{FS}

@item
Dynamic regexps as operands of the @samp{~} and @samp{!~} operators
(@pxref{Computed Regexps})

@item
The escape sequences @samp{\b}, @samp{\f}, and @samp{\r}
(@pxref{Escape Sequences})

@item
Redirection of input for the @code{getline} function
(@pxref{Getline})

@item
Multiple @code{BEGIN} and @code{END} rules
(@pxref{BEGIN/END})

@item
Multidimensional arrays
(@pxref{Multidimensional})
@end itemize

@node SVR4
@appendixsec Changes Between SVR3.1 and SVR4

@cindex @command{awk}, versions of, changes between SVR3.1 and SVR4
The System V Release 4 (1989) version of Unix @command{awk} added these features
(some of which originated in @command{gawk}):

@itemize @value{BULLET}
@item
The @code{ENVIRON} array (@pxref{Built-in Variables})
@c gawk and MKS awk

@item
Multiple @option{-f} options on the command line
(@pxref{Options})
@c MKS awk

@item
The @option{-v} option for assigning variables before program execution begins
(@pxref{Options})
@c GNU, Bell Laboratories & MKS together

@item
The @option{--} signal for terminating command-line options

@item
The @samp{\a}, @samp{\v}, and @samp{\x} escape sequences
(@pxref{Escape Sequences})
@c GNU, for ANSI C compat

@item
A defined return value for the @code{srand()} built-in function
(@pxref{Numeric Functions})

@item
The @code{toupper()} and @code{tolower()} built-in string functions
for case translation
(@pxref{String Functions})

@item
A cleaner specification for the @samp{%c} format-control letter in the
@code{printf} function
(@pxref{Control Letters})

@item
The ability to dynamically pass the field width and precision (@code{"%*.*d"})
in the argument list of @code{printf} and @code{sprintf()}
(@pxref{Control Letters})

@item
The use of regexp constants, such as @code{/foo/}, as expressions, where
they are equivalent to using the matching operator, as in @samp{$0 ~ /foo/}
(@pxref{Using Constant Regexps})

@item
Processing of escape sequences inside command-line variable assignments
(@pxref{Assignment Options})
@end itemize

@node POSIX
@appendixsec Changes Between SVR4 and POSIX @command{awk}
@cindex @command{awk}, versions of, changes between SVR4 and POSIX @command{awk}
@cindex POSIX @command{awk}, changes in @command{awk} versions

The POSIX Command Language and Utilities standard for @command{awk} (1992)
introduced the following changes into the language:

@itemize @value{BULLET}
@item
The use of @option{-W} for implementation-specific options
(@pxref{Options})

@item
The use of @code{CONVFMT} for controlling the conversion of numbers
to strings (@pxref{Conversion})

@item
The concept of a numeric string and tighter comparison rules to go
with it (@pxref{Typing and Comparison})

@item
The use of predefined variables as function parameter names is forbidden
(@pxref{Definition Syntax})

@item
More complete documentation of many of the previously undocumented
features of the language
@end itemize

In 2012, a number of extensions that had been commonly available for
many years were finally added to POSIX. They are:

@itemize @value{BULLET}
@item
The @code{fflush()} built-in function for flushing buffered output
(@pxref{I/O Functions})

@item
The @code{nextfile} statement
(@pxref{Nextfile Statement})

@item
The ability to delete all of an array at once with @samp{delete @var{array}}
(@pxref{Delete})

@end itemize

@xref{Common Extensions} for a list of common extensions
not permitted by the POSIX standard.

The 2008 POSIX standard can be found online at
@url{http://pubs.opengroup.org/onlinepubs/9699919799/}.


@node BTL
@appendixsec Extensions in Brian Kernighan's @command{awk}

@cindex @command{awk}, versions of, See Also Brian Kernighan's @command{awk}
@cindex extensions, Brian Kernighan's @command{awk}
@cindex Brian Kernighan's @command{awk}, extensions
@cindex Kernighan, Brian
Brian Kernighan
has made his version available via his home page
(@pxref{Other Versions}).

This @value{SECTION} describes common extensions that
originally appeared in his version of @command{awk}:

@itemize @value{BULLET}
@item
The @samp{**} and @samp{**=} operators
(@pxref{Arithmetic Ops}
and
@ref{Assignment Ops})

@item
The use of @code{func} as an abbreviation for @code{function}
(@pxref{Definition Syntax})

@item
The @code{fflush()} built-in function for flushing buffered output
(@pxref{I/O Functions})

@ignore
@item
The @code{SYMTAB} array, that allows access to @command{awk}'s internal symbol
table. This feature was never documented for his @command{awk}, largely because
it is somewhat shakily implemented. For instance, you cannot access arrays
or array elements through it
@end ignore
@end itemize

@xref{Common Extensions} for a full list of the extensions
available in his @command{awk}.

@node POSIX/GNU
@appendixsec Extensions in @command{gawk} Not in POSIX @command{awk}

@cindex compatibility mode (@command{gawk}), extensions
@cindex extensions, in @command{gawk}, not in POSIX @command{awk}
@cindex POSIX, @command{gawk} extensions not included in
The GNU implementation, @command{gawk}, adds a large number of features.
They can all be disabled with either the @option{--traditional} or
@option{--posix} options
(@pxref{Options}).

A number of features have come and gone over the years. This @value{SECTION}
summarizes the additional features over POSIX @command{awk} that are
in the current version of @command{gawk}.

@itemize @value{BULLET}

@item
Additional predefined variables:

@itemize @value{MINUS}
@item
The
@code{ARGIND},
@code{BINMODE},
@code{ERRNO},
@code{FIELDWIDTHS},
@code{FPAT},
@code{IGNORECASE},
@code{LINT},
@code{PROCINFO},
@code{RT},
and
@code{TEXTDOMAIN}
variables
(@pxref{Built-in Variables})
@end itemize

@item
Special files in I/O redirections:

@itemize @value{MINUS}
@item
The @file{/dev/stdin}, @file{/dev/stdout}, @file{/dev/stderr}, and
@file{/dev/fd/@var{N}} special @value{FN}s
(@pxref{Special Files})

@item
The @file{/inet}, @file{/inet4}, and @file{/inet6} special files for
TCP/IP networking using @samp{|&} to specify which version of the
IP protocol to use
(@pxref{TCP/IP Networking})
@end itemize

@item
Changes and/or additions to the language:

@itemize @value{MINUS}
@item
The @samp{\x} escape sequence
(@pxref{Escape Sequences})

@item
Full support for both POSIX and GNU regexps
(@pxref{Regexp})

@item
The ability for @code{FS} and for the third
argument to @code{split()} to be null strings
(@pxref{Single Character Fields})

@item
The ability for @code{RS} to be a regexp
(@pxref{Records})

@item
The ability to use octal and hexadecimal constants in @command{awk}
program source code
(@pxref{Nondecimal-numbers})

@item
The @samp{|&} operator for two-way I/O to a coprocess
(@pxref{Two-way I/O})

@item
Indirect function calls
(@pxref{Indirect Calls})

@item
Directories on the command line produce a warning and are skipped
(@pxref{Command-line directories})

@item
Output with @code{print} and @code{printf} need not be fatal
(@pxref{Nonfatal})
@end itemize

@item
New keywords:

@itemize @value{MINUS}
@item
The @code{BEGINFILE} and @code{ENDFILE} special patterns
(@pxref{BEGINFILE/ENDFILE})

@item
The @code{switch} statement
(@pxref{Switch Statement})
@end itemize

@item
Changes to standard @command{awk} functions:

@itemize @value{MINUS}
@item
The optional second argument to @code{close()} that allows closing one end
of a two-way pipe to a coprocess
(@pxref{Two-way I/O})

@item
POSIX compliance for @code{gsub()} and @code{sub()} with @option{--posix}

@item
The @code{length()} function accepts an array argument
and returns the number of elements in the array
(@pxref{String Functions})

@item
The optional third argument to the @code{match()} function
for capturing text-matching subexpressions within a regexp
(@pxref{String Functions})

@item
Positional specifiers in @code{printf} formats for
making translations easier
(@pxref{Printf Ordering})

@item
The @code{split()} function's additional optional fourth
argument, which is an array to hold the text of the field separators
(@pxref{String Functions})
@end itemize

@item
Additional functions only in @command{gawk}:

@itemize @value{MINUS}
@item
The @code{gensub()}, @code{patsplit()}, and @code{strtonum()} functions
for more powerful text manipulation
(@pxref{String Functions})

@item
The @code{asort()} and @code{asorti()} functions for sorting arrays
(@pxref{Array Sorting})

@item
The @code{mktime()}, @code{systime()}, and @code{strftime()}
functions for working with timestamps
(@pxref{Time Functions})

@item
The
@code{and()},
@code{compl()},
@code{lshift()},
@code{or()},
@code{rshift()},
and
@code{xor()}
functions for bit manipulation
(@pxref{Bitwise Functions})
@c In 4.1, and(), or() and xor() grew the ability to take > 2 arguments

@item
The @code{isarray()} function to check if a variable is an array or not
(@pxref{Type Functions})

@item
The @code{bindtextdomain()}, @code{dcgettext()}, and @code{dcngettext()}
functions for internationalization
(@pxref{Programmer i18n})

@ifset INTDIV
@item
The @code{intdiv0()} function for doing integer
division and remainder
(@pxref{Numeric Functions})
@end ifset
@end itemize

@item
Changes and/or additions in the command-line options:

@itemize @value{MINUS}
@item
The @env{AWKPATH} environment variable for specifying a path search for
the @option{-f} command-line option
(@pxref{Options})

@item
The @env{AWKLIBPATH} environment variable for specifying a path search for
the @option{-l} command-line option
(@pxref{Options})

@item
The
@option{-b},
@option{-c},
@option{-C},
@option{-d},
@option{-D},
@option{-e},
@option{-E},
@option{-g},
@option{-h},
@option{-i},
@option{-l},
@option{-L},
@option{-M},
@option{-n},
@option{-N},
@option{-o},
@option{-O},
@option{-p},
@option{-P},
@option{-r},
@option{-s},
@option{-S},
@option{-t},
and
@option{-V}
short options. Also, the
ability to use GNU-style long-named options that start with @option{--},
and the
@option{--assign},
@option{--bignum},
@option{--characters-as-bytes},
@option{--copyright},
@option{--debug},
@option{--dump-variables},
@option{--exec},
@option{--field-separator},
@option{--file},
@option{--gen-pot},
@option{--help},
@option{--include},
@option{--lint},
@option{--lint-old},
@option{--load},
@option{--non-decimal-data},
@option{--optimize},
@option{--no-optimize},
@option{--posix},
@option{--pretty-print},
@option{--profile},
@option{--re-interval},
@option{--sandbox},
@option{--source},
@option{--traditional},
@option{--use-lc-numeric},
and
@option{--version}
long options
(@pxref{Options}).
@end itemize

@c       new ports

@item
Support for the following obsolete systems was removed from the code
and the documentation for @command{gawk} @value{PVERSION} 4.0:

@c nested table
@itemize @value{MINUS}
@item
Amiga

@item
Atari

@item
BeOS

@item
Cray

@item
MIPS RiscOS

@item
MS-DOS with the Microsoft Compiler

@item
MS-Windows with the Microsoft Compiler

@item
NeXT

@item
SunOS 3.x, Sun 386 (Road Runner)

@item
Tandem (non-POSIX)

@item
Prestandard VAX C compiler for VAX/VMS

@item
GCC for VAX and Alpha has not been tested for a while.

@end itemize

@item
Support for the following obsolete system was removed from the code
for @command{gawk} @value{PVERSION} 4.1:

@c nested table
@itemize @value{MINUS}
@item
Ultrix
@end itemize

@item
Support for the following systems was removed from the code
for @command{gawk} @value{PVERSION} 4.2:

@c nested table
@itemize @value{MINUS}
@item
MirBSD

@item
GNU/Linux on Alpha
@end itemize

@end itemize

@c XXX ADD MORE STUFF HERE


@c This does not need to be in the formal book.
@ifclear FOR_PRINT
@node Feature History
@appendixsec History of @command{gawk} Features

@ignore
See the thread:
https://groups.google.com/forum/#!topic/comp.lang.awk/SAUiRuff30c
This motivated me to add this section.
@end ignore

@ignore
I've tried to follow this general order, esp.@: for the 3.0 and 3.1 sections:
       variables
       special files
       language changes (e.g., hex constants)
       differences in standard awk functions
       new gawk functions
       new keywords
       new command-line options
       behavioral changes
       extension API changes
       new / deprecated / removed ports
       installation time stuff
Within each category, be alphabetical.
@end ignore

This @value{SECTION} describes the features in @command{gawk}
over and above those in POSIX @command{awk},
in the order they were added to @command{gawk}.

Version 2.10 of @command{gawk} introduced the following features:

@itemize @value{BULLET}
@item
The @env{AWKPATH} environment variable for specifying a path search for
the @option{-f} command-line option
(@pxref{Options}).

@item
The @code{IGNORECASE} variable and its effects
(@pxref{Case-sensitivity}).

@item
The @file{/dev/stdin}, @file{/dev/stdout}, @file{/dev/stderr} and
@file{/dev/fd/@var{N}} special @value{FN}s
(@pxref{Special Files}).
@end itemize

Version 2.13 of @command{gawk} introduced the following features:

@itemize @value{BULLET}
@item
The @code{FIELDWIDTHS} variable and its effects
(@pxref{Constant Size}).

@item
The @code{systime()} and @code{strftime()} built-in functions for obtaining
and printing timestamps
(@pxref{Time Functions}).

@item
Additional command-line options
(@pxref{Options}):

@itemize @value{MINUS}
@item
The @option{-W lint} option to provide error and portability checking
for both the source code and at runtime.

@item
The @option{-W compat} option to turn off the GNU extensions.

@item
The @option{-W posix} option for full POSIX compliance.
@end itemize
@end itemize

Version 2.14 of @command{gawk} introduced the following feature:

@itemize @value{BULLET}
@item
The @code{next file} statement for skipping to the next @value{DF}
(@pxref{Nextfile Statement}).
@end itemize

Version 2.15 of @command{gawk} introduced the following features:

@itemize @value{BULLET}
@item
New variables (@pxref{Built-in Variables}):

@itemize @value{MINUS}
@item
@code{ARGIND}, which tracks the movement of @code{FILENAME}
through @code{ARGV}.

@item
@code{ERRNO}, which contains the system error message when
@code{getline} returns @minus{}1 or @code{close()} fails.
@end itemize

@item
The @file{/dev/pid}, @file{/dev/ppid}, @file{/dev/pgrpid}, and
@file{/dev/user} special @value{FN}s. These have since been removed.

@item
The ability to delete all of an array at once with @samp{delete @var{array}}
(@pxref{Delete}).

@item
Command-line option changes
(@pxref{Options}):

@itemize @value{MINUS}
@item
The ability to use GNU-style long-named options that start with @option{--}.

@item
The @option{--source} option for mixing command-line and library-file
source code.
@end itemize
@end itemize

Version 3.0 of @command{gawk} introduced the following features:

@itemize @value{BULLET}
@item
New or changed variables:

@itemize @value{MINUS}
@item
@code{IGNORECASE} changed, now applying to string comparison as well
as regexp operations
(@pxref{Case-sensitivity}).

@item
@code{RT}, which contains the input text that matched @code{RS}
(@pxref{Records}).
@end itemize

@item
Full support for both POSIX and GNU regexps
(@pxref{Regexp}).

@item
The @code{gensub()} function for more powerful text manipulation
(@pxref{String Functions}).

@item
The @code{strftime()} function acquired a default time format,
allowing it to be called with no arguments
(@pxref{Time Functions}).

@item
The ability for @code{FS} and for the third
argument to @code{split()} to be null strings
(@pxref{Single Character Fields}).

@item
The ability for @code{RS} to be a regexp
(@pxref{Records}).

@item
The @code{next file} statement became @code{nextfile}
(@pxref{Nextfile Statement}).

@item
The @code{fflush()} function from
BWK @command{awk}
(then at Bell Laboratories;
@pxref{I/O Functions}).

@item
New command-line options:

@itemize @value{MINUS}
@item
The @option{--lint-old} option to
warn about constructs that are not available in
the original Version 7 Unix version of @command{awk}
(@pxref{V7/SVR3.1}).

@item
The @option{-m} option from BWK @command{awk}.  (Brian was
still at Bell Laboratories at the time.)  This was later removed from
both his @command{awk} and from @command{gawk}.

@item
The @option{--re-interval} option to provide interval expressions in regexps
(@pxref{Regexp Operators}).

@item
The @option{--traditional} option was added as a better name for
@option{--compat} (@pxref{Options}).
@end itemize

@item
The use of GNU Autoconf to control the configuration process
(@pxref{Quick Installation}).

@item
Amiga support.
This has since been removed.

@end itemize

Version 3.1 of @command{gawk} introduced the following features:

@itemize @value{BULLET}
@item
New variables
(@pxref{Built-in Variables}):

@itemize @value{MINUS}
@item
@code{BINMODE}, for non-POSIX systems,
which allows binary I/O for input and/or output files
(@pxref{PC Using}).

@item
@code{LINT}, which dynamically controls lint warnings.

@item
@code{PROCINFO}, an array for providing process-related information.

@item
@code{TEXTDOMAIN}, for setting an application's internationalization text domain
(@pxref{Internationalization}).
@end itemize

@item
The ability to use octal and hexadecimal constants in @command{awk}
program source code
(@pxref{Nondecimal-numbers}).

@item
The @samp{|&} operator for two-way I/O to a coprocess
(@pxref{Two-way I/O}).

@item
The @file{/inet} special files for TCP/IP networking using @samp{|&}
(@pxref{TCP/IP Networking}).

@item
The optional second argument to @code{close()} that allows closing one end
of a two-way pipe to a coprocess
(@pxref{Two-way I/O}).

@item
The optional third argument to the @code{match()} function
for capturing text-matching subexpressions within a regexp
(@pxref{String Functions}).

@item
Positional specifiers in @code{printf} formats for
making translations easier
(@pxref{Printf Ordering}).

@item
A number of new built-in functions:

@itemize @value{MINUS}
@item
The @code{asort()} and @code{asorti()} functions for sorting arrays
(@pxref{Array Sorting}).

@item
The @code{bindtextdomain()}, @code{dcgettext()} and @code{dcngettext()} functions
for internationalization
(@pxref{Programmer i18n}).

@item
The @code{extension()} function and the ability to add
new built-in functions dynamically
(@pxref{Dynamic Extensions}).

@item
The @code{mktime()} function for creating timestamps
(@pxref{Time Functions}).

@item
The @code{and()}, @code{or()}, @code{xor()}, @code{compl()},
@code{lshift()}, @code{rshift()}, and @code{strtonum()} functions
(@pxref{Bitwise Functions}).
@end itemize

@item
@cindex @code{next file} statement
The support for @samp{next file} as two words was removed completely
(@pxref{Nextfile Statement}).

@item
Additional command-line options
(@pxref{Options}):

@itemize @value{MINUS}
@item
The @option{--dump-variables} option to print a list of all global variables.

@item
The @option{--exec} option, for use in CGI scripts.

@item
The @option{--gen-po} command-line option and the use of a leading
underscore to mark strings that should be translated
(@pxref{String Extraction}).

@item
The @option{--non-decimal-data} option to allow non-decimal
input data
(@pxref{Nondecimal Data}).

@item
The @option{--profile} option and @command{pgawk}, the
profiling version of @command{gawk}, for producing execution
profiles of @command{awk} programs
(@pxref{Profiling}).

@item
The @option{--use-lc-numeric} option to force @command{gawk}
to use the locale's decimal point for parsing input data
(@pxref{Conversion}).
@end itemize

@item
The use of GNU Automake to help in standardizing the configuration process
(@pxref{Quick Installation}).

@item
The use of GNU @command{gettext} for @command{gawk}'s own message output
(@pxref{Gawk I18N}).

@item
BeOS support. This was later removed.

@item
Tandem support. This was later removed.

@item
The Atari port became officially unsupported and was
later removed entirely.

@item
The source code changed to use ISO C standard-style function definitions.

@item
POSIX compliance for @code{sub()} and @code{gsub()}
(@pxref{Gory Details}).

@item
The @code{length()} function was extended to accept an array argument
and return the number of elements in the array
(@pxref{String Functions}).

@item
The @code{strftime()} function acquired a third argument to
enable printing times as UTC
(@pxref{Time Functions}).
@end itemize

Version 4.0 of @command{gawk} introduced the following features:

@itemize @value{BULLET}

@item
Variable additions:

@itemize @value{MINUS}
@item
@code{FPAT}, which allows you to specify a regexp that matches
the fields, instead of matching the field separator
(@pxref{Splitting By Content}).

@item
If @code{PROCINFO["sorted_in"]} exists, @samp{for(iggy in foo)} loops sort the
indices before looping over them.  The value of this element
provides control over how the indices are sorted before the loop
traversal starts
(@pxref{Controlling Scanning}).

@item
@code{PROCINFO["strftime"]}, which holds
the default format for @code{strftime()}
(@pxref{Time Functions}).
@end itemize

@item
The special files @file{/dev/pid}, @file{/dev/ppid}, @file{/dev/pgrpid}
and @file{/dev/user} were removed.

@item
Support for IPv6 was added via the @file{/inet6} special file.
@file{/inet4} forces IPv4 and @file{/inet} chooses the system
default, which is probably IPv4
(@pxref{TCP/IP Networking}).

@item
The use of @samp{\s} and @samp{\S} escape sequences in regular expressions
(@pxref{GNU Regexp Operators}).

@item
Interval expressions became part of default regular expressions
(@pxref{Regexp Operators}).

@item
POSIX character classes work even with @option{--traditional}
(@pxref{Regexp Operators}).

@item
@code{break} and @code{continue} became invalid outside a loop,
even with @option{--traditional}
(@pxref{Break Statement}, and also see
@ref{Continue Statement}).

@item
@code{fflush()}, @code{nextfile}, and @samp{delete @var{array}}
are allowed if @option{--posix} or @option{--traditional}, since they
are all now part of POSIX.

@item
An optional third argument to
@code{asort()} and @code{asorti()}, specifying how to sort
(@pxref{String Functions}).

@item
The behavior of @code{fflush()} changed to match BWK @command{awk}
and for POSIX; now both @samp{fflush()} and @samp{fflush("")}
flush all open output redirections
(@pxref{I/O Functions}).

@item
The @code{isarray()}
function which distinguishes if an item is an array
or not, to make it possible to traverse arrays of arrays
(@pxref{Type Functions}).

@item
The @code{patsplit()}
function which gives the same capability as @code{FPAT}, for splitting
(@pxref{String Functions}).

@item
An optional fourth argument to the @code{split()} function,
which is an array to hold the values of the separators
(@pxref{String Functions}).

@item
Arrays of arrays
(@pxref{Arrays of Arrays}).

@item
The @code{BEGINFILE} and @code{ENDFILE} special patterns
(@pxref{BEGINFILE/ENDFILE}).

@item
Indirect function calls
(@pxref{Indirect Calls}).

@item
@code{switch} / @code{case} are enabled by default
(@pxref{Switch Statement}).

@item
Command-line option changes
(@pxref{Options}):

@itemize @value{MINUS}
@item
The @option{-b} and @option{--characters-as-bytes} options
which prevent @command{gawk} from treating input as a multibyte string.

@item
The redundant @option{--compat}, @option{--copyleft}, and @option{--usage}
long options were removed.

@item
The @option{--gen-po} option was finally renamed to the correct @option{--gen-pot}.

@item
The @option{--sandbox} option which disables certain features.

@item
All long options acquired corresponding short options, for use in @samp{#!} scripts.
@end itemize

@item
Directories named on the command line now produce a warning, not a fatal
error, unless @option{--posix} or @option{--traditional} are used
(@pxref{Command-line directories}).

@item
The @command{gawk} internals were rewritten, bringing the @command{dgawk}
debugger and possibly improved performance
(@pxref{Debugger}).

@item
Per the GNU Coding Standards, dynamic extensions must now define
a global symbol indicating that they are GPL-compatible
(@pxref{Plugin License}).

@item
In POSIX mode, string comparisons use @code{strcoll()} / @code{wcscoll()}
(@pxref{POSIX String Comparison}).

@item
The option for raw sockets was removed, since it was never implemented
(@pxref{TCP/IP Networking}).

@item
Ranges of the form @samp{[d-h]} are treated as if they were in the
C locale, no matter what kind of regexp is being used, and even if
@option{--posix}
(@pxref{Ranges and Locales}).

@item
Support was removed for the following systems:

@itemize @value{MINUS}
@item
Atari

@item
Amiga

@item
BeOS

@item
Cray

@item
MIPS RiscOS

@item
MS-DOS with the Microsoft Compiler

@item
MS-Windows with the Microsoft Compiler

@item
NeXT

@item
SunOS 3.x, Sun 386 (Road Runner)

@item
Tandem (non-POSIX)

@item
Prestandard VAX C compiler for VAX/VMS
@end itemize
@end itemize

Version 4.1 of @command{gawk} introduced the following features:

@itemize @value{BULLET}

@item
Three new arrays:
@code{SYMTAB}, @code{FUNCTAB}, and @code{PROCINFO["identifiers"]}
(@pxref{Auto-set}).

@item
The three executables @command{gawk}, @command{pgawk}, and @command{dgawk}, were merged into
one, named just @command{gawk}.  As a result the command-line options changed.

@item
Command-line option changes
(@pxref{Options}):

@itemize @value{MINUS}
@item
The @option{-D} option invokes the debugger.

@item
The @option{-i} and @option{--include} options
load @command{awk} library files.

@item
The @option{-l} and @option{--load} options load compiled dynamic extensions.

@item
The @option{-M} and @option{--bignum} options enable MPFR.

@item
The @option{-o} option only does pretty-printing.

@item
The @option{-p} option is used for profiling.

@item
The @option{-R} option was removed.
@end itemize

@item
Support for high precision arithmetic with MPFR
(@pxref{Arbitrary Precision Arithmetic}).

@item
The @code{and()}, @code{or()} and @code{xor()} functions
changed to allow any number of arguments,
with a minimum of two
(@pxref{Bitwise Functions}).

@item
The dynamic extension interface was completely redone
(@pxref{Dynamic Extensions}).

@item
Redirected @code{getline} became allowed inside
@code{BEGINFILE} and @code{ENDFILE}
(@pxref{BEGINFILE/ENDFILE}).

@item
The @code{where} command was added to the debugger
(@pxref{Execution Stack}).

@item
Support for Ultrix was removed.

@end itemize

Version 4.2 of @command{gawk} introduced the following changes:

@itemize @bullet
@item
Changes to @code{ENVIRON} are reflected into @command{gawk}'s
environment and that of programs that it runs.
@xref{Auto-set}.

@item
@code{FIELDWIDTHS} was enhanced to allow skipping characters
before assigning a value to a field
(@pxref{Splitting By Content}).

@item
The @code{PROCINFO["argv"]} array.
@xref{Auto-set}.

@item
The maximum number of hexadecimal digits in @samp{\x} escapes
is now two.
@xref{Escape Sequences}.

@item
Strongly typed regexp constants of the form @samp{@@/@dots{}/}
(@pxref{Strong Regexp Constants}).

@item
The bitwise functions changed, making negative arguments into
a fatal error (@pxref{Bitwise Functions}).

@ifset INTDIV
@item
The @code{intdiv0()} function.
@xref{Numeric Functions}.
@end ifset

@item
The @code{mktime()} function now accepts an optional
second argument
(@pxref{Time Functions}).

@item
The @code{typeof()} function (@pxref{Type Functions}).

@item
Optimizations are enabled by default. Use @option{-s} /
@option{--no-optimize} to disable optimizations.

@item
For many years, POSIX specified that default field splitting
only allowed spaces and tabs to separate fields, and this was
how @command{gawk} behaved with @option{--posix}. As of 2013,
the standard restored historical behavior, and now default
field splitting with @option{--posix} also allows newlines to
separate fields.

@item
Nonfatal output with @code{print} and @code{printf}.
@xref{Nonfatal}.

@item
Retryable I/O via @code{PROCINFO[@var{input-file}, "RETRY"]};
(@pxref{Retrying Input}).

@item
Changes to the pretty-printer (@pxref{Profiling}):

@c nested table
@itemize @value{MINUS}
@item
The @option{--pretty-print} option no longer runs the @command{awk}
program too.

@item
Comments in the source program are preserved and placed into the
output file.

@item
Explicit parentheses for expressions
in the input are preserved in the generated output.
@end itemize

@item
Improvements to the extension API
(@pxref{Dynamic Extensions}):

@c nested
@itemize @value{MINUS}
@item
The @code{get_file()} function to access open redirections.

@item
The @code{nonfatal()} function for generating nonfatal error messages.

@item
Support for GMP and MPFR values.

@item
Input parsers can now override the default field parsing mechanism
by specifying explicit locations.
@end itemize

@item
Shell startup files are supplied with the distribution and
installed by @samp{make install} (@pxref{Shell Startup Files}).

@item
The @command{igawk} program and its manual page are no longer
installed when @command{gawk} is built.
@xref{Igawk Program}.

@item
Support for MirBSD was removed.

@item
Support for GNU/Linux on Alpha was removed.

@end itemize

@c XXX ADD MORE STUFF HERE
@end ifclear

@node Common Extensions
@appendixsec Common Extensions Summary

@cindex extensions, Brian Kernighan's @command{awk}
@cindex extensions, @command{mawk}
The following table summarizes the common extensions supported
by @command{gawk}, Brian Kernighan's @command{awk}, and @command{mawk},
the three most widely used freely available versions of @command{awk}
(@pxref{Other Versions}).

@multitable {@file{/dev/stderr} special file} {BWK @command{awk}} {@command{mawk}} {@command{gawk}} {Now standard}
@headitem Feature @tab BWK @command{awk} @tab @command{mawk} @tab @command{gawk} @tab Now standard
@item @samp{\x} escape sequence @tab X @tab X @tab X @tab
@item @code{FS} as null string @tab X @tab X @tab X @tab
@item @file{/dev/stdin} special file @tab X @tab X @tab X @tab
@item @file{/dev/stdout} special file @tab X @tab X @tab X @tab
@item @file{/dev/stderr} special file @tab X @tab X @tab X @tab
@item @code{delete} without subscript @tab X @tab X @tab X @tab X
@item @code{fflush()} function @tab X @tab X @tab X @tab X
@item @code{length()} of an array @tab X @tab X @tab X @tab
@item @code{nextfile} statement @tab X @tab X @tab X @tab X
@item @code{**} and @code{**=} operators @tab X @tab @tab X @tab
@item @code{func} keyword @tab X @tab @tab X @tab
@item @code{BINMODE} variable @tab @tab X @tab X @tab
@item @code{RS} as regexp @tab @tab X @tab X @tab
@item Time-related functions @tab @tab X @tab X @tab
@end multitable

@node Ranges and Locales
@appendixsec Regexp Ranges and Locales: A Long Sad Story

This @value{SECTION} describes the confusing history of ranges within
regular expressions and their interactions with locales, and how this
affected different versions of @command{gawk}.

The original Unix tools that worked with regular expressions defined
character ranges (such as @samp{[a-z]}) to match any character between
the first character in the range and the last character in the range,
inclusive.  Ordering was based on the numeric value of each character
in the machine's native character set.  Thus, on ASCII-based systems,
@samp{[a-z]} matched all the lowercase letters, and only the lowercase
letters, as the numeric values for the letters from @samp{a} through
@samp{z} were contiguous.  (On an EBCDIC system, the range @samp{[a-z]}
includes additional nonalphabetic characters as well.)

Almost all introductory Unix literature explained range expressions
as working in this fashion, and in particular, would teach that the
``correct'' way to match lowercase letters was with @samp{[a-z]}, and
that @samp{[A-Z]} was the ``correct'' way to match uppercase letters.
And indeed, this was true.@footnote{And Life was good.}

The 1992 POSIX standard introduced the idea of locales (@pxref{Locales}).
Because many locales include other letters besides the plain 26
letters of the English alphabet, the POSIX standard added
character classes (@pxref{Bracket Expressions}) as a way to match
different kinds of characters besides the traditional ones in the ASCII
character set.

However, the standard @emph{changed} the interpretation of range expressions.
In the @code{"C"} and @code{"POSIX"} locales, a range expression like
@samp{[a-dx-z]} is still equivalent to @samp{[abcdxyz]}, as in ASCII.
But outside those locales, the ordering was defined to be based on
@dfn{collation order}.

What does that mean?
In many locales, @samp{A} and @samp{a} are both less than @samp{B}.
In other words, these locales sort characters in dictionary order,
and @samp{[a-dx-z]} is typically not equivalent to @samp{[abcdxyz]};
instead, it might be equivalent to @samp{[ABCXYabcdxyz]}, for example.

This point needs to be emphasized: much literature teaches that you should
use @samp{[a-z]} to match a lowercase character.  But on systems with
non-ASCII locales, this also matches all of the uppercase characters
except @samp{A} or @samp{Z}!  This was a continuous cause of confusion, even well
into the twenty-first century.

To demonstrate these issues, the following example uses the @code{sub()}
function, which does text replacement (@pxref{String Functions}).  Here,
the intent is to remove trailing uppercase characters:

@example
$ @kbd{echo something1234abc | gawk-3.1.8 '@{ sub("[A-Z]*$", ""); print @}'}
@print{} something1234a
@end example

@noindent
This output is unexpected, as the @samp{bc} at the end of
@samp{something1234abc} should not normally match @samp{[A-Z]*}.
This result is due to the locale setting (and thus you may not see
it on your system).

@cindex Unicode
Similar considerations apply to other ranges.  For example, @samp{["-/]}
is perfectly valid in ASCII, but is not valid in many Unicode locales,
such as @code{en_US.UTF-8}.

Early versions of @command{gawk} used regexp matching code that was not
locale-aware, so ranges had their traditional interpretation.

When @command{gawk} switched to using locale-aware regexp matchers,
the problems began; especially as both GNU/Linux and commercial Unix
vendors started implementing non-ASCII locales, @emph{and making them
the default}.  Perhaps the most frequently asked question became something
like, ``Why does @samp{[A-Z]} match lowercase letters?!?''

@cindex Berry, Karl
This situation existed for close to 10 years, if not more, and
the @command{gawk} maintainer grew weary of trying to explain that
@command{gawk} was being nicely standards-compliant, and that the issue
was in the user's locale.  During the development of @value{PVERSION} 4.0,
he modified @command{gawk} to always treat ranges in the original,
pre-POSIX fashion, unless @option{--posix} was used (@pxref{Options}).@footnote{And
thus was born the Campaign for Rational Range Interpretation (or
RRI). A number of GNU tools have already implemented this change,
or will soon.  Thanks to Karl Berry for coining the phrase ``Rational
Range Interpretation.''}

Fortunately, shortly before the final release of @command{gawk} 4.0,
the maintainer learned that the 2008 standard had changed the
definition of ranges, such that outside the @code{"C"} and @code{"POSIX"}
locales, the meaning of range expressions was @emph{undefined}.@footnote{See
@uref{http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap09.html#tag_09_03_05, the standard}
and
@uref{http://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap09.html#tag_21_09_03_05, its rationale}.}

By using this lovely technical term, the standard gives license
to implementers to implement ranges in whatever way they choose.
The @command{gawk} maintainer chose to apply the pre-POSIX meaning
both with the default regexp matching and when @option{--traditional} or
@option{--posix} are used.
In all cases @command{gawk} remains POSIX-compliant.

@node Contributors
@appendixsec Major Contributors to @command{gawk}
@cindex @command{gawk}, list of contributors to
@quotation
@i{Always give credit where credit is due.}
@author Anonymous
@end quotation

This @value{SECTION} names the major contributors to @command{gawk}
and/or this @value{DOCUMENT}, in approximate chronological order:

@itemize @value{BULLET}
@item
@cindex Aho, Alfred
@cindex Weinberger, Peter
@cindex Kernighan, Brian
Dr.@: Alfred V.@: Aho,
Dr.@: Peter J.@: Weinberger, and
Dr.@: Brian W.@: Kernighan, all of Bell Laboratories,
designed and implemented Unix @command{awk},
from which @command{gawk} gets the majority of its feature set.

@item
@cindex Rubin, Paul
Paul Rubin
did the initial design and implementation in 1986, and wrote
the first draft (around 40 pages) of this @value{DOCUMENT}.

@item
@cindex Fenlason, Jay
Jay Fenlason
finished the initial implementation.

@item
@cindex Close, Diane
Diane Close
revised the first draft of this @value{DOCUMENT}, bringing it
to around 90 pages.

@item
@cindex Stallman, Richard
Richard Stallman
helped finish the implementation and the initial draft of this
@value{DOCUMENT}.
He is also the founder of the FSF and the GNU Project.

@item
@cindex Woods, John
John Woods
contributed parts of the code (mostly fixes) in
the initial version of @command{gawk}.

@item
@cindex Trueman, David
In 1988,
David Trueman
took over primary maintenance of @command{gawk},
making it compatible with ``new'' @command{awk}, and
greatly improving its performance.

@item
@cindex Kwok, Conrad
@cindex Garfinkle, Scott
@cindex Williams, Kent
Conrad Kwok,
Scott Garfinkle,
and
Kent Williams
did the initial ports to MS-DOS with various versions of MSC.

@item
@cindex Rankin, Pat
Pat Rankin
provided the VMS port and its documentation.

@item
@cindex Peterson, Hal
Hal Peterson
provided help in porting @command{gawk} to Cray systems.
(This is no longer supported.)

@item
@cindex Rommel, Kai Uwe
Kai Uwe Rommel
provided the initial port to OS/2 and its documentation.

@item
@cindex Jaegermann, Michal
Michal Jaegermann
provided the port to Atari systems and its documentation.
(This port is no longer supported.)
He continues to provide portability checking,
and has done a lot of work to make sure @command{gawk}
works on non-32-bit systems.

@item
@cindex Fish, Fred
Fred Fish
provided the port to Amiga systems and its documentation.
(With Fred's sad passing, this is no longer supported.)

@item
@cindex Deifik, Scott
Scott Deifik
formerly maintained the MS-DOS port using DJGPP.

@item
@cindex Zaretskii, Eli
Eli Zaretskii
currently maintains the MS-Windows port using MinGW.

@item
@cindex Grigera, Juan
Juan Grigera
provided a port to Windows32 systems.
(This is no longer supported.)


@item
@cindex Hankerson, Darrel
For many years,
Dr.@: Darrel Hankerson
acted as coordinator for the various ports to different PC platforms
and created binary distributions for various PC operating systems.
He was also instrumental in keeping the documentation up to date for
the various PC platforms.

@item
@cindex Zoulas, Christos
Christos Zoulas
provided the @code{extension()}
built-in function for dynamically adding new functions.
(This was obsoleted at @command{gawk} 4.1.)

@item
@cindex Kahrs, J@"urgen
J@"urgen Kahrs
contributed the initial version of the TCP/IP networking
code and documentation, and motivated the inclusion of the @samp{|&} operator.

@item
@cindex Davies, Stephen
Stephen Davies
provided the initial port to Tandem systems and its documentation.
(However, this is no longer supported.)
He was also instrumental in the initial work to integrate the
byte-code internals into the @command{gawk} code base.
Additionally, he did most of the work enabling the pretty-printer
to preserve and output comments.

@item
@cindex Woehlke, Matthew
Matthew Woehlke
provided improvements for Tandem's POSIX-compliant systems.

@item
@cindex Brown, Martin
Martin Brown
provided the port to BeOS and its documentation.
(This is no longer supported.)

@item
@cindex Peters, Arno
Arno Peters
did the initial work to convert @command{gawk} to use
GNU Automake and GNU @command{gettext}.

@item
@cindex Broder, Alan J.@:
Alan J.@: Broder
provided the initial version of the @code{asort()} function
as well as the code for the optional third argument to the
@code{match()} function.

@item
@cindex Buening, Andreas
Andreas Buening
updated the @command{gawk} port for OS/2.

@item
@cindex Hasegawa, Isamu
Isamu Hasegawa,
of IBM in Japan, contributed support for multibyte characters.

@item
@cindex Benzinger, Michael
Michael Benzinger contributed the initial code for @code{switch} statements.

@item
@cindex McPhee, Patrick
Patrick T.J.@: McPhee contributed the code for dynamic loading in Windows32
environments.
(This is no longer supported.)

@item
@cindex Wallin, Anders
Anders Wallin helped keep the VMS port going for several years.

@item
@cindex Gordon, Assaf
Assaf Gordon contributed the code to implement the
@option{--sandbox} option.

@item
@cindex Haque, John
John Haque made the following contributions:

@itemize @value{MINUS}
@item
The modifications to convert @command{gawk}
into a byte-code interpreter, including the debugger

@item
The addition of true arrays of arrays

@item
The additional modifications for support of arbitrary-precision arithmetic

@item
The initial text of
@ref{Arbitrary Precision Arithmetic}

@item
The work to merge the three versions of @command{gawk}
into one, for the 4.1 release

@item
Improved array internals for arrays indexed by integers

@item
The improved array sorting features were also driven by John, together
with Pat Rankin
@end itemize

@cindex Papadopoulos, Panos
@item
Panos Papadopoulos contributed the original text for @ref{Include Files}.

@item
@cindex Yawitz, Efraim
Efraim Yawitz contributed the original text for @ref{Debugger}.

@item
@cindex Schorr, Andrew
The development of the extension API first released with
@command{gawk} 4.1 was driven primarily by
Arnold Robbins and Andrew Schorr, with notable contributions from
the rest of the development team.

@cindex Malmberg, John E.
@item
John Malmberg contributed significant improvements to the
OpenVMS port and the related documentation.

@item
@cindex Colombo, Antonio
Antonio Giovanni Colombo rewrote a number of examples in the early
chapters that were severely dated, for which I am incredibly grateful.
He also provided and maintains the Italian translation.

@item
@cindex Curreli, Marco
Marco Curreli, together with Antonio Colombo, translated this
@value{DOCUMENT} into Italian.  It is included in the @command{gawk}
distribution.

@item
@cindex Guerrero, Juan Manuel
Juan Manuel Guerrero took over maintenance of the DJGPP port.

@item
@cindex Robbins, Arnold
Arnold Robbins
has been working on @command{gawk} since 1988, at first
helping David Trueman, and as the primary maintainer since around 1994.
@end itemize

@node History summary
@appendixsec Summary

@itemize @value{BULLET}
@item
The @command{awk} language has evolved over time. The first release
was with V7 Unix, circa 1978.  In 1987, for System V Release 3.1,
major additions, including user-defined functions, were made to the language.
Additional changes were made for System V Release 4, in 1989.
Since then, further minor changes have happened under the auspices of the
POSIX standard.

@item
Brian Kernighan's @command{awk} provides a small number of extensions
that are implemented in common with other versions of @command{awk}.

@item
@command{gawk} provides a large number of extensions over POSIX @command{awk}.
They can be disabled with either the @option{--traditional} or @option{--posix}
options.

@item
The interaction of POSIX locales and regexp matching in @command{gawk} has been confusing over
the years. Today, @command{gawk} implements Rational Range Interpretation, where
ranges of the form @samp{[a-z]} match @emph{only} the characters numerically between
@samp{a} through @samp{z} in the machine's native character set.  Usually this is ASCII,
but it can be EBCDIC on IBM S/390 systems.

@item
Many people have contributed to @command{gawk} development over the years.
We hope that the list provided in this @value{CHAPTER} is complete and gives
the appropriate credit where credit is due.

@end itemize

@node Installation
@appendix Installing @command{gawk}

@c last two commas are part of see also
@cindex operating systems, See Also GNU/Linux@comma{} PC operating systems@comma{} Unix
@cindex @command{gawk}, installing
@cindex installing @command{gawk}
This appendix provides instructions for installing @command{gawk} on the
various platforms that are supported by the developers.  The primary
developer supports GNU/Linux (and Unix), whereas the other ports are
contributed.
@xref{Bugs}
for the email addresses of the people who maintain
the respective ports.

@menu
* Gawk Distribution::           What is in the @command{gawk} distribution.
* Unix Installation::           Installing @command{gawk} under various
                                versions of Unix.
* Non-Unix Installation::       Installation on Other Operating Systems.
* Bugs::                        Reporting Problems and Bugs.
* Other Versions::              Other freely available @command{awk}
                                implementations.
* Installation summary::        Summary of installation.
@end menu

@node Gawk Distribution
@appendixsec The @command{gawk} Distribution
@cindex source code, @command{gawk}

This @value{SECTION} describes how to get the @command{gawk}
distribution, how to extract it, and then what is in the various files and
subdirectories.

@menu
* Getting::                     How to get the distribution.
* Extracting::                  How to extract the distribution.
* Distribution contents::       What is in the distribution.
@end menu

@node Getting
@appendixsubsec Getting the @command{gawk} Distribution
@cindex @command{gawk}, source code@comma{} obtaining
There are two ways to get GNU software:

@itemize @value{BULLET}
@item
Copy it from someone else who already has it.

@cindex FSF (Free Software Foundation)
@cindex Free Software Foundation (FSF)
@item
Retrieve @command{gawk}
from the Internet host
@code{ftp.gnu.org}, in the directory @file{/gnu/gawk}.
Both anonymous @command{ftp} and @code{http} access are supported.
If you have the @command{wget} program, you can use a command like
the following:

@example
wget https://ftp.gnu.org/gnu/gawk/gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz
@end example
@end itemize

The GNU software archive is mirrored around the world.
The up-to-date list of mirror sites is available from
@uref{https://www.gnu.org/order/ftp.html, the main FSF website}.
Try to use one of the mirrors; they
will be less busy, and you can usually find one closer to your site.

You may also retrieve the @command{gawk} source code from the official
Git repository; for more information see @ref{Accessing The Source}.

@node Extracting
@appendixsubsec Extracting the Distribution
@command{gawk} is distributed as several @code{tar} files compressed with
different compression programs: @command{gzip}, @command{bzip2},
and @command{xz}. For simplicity, the rest of these instructions assume
you are using the one compressed with the GNU Gzip program (@command{gzip}).

Once you have the distribution (e.g.,
@file{gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz}),
use @code{gzip} to expand the
file and then use @code{tar} to extract it.  You can use the following
pipeline to produce the @command{gawk} distribution:

@example
gzip -d -c gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz | tar -xvpf -
@end example

On a system with GNU @command{tar}, you can let @command{tar}
do the decompression for you:

@example
tar -xvpzf gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz
@end example

@noindent
Extracting the archive
creates a directory named @file{gawk-@value{VERSION}.@value{PATCHLEVEL}}
in the current directory.

The distribution @value{FN} is of the form
@file{gawk-@var{V}.@var{R}.@var{P}.tar.gz}.
The @var{V} represents the major version of @command{gawk},
the @var{R} represents the current release of version @var{V}, and
the @var{P} represents a @dfn{patch level}, meaning that minor bugs have
been fixed in the release.  The current patch level is @value{PATCHLEVEL},
but when retrieving distributions, you should get the version with the highest
version, release, and patch level.  (Note, however, that patch levels greater than
or equal to 70 denote ``beta'' or nonproduction software; you might not want
to retrieve such a version unless you don't mind experimenting.)
If you are not on a Unix or GNU/Linux system, you need to make other arrangements
for getting and extracting the @command{gawk} distribution.  You should consult
a local expert.

@node Distribution contents
@appendixsubsec Contents of the @command{gawk} Distribution
@cindex @command{gawk}, distribution

The @command{gawk} distribution has a number of C source files,
documentation files,
subdirectories, and files related to the configuration process
(@pxref{Unix Installation}),
as well as several subdirectories related to different non-Unix
operating systems:

@table @asis
@item Various @samp{.c}, @samp{.y}, and @samp{.h} files
These files contain the actual @command{gawk} source code.
@end table

@table @file
@item support/*
C header and source files for routines that @command{gawk}
uses, but that are not part of its core functionality.
For example, argument parsing, regular expression matching,
and random number generating routines are all kept here.

@item ABOUT-NLS
A file containing information about GNU @command{gettext} and translations.

@item AUTHORS
A file with some information about the authorship of @command{gawk}.
It exists only to satisfy the pedants at the Free Software Foundation.

@item README
@itemx README_d/README.*
Descriptive files: @file{README} for @command{gawk} under Unix and the
rest for the various hardware and software combinations.

@item INSTALL
A file providing an overview of the configuration and installation process.

@item ChangeLog
A detailed list of source code changes as bugs are fixed or improvements made.

@item ChangeLog.0
An older list of source code changes.

@item NEWS
A list of changes to @command{gawk} since the last release or patch.

@item NEWS.0
An older list of changes to @command{gawk}.

@item COPYING
The GNU General Public License.

@item POSIX.STD
A description of behaviors in the POSIX standard for @command{awk} that
are left undefined, or where @command{gawk} may not comply fully, as well
as a list of things that the POSIX standard should describe but does not.

@cindex artificial intelligence@comma{} @command{gawk} and
@item doc/awkforai.txt
Pointers to the original draft of
a short article describing why @command{gawk} is a good language for
artificial intelligence (AI) programming.

@item doc/bc_notes
A brief description of @command{gawk}'s ``byte code'' internals.

@item doc/README.card
@itemx doc/ad.block
@itemx doc/awkcard.in
@itemx doc/cardfonts
@itemx doc/colors
@itemx doc/macros
@itemx doc/no.colors
@itemx doc/setter.outline
The @command{troff} source for a five-color @command{awk} reference card.
A modern version of @command{troff} such as GNU @command{troff} (@command{groff}) is
needed to produce the color version. See the file @file{README.card}
for instructions if you have an older @command{troff}.

@item doc/gawk.1
The @command{troff} source for a manual page describing @command{gawk}.
This is distributed for the convenience of Unix users.

@cindex Texinfo
@item doc/gawktexi.in
@itemx doc/sidebar.awk
The Texinfo source file for this @value{DOCUMENT}.
It should be processed by @file{doc/sidebar.awk}
before processing with @command{texi2dvi} or @command{texi2pdf}
to produce a printed document, and
with @command{makeinfo} to produce an Info or HTML file.
The @file{Makefile} takes care of this processing and produces
printable output via @command{texi2dvi} or @command{texi2pdf}.

@item doc/gawk.texi
The file produced after processing @file{gawktexi.in}
with @file{sidebar.awk}.

@item doc/gawk.info
The generated Info file for this @value{DOCUMENT}.

@item doc/gawkinet.texi
The Texinfo source file for
@ifinfo
@inforef{Top, , General Introduction, gawkinet, @value{GAWKINETTITLE}}.
@end ifinfo
@ifnotinfo
@cite{@value{GAWKINETTITLE}}.
@end ifnotinfo
It should be processed with @TeX{}
(via @command{texi2dvi} or @command{texi2pdf})
to produce a printed document and
with @command{makeinfo} to produce an Info or HTML file.

@item doc/gawkinet.info
The generated Info file for
@cite{@value{GAWKINETTITLE}}.

@item doc/igawk.1
The @command{troff} source for a manual page describing the @command{igawk}
program presented in
@ref{Igawk Program}.
(Since @command{gawk} can do its own @code{@@include} processing,
neither @command{igawk} nor @file{igawk.1} are installed.)

@item doc/it/*
Files for the Italian translation of this @value{DOCUMENT}, produced and
contributed by Antonio Colombo and Marco Curreli.

@item doc/Makefile.in
The input file used during the configuration process to generate the
actual @file{Makefile} for creating the documentation.

@item Makefile.am
@itemx */Makefile.am
Files used by the GNU Automake software for generating
the @file{Makefile.in} files used by Autoconf and
@command{configure}.

@item Makefile.in
@itemx aclocal.m4
@itemx bisonfix.awk
@itemx config.guess
@itemx configh.in
@itemx configure.ac
@itemx configure
@itemx custom.h
@itemx depcomp
@itemx install-sh
@itemx missing_d/*
@itemx mkinstalldirs
@itemx m4/*
These files and subdirectories are used when configuring and compiling
@command{gawk} for various Unix systems.  Most of them are explained
in @ref{Unix Installation}. The rest are there to support the main
infrastructure.

@item po/*
The @file{po} library contains message translations.

@item awklib/extract.awk
@itemx awklib/Makefile.am
@itemx awklib/Makefile.in
@itemx awklib/eg/*
The @file{awklib} directory contains a copy of @file{extract.awk}
(@pxref{Extract Program}),
which can be used to extract the sample programs from the Texinfo
source file for this @value{DOCUMENT}. It also contains a @file{Makefile.in} file, which
@command{configure} uses to generate a @file{Makefile}.
@file{Makefile.am} is used by GNU Automake to create @file{Makefile.in}.
The library functions from
@ref{Library Functions},
are included as ready-to-use files in the @command{gawk} distribution.
They are installed as part of the installation process.
The rest of the programs in this @value{DOCUMENT} are available in appropriate
subdirectories of @file{awklib/eg}.

@item extension/*
The source code, manual pages, and infrastructure files for
the sample extensions included with @command{gawk}.
@xref{Dynamic Extensions}, for more information.

@item extras/*
Additional non-essential files.  Currently, this directory contains some shell
startup files to be installed in @file{/etc/profile.d} to aid in manipulating
the @env{AWKPATH} and @env{AWKLIBPATH} environment variables.
@xref{Shell Startup Files}, for more information.

@item posix/*
Files needed for building @command{gawk} on POSIX-compliant systems.

@item pc/*
Files needed for building @command{gawk} under MS-Windows
(@pxref{PC Installation} for details).

@item vms/*
Files needed for building @command{gawk} under Vax/VMS and OpenVMS
(@pxref{VMS Installation} for details).

@item test/*
A test suite for
@command{gawk}.  You can use @samp{make check} from the top-level @command{gawk}
directory to run your version of @command{gawk} against the test suite.
If @command{gawk} successfully passes @samp{make check}, then you can
be confident of a successful port.
@end table

@node Unix Installation
@appendixsec Compiling and Installing @command{gawk} on Unix-Like Systems

Usually, you can compile and install @command{gawk} by typing only two
commands.  However, if you use an unusual system, you may need
to configure @command{gawk} for your system yourself.

@menu
* Quick Installation::               Compiling @command{gawk} under Unix.
* Shell Startup Files::              Shell convenience functions.
* Additional Configuration Options:: Other compile-time options.
* Configuration Philosophy::         How it's all supposed to work.
@end menu

@node Quick Installation
@appendixsubsec Compiling @command{gawk} for Unix-Like Systems

The normal installation steps should work on all modern commercial
Unix-derived systems, GNU/Linux, BSD-based systems, and the Cygwin
environment for MS-Windows.

After you have extracted the @command{gawk} distribution, @command{cd}
to @file{gawk-@value{VERSION}.@value{PATCHLEVEL}}.  As with most GNU
software, you configure @command{gawk} for your system by running the
@command{configure} program.  This program is a Bourne shell script that
is generated automatically using GNU Autoconf.
@ifnotinfo
(The Autoconf software is
described fully in
@cite{Autoconf---Generating Automatic Configuration Scripts},
which can be found online at
@uref{https://www.gnu.org/software/autoconf/manual/index.html,
the Free Software Foundation's website}.)
@end ifnotinfo
@ifinfo
(The Autoconf software is described fully starting with
@inforef{Top, , Autoconf, autoconf,Autoconf---Generating Automatic Configuration Scripts}.)
@end ifinfo

To configure @command{gawk}, simply run @command{configure}:

@example
sh ./configure
@end example

This produces a @file{Makefile} and @file{config.h} tailored to your system.
The @file{config.h} file describes various facts about your system.
You might want to edit the @file{Makefile} to
change the @code{CFLAGS} variable, which controls
the command-line options that are passed to the C compiler (such as
optimization levels or compiling for debugging).

Alternatively, you can add your own values for most @command{make}
variables on the command line, such as @code{CC} and @code{CFLAGS}, when
running @command{configure}:

@example
CC=cc CFLAGS=-g sh ./configure
@end example

@noindent
See the file @file{INSTALL} in the @command{gawk} distribution for
all the details.

After you have run @command{configure} and possibly edited the @file{Makefile},
type:

@example
make
@end example

@noindent
Shortly thereafter, you should have an executable version of @command{gawk}.
That's all there is to it!
To verify that @command{gawk} is working properly,
run @samp{make check}.  All of the tests should succeed.
If these steps do not work, or if any of the tests fail,
check the files in the @file{README_d} directory to see if you've
found a known problem.  If the failure is not described there,
send in a bug report (@pxref{Bugs}).

Of course, once you've built @command{gawk}, it is likely that you will
wish to install it.  To do so, you need to run the command @samp{make
install}, as a user with the appropriate permissions.  How to do this
varies by system, but on many systems you can use the @command{sudo}
command to do so.  The command then becomes @samp{sudo make install}. It
is likely that you will be asked for your password, and you will have
to have been set up previously as a user who is allowed to run the
@command{sudo} command.

@node Shell Startup Files
@appendixsubsec Shell Startup Files

The distribution contains shell startup files @file{gawk.sh} and
@file{gawk.csh}, containing functions to aid in manipulating
the @env{AWKPATH} and @env{AWKLIBPATH} environment variables.
On a Fedora GNU/Linux system, these files should be installed in @file{/etc/profile.d};
on other platforms, the appropriate location may be different.

@table @command

@cindex @command{gawkpath_default} shell function
@item gawkpath_default
Reset the @env{AWKPATH} environment variable to its default value.

@cindex @command{gawkpath_prepend} shell function
@item gawkpath_prepend
Add the argument to the front of the @env{AWKPATH} environment variable.

@cindex @command{gawkpath_append} shell function
@item gawkpath_append
Add the argument to the end of the @env{AWKPATH} environment variable.

@cindex @command{gawklibpath_default} shell function
@item gawklibpath_default
Reset the @env{AWKLIBPATH} environment variable to its default value.

@cindex @command{gawklibpath_prepend} shell function
@item gawklibpath_prepend
Add the argument to the front of the @env{AWKLIBPATH} environment variable.

@cindex @command{gawklibpath_append} shell function
@item gawklibpath_append
Add the argument to the end of the @env{AWKLIBPATH} environment variable.

@end table


@node Additional Configuration Options
@appendixsubsec Additional Configuration Options
@cindex @command{gawk}, configuring, options
@cindex configuration options@comma{} @command{gawk}

There are several additional options you may use on the @command{configure}
command line when compiling @command{gawk} from scratch, including:

@table @code

@cindex @option{--disable-extensions} configuration option
@cindex configuration option, @code{--disable-extensions}
@item --disable-extensions
Disable configuring and building the sample extensions in the
@file{extension} directory. This is useful for cross-compiling.
The default action is to dynamically check if the extensions
can be configured and compiled.

@cindex @option{--disable-lint} configuration option
@cindex configuration option, @code{--disable-lint}
@item --disable-lint
Disable all lint checking within @command{gawk}.  The
@option{--lint} and @option{--lint-old} options
(@pxref{Options})
are accepted, but silently do nothing.
Similarly, setting the @code{LINT} variable
(@pxref{User-modified})
has no effect on the running @command{awk} program.

When used with the GNU Compiler Collection's (GCC's)
automatic dead-code-elimination, this option
cuts almost 23K bytes off the size of the @command{gawk}
executable on GNU/Linux x86_64 systems.  Results on other systems and
with other compilers are likely to vary.
Using this option may bring you some slight performance improvement.

@quotation CAUTION
Using this option will cause some of the tests in the test suite
to fail.  This option may be removed at a later date.
@end quotation

@cindex @option{--disable-mpfr} configuration option
@cindex configuration option, @code{--disable-mpfr}
@item --disable-mpfr
Skip checking for the MPFR and GMP libraries. This is useful
mainly for the developers, to make sure nothing breaks if
MPFR support is not available.

@cindex @option{--disable-nls} configuration option
@cindex configuration option, @code{--disable-nls}
@item --disable-nls
Disable all message-translation facilities.
This is usually not desirable, but it may bring you some slight performance
improvement.

@cindex @option{--enable-versioned-extension-dir} configuration option
@cindex configuration option, @code{--enable-versioned-extension-dir}
@item --enable-versioned-extension-dir
Use a versioned directory for extensions.  The directory name will
include the major and minor API versions in it. This makes it possible
to keep extensions for different API versions on the same system
without their conflicting with one another.

@cindex @option{--with-whiny-user-strftime} configuration option
@cindex configuration option, @code{--with-whiny-user-strftime}
@item  --with-whiny-user-strftime
Force use of the included version of the C @code{strftime()}
function for deficient systems.
@end table

Use the command @samp{./configure --help} to see the full list of
options supplied by @command{configure}.

@node Configuration Philosophy
@appendixsubsec The Configuration Process

@cindex @command{gawk}, configuring
This @value{SECTION} is of interest only if you know something about using the
C language and Unix-like operating systems.

The source code for @command{gawk} generally attempts to adhere to formal
standards wherever possible.  This means that @command{gawk} uses library
routines that are specified by the ISO C standard and by the POSIX
operating system interface standard.
The @command{gawk} source code requires using an ISO C compiler (the 1990
standard).

Many Unix systems do not support all of either the ISO or the
POSIX standards.  The @file{missing_d} subdirectory in the @command{gawk}
distribution contains replacement versions of those functions that are
most likely to be missing.

The @file{config.h} file that @command{configure} creates contains
definitions that describe features of the particular operating system
where you are attempting to compile @command{gawk}.  The three things
described by this file are: what header files are available, so that
they can be correctly included, what (supposedly) standard functions
are actually available in your C libraries, and various miscellaneous
facts about your operating system.  For example, there may not be an
@code{st_blksize} element in the @code{stat} structure.  In this case,
@samp{HAVE_STRUCT_STAT_ST_BLKSIZE} is undefined.

@cindex @code{custom.h} file
It is possible for your C compiler to lie to @command{configure}. It may
do so by not exiting with an error when a library function is not
available.  To get around this, edit the @file{custom.h} file.
Use an @samp{#ifdef} that is appropriate for your system, and either
@code{#define} any constants that @command{configure} should have defined but
didn't, or @code{#undef} any constants that @command{configure} defined and
should not have.  The @file{custom.h} file is automatically included by
the @file{config.h} file.

It is also possible that the @command{configure} program generated by
Autoconf will not work on your system in some other fashion.
If you do have a problem, the @file{configure.ac} file is the input for
Autoconf.  You may be able to change this file and generate a
new version of @command{configure} that works on your system
(@pxref{Bugs}
for information on how to report problems in configuring @command{gawk}).
The same mechanism may be used to send in updates to @file{configure.ac}
and/or @file{custom.h}.

@node Non-Unix Installation
@appendixsec Installation on Other Operating Systems

This @value{SECTION} describes how to install @command{gawk} on
various non-Unix systems.

@menu
* PC Installation::             Installing and Compiling @command{gawk} on
                                Microsoft Windows.
* VMS Installation::            Installing @command{gawk} on VMS.
@end menu

@node PC Installation
@appendixsubsec Installation on MS-Windows

@cindex PC operating systems@comma{} @command{gawk} on, installing
@cindex operating systems, PC@comma{} @command{gawk} on, installing
This @value{SECTION} covers installation and usage of @command{gawk}
on Intel architecture machines running any version of MS-Windows.
In this @value{SECTION}, the term ``Windows32''
refers to any of Microsoft Windows 95/98/ME/NT/2000/XP/Vista/7/8/10.

See also the @file{README_d/README.pc} file in the distribution.

@menu
* PC Binary Installation::      Installing a prepared distribution.
* PC Compiling::                Compiling @command{gawk} for Windows32.
* PC Using::                    Running @command{gawk} on Windows32.
* Cygwin::                      Building and running @command{gawk} for
                                Cygwin.
* MSYS::                        Using @command{gawk} In The MSYS Environment.
@end menu

@node PC Binary Installation
@appendixsubsubsec Installing a Prepared Distribution for MS-Windows Systems

The only supported binary distribution for MS-Windows systems
is that provided by Eli Zaretskii's @uref{https://sourceforge.net/projects/ezwinports/,
``ezwinports''} project.  Install the compiled @command{gawk} from there.

@node PC Compiling
@appendixsubsubsec Compiling @command{gawk} for PC Operating Systems

@command{gawk} can be compiled for Windows32 using MinGW (Windows32).
The file @file{README_d/README.pc} in the @command{gawk} distribution
contains additional notes, and @file{pc/Makefile} contains important
information on compilation options.

@cindex compiling @command{gawk} for MS-Windows
To build @command{gawk} for Windows32, copy the files in
the @file{pc} directory (@emph{except} for @file{ChangeLog}) to the
directory with the rest of the @command{gawk} sources, then invoke
@command{make} with the appropriate target name as an argument to
build @command{gawk}.  The @file{Makefile} copied from the @file{pc}
directory contains a configuration section with comments and may need
to be edited in order to work with your @command{make} utility.

The @file{Makefile} supports a number of targets for building various
MS-DOS and Windows32 versions.  A list of targets is printed if the
@command{make} command is given without a target.  As an example,
to build a native MS-Windows binary of @command{gawk} using the MinGW tools,
type @samp{make mingw32}.

@node PC Using
@appendixsubsubsec Using @command{gawk} on PC Operating Systems
@cindex operating systems, PC, @command{gawk} on
@cindex PC operating systems, @command{gawk} on

Information in this section applies to the MinGW and
DJGPP ports of @command{gawk}. @xref{Cygwin} for information
about the Cygwin port.

Under MS-Windows, the MinGW environment supports
both the @samp{|&} operator and TCP/IP networking
(@pxref{TCP/IP Networking}).
The DJGPP environment does not support @samp{|&}.

@cindex search paths
@cindex search paths, for source files
@cindex @command{gawk}, MS-Windows version of
@cindex @code{;} (semicolon), @env{AWKPATH} variable and
@cindex semicolon (@code{;}), @env{AWKPATH} variable and
@cindex @env{AWKPATH} environment variable
The MS-Windows version of @command{gawk} searches for
program files as described in @ref{AWKPATH Variable}.  However,
semicolons (rather than colons) separate elements in the @env{AWKPATH}
variable.  If @env{AWKPATH} is not set or is empty, then the default
search path is @samp{@w{.;c:/lib/awk;c:/gnu/lib/awk}}.

@cindex common extensions, @code{BINMODE} variable
@cindex extensions, common@comma{} @code{BINMODE} variable
@cindex differences in @command{awk} and @command{gawk}, @code{BINMODE} variable
@cindex @code{BINMODE} variable
Under MS-Windows,
@command{gawk} (and many other text programs) silently
translates end-of-line @samp{\r\n} to @samp{\n} on input and @samp{\n}
to @samp{\r\n} on output.  A special @code{BINMODE} variable @value{COMMONEXT}
allows control over these translations and is interpreted as follows:

@itemize @value{BULLET}
@item
If @code{BINMODE} is @code{"r"} or one,
then
binary mode is set on read (i.e., no translations on reads).

@item
If @code{BINMODE} is @code{"w"} or two,
then
binary mode is set on write (i.e., no translations on writes).

@item
If @code{BINMODE} is @code{"rw"} or @code{"wr"} or three,
binary mode is set for both read and write.

@item
@code{BINMODE=@var{non-null-string}} is
the same as @samp{BINMODE=3} (i.e., no translations on
reads or writes).  However, @command{gawk} issues a warning
message if the string is not one of @code{"rw"} or @code{"wr"}.
@end itemize

@noindent
The modes for standard input and standard output are set one time
only (after the
command line is read, but before processing any of the @command{awk} program).
Setting @code{BINMODE} for standard input or
standard output is accomplished by using an
appropriate @samp{-v BINMODE=@var{N}} option on the command line.
@code{BINMODE} is set at the time a file or pipe is opened and cannot be
changed midstream.

The name @code{BINMODE} was chosen to match @command{mawk}
(@pxref{Other Versions}).
@command{mawk} and @command{gawk} handle @code{BINMODE} similarly; however,
@command{mawk} adds a @samp{-W BINMODE=@var{N}} option and an environment
variable that can set @code{BINMODE}, @code{RS}, and @code{ORS}.  The
files @file{binmode[1-3].awk} (under @file{gnu/lib/awk} in some of the
prepared binary distributions) have been chosen to match @command{mawk}'s @samp{-W
BINMODE=@var{N}} option.  These can be changed or discarded; in particular,
the setting of @code{RS} giving the fewest ``surprises'' is open to debate.
@command{mawk} uses @samp{RS = "\r\n"} if binary mode is set on read, which is
appropriate for files with the MS-DOS-style end-of-line.

To illustrate, the following examples set binary mode on writes for standard
output and other files, and set @code{ORS} as the ``usual'' MS-DOS-style
end-of-line:

@example
gawk -v BINMODE=2 -v ORS="\r\n" @dots{}
@end example

@noindent
or:

@example
gawk -v BINMODE=w -f binmode2.awk @dots{}
@end example

@noindent
These give the same result as the @samp{-W BINMODE=2} option in
@command{mawk}.
The following changes the record separator to @code{"\r\n"} and sets binary
mode on reads, but does not affect the mode on standard input:

@example
gawk -v RS="\r\n" -e "BEGIN @{ BINMODE = 1 @}" @dots{}
@end example

@noindent
or:

@example
gawk -f binmode1.awk @dots{}
@end example

@noindent
With proper quoting, in the first example the setting of @code{RS} can be
moved into the @code{BEGIN} rule.

@node Cygwin
@appendixsubsubsec Using @command{gawk} In The Cygwin Environment
@cindex compiling @command{gawk} for Cygwin

@command{gawk} can be built and used ``out of the box'' under MS-Windows
if you are using the @uref{http://www.cygwin.com, Cygwin environment}.
This environment provides an excellent simulation of GNU/Linux, using
Bash, GCC, GNU Make,
and other GNU programs.  Compilation and installation for Cygwin is the
same as for a Unix system:

@example
tar -xvpzf gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz
cd gawk-@value{VERSION}.@value{PATCHLEVEL}
./configure
make && make check
@end example

When compared to GNU/Linux on the same system, the @samp{configure}
step on Cygwin takes considerably longer.  However, it does finish,
and then the @samp{make} proceeds as usual.

Recent versions of Cygwin open all files in binary mode. This means
that you should use @samp{RS = "\r?\n"} in order to be able to
handle standard MS-Windows text files with carriage-return plus
line-feed line endings.

The Cygwin environment supports
both the @samp{|&} operator and TCP/IP networking
(@pxref{TCP/IP Networking}).

@node MSYS
@appendixsubsubsec Using @command{gawk} In The MSYS Environment

In the MSYS environment under MS-Windows, @command{gawk} automatically
uses binary mode for reading and writing files.  Thus, there is no
need to use the @code{BINMODE} variable.

This can cause problems with other Unix-like components that have
been ported to MS-Windows that expect @command{gawk} to do automatic
translation of @code{"\r\n"}, because it won't.

@node VMS Installation
@appendixsubsec Compiling and Installing @command{gawk} on Vax/VMS and OpenVMS

@c based on material from Pat Rankin <rankin@eql.caltech.edu>
@c now rankin@pactechdata.com
@c now r.pat.rankin@gmail.com

@cindex @command{gawk}, VMS version of
@cindex installation, VMS
This @value{SUBSECTION} describes how to compile and install @command{gawk} under VMS.
The older designation ``VMS'' is used throughout to refer to OpenVMS.

@menu
* VMS Compilation::             How to compile @command{gawk} under VMS.
* VMS Dynamic Extensions::      Compiling @command{gawk} dynamic extensions on
                                VMS.
* VMS Installation Details::    How to install @command{gawk} under VMS.
* VMS Running::                 How to run @command{gawk} under VMS.
* VMS GNV::                     The VMS GNV Project.
* VMS Old Gawk::                An old version comes with some VMS systems.
@end menu

@node VMS Compilation
@appendixsubsubsec Compiling @command{gawk} on VMS
@cindex compiling @command{gawk} for VMS

To compile @command{gawk} under VMS, there is a @code{DCL} command procedure
that issues all the necessary @code{CC} and @code{LINK} commands. There is
also a @file{Makefile} for use with the @code{MMS} and @code{MMK} utilities.
From the source directory, use either:

@example
$ @kbd{@@[.vms]vmsbuild.com}
@end example

@noindent
or:

@example
$ @kbd{MMS/DESCRIPTION=[.vms]descrip.mms gawk}
@end example

@noindent
or:

@example
$ @kbd{MMK/DESCRIPTION=[.vms]descrip.mms gawk}
@end example

@command{MMK} is an open source, free, near-clone of @command{MMS} and
can better handle ODS-5 volumes with upper- and lowercase @value{FN}s.
@command{MMK} is available from @uref{https://github.com/endlesssoftware/mmk}.

With ODS-5 volumes and extended parsing enabled, the case of the target
parameter may need to be exact.

@command{gawk} has been tested under VAX/VMS 7.3 and Alpha/VMS 7.3-1
using Compaq C V6.4, and under Alpha/VMS 7.3, Alpha/VMS 7.3-2, and IA64/VMS 8.3.
The most recent builds used HP C V7.3 on Alpha VMS 8.3 and both
Alpha and IA64 VMS 8.4 used HP C 7.3.@footnote{The IA64 architecture
is also known as ``Itanium.''}

@xref{VMS GNV} for information on building
@command{gawk} as a PCSI kit that is compatible with the GNV product.

@node VMS Dynamic Extensions
@appendixsubsubsec Compiling @command{gawk} Dynamic Extensions on VMS

The extensions that have been ported to VMS can be built using one of
the following commands:

@example
$ @kbd{MMS/DESCRIPTION=[.vms]descrip.mms extensions}
@end example

@noindent
or:

@example
$ @kbd{MMK/DESCRIPTION=[.vms]descrip.mms extensions}
@end example

@command{gawk} uses @code{AWKLIBPATH} as either an environment variable
or a logical name to find the dynamic extensions.

Dynamic extensions need to be compiled with the same compiler options for
floating-point, pointer size, and symbol name handling as were used
to compile @command{gawk} itself.
Alpha and Itanium should use IEEE floating point.  The pointer size is 32 bits,
and the symbol name handling should be exact case with CRC shortening for
symbols longer than 32 bits.

For Alpha and Itanium:

@example
/name=(as_is,short)
/float=ieee/ieee_mode=denorm_results
@end example

For VAX:

@example
/name=(as_is,short)
@end example

Compile-time macros need to be defined before the first VMS-supplied
header file is included, as follows:

@example
#if (__CRTL_VER >= 70200000) && !defined (__VAX)
#define _LARGEFILE 1
#endif

#ifndef __VAX
#ifdef __CRTL_VER
#if __CRTL_VER >= 80200000
#define _USE_STD_STAT 1
#endif
#endif
#endif
@end example

If you are writing your own extensions to run on VMS, you must supply these
definitions yourself. The @file{config.h} file created when building @command{gawk}
on VMS does this for you; if instead you use that file or a similar one, then you
must remember to include it before any VMS-supplied header files.

@node VMS Installation Details
@appendixsubsubsec Installing @command{gawk} on VMS

To use @command{gawk}, all you need is a ``foreign'' command, which is a
@code{DCL} symbol whose value begins with a dollar sign. For example:

@example
$ @kbd{GAWK :== $disk1:[gnubin]gawk}
@end example

@noindent
Substitute the actual location of @command{gawk.exe} for
@samp{$disk1:[gnubin]}. The symbol should be placed in the
@file{login.com} of any user who wants to run @command{gawk},
so that it is defined every time the user logs on.
Alternatively, the symbol may be placed in the system-wide
@file{sylogin.com} procedure, which allows all users
to run @command{gawk}.

If your @command{gawk} was installed by a PCSI kit into the
@file{GNV$GNU:} directory tree, the program will be known as
@file{GNV$GNU:[bin]gnv$gawk.exe} and the help file will be
@file{GNV$GNU:[vms_help]gawk.hlp}.

The PCSI kit also installs a @file{GNV$GNU:[vms_bin]gawk_verb.cld} file
that can be used to add @command{gawk} and @command{awk} as DCL commands.

For just the current process you can use:

@example
$ @kbd{set command gnv$gnu:[vms_bin]gawk_verb.cld}
@end example

Or the system manager can use @file{GNV$GNU:[vms_bin]gawk_verb.cld} to
add the @command{gawk} and @command{awk} to the system-wide @samp{DCLTABLES}.

The DCL syntax is documented in the @file{gawk.hlp} file.

Optionally, the @file{gawk.hlp} entry can be loaded into a VMS help library:

@example
$ @kbd{LIBRARY/HELP sys$help:helplib [.vms]gawk.hlp}
@end example

@noindent
(You may want to substitute a site-specific help library rather than
the standard VMS library @samp{HELPLIB}.)  After loading the help text,
the command:

@example
$ @kbd{HELP GAWK}
@end example

@noindent
provides information about both the @command{gawk} implementation and the
@command{awk} programming language.

The logical name @samp{AWK_LIBRARY} can designate a default location
for @command{awk} program files.  For the @option{-f} option, if the specified
@value{FN} has no device or directory path information in it, @command{gawk}
looks in the current directory first, then in the directory specified
by the translation of @samp{AWK_LIBRARY} if the file is not found.
If, after searching in both directories, the file still is not found,
@command{gawk} appends the suffix @samp{.awk} to the @value{FN} and retries
the file search.  If @samp{AWK_LIBRARY} has no definition, a default value
of @samp{SYS$LIBRARY:} is used for it.

@node VMS Running
@appendixsubsubsec Running @command{gawk} on VMS

Command-line parsing and quoting conventions are significantly different
on VMS, so examples in this @value{DOCUMENT} or from other sources often need minor
changes.  They @emph{are} minor though, and all @command{awk} programs
should run correctly.

Here are a couple of trivial tests:

@example
$ @kbd{gawk -- "BEGIN @{print ""Hello, World!""@}"}
$ @kbd{gawk -"W" version}
! could also be -"W version" or "-W version"
@end example

@noindent
Note that uppercase and mixed-case text must be quoted.

The VMS port of @command{gawk} includes a @code{DCL}-style interface in addition
to the original shell-style interface (see the help entry for details).
One side effect of dual command-line parsing is that if there is only a
single parameter (as in the quoted string program), the command
becomes ambiguous.  To work around this, the normally optional @option{--}
flag is required to force Unix-style parsing rather than @code{DCL} parsing.
If any other dash-type options (or multiple parameters such as @value{DF}s to
process) are present, there is no ambiguity and @option{--} can be omitted.

@cindex exit status, of VMS
The @code{exit} value is a Unix-style value and is encoded into a VMS exit
status value when the program exits.

The VMS severity bits will be set based on the @code{exit} value.
A failure is indicated by 1, and VMS sets the @code{ERROR} status.
A fatal error is indicated by 2, and VMS sets the @code{FATAL} status.
All other values will have the @code{SUCCESS} status.  The exit value is
encoded to comply with VMS coding standards and will have the
@code{C_FACILITY_NO} of @code{0x350000} with the constant @code{0xA000}
added to the number shifted over by 3 bits to make room for the severity codes.

To extract the actual @command{gawk} exit code from the VMS status, use:

@example
unix_status = (vms_status .and. %x7f8) / 8
@end example

@noindent
A C program that uses @code{exec()} to call @command{gawk} will get the original
Unix-style exit value.

Older versions of @command{gawk} for VMS treated a Unix exit code 0 as 1,
a failure as 2, a fatal error as 4, and passed all the other numbers through.
This violated the VMS exit status coding requirements.

@cindex floating-point, VAX/VMS
VAX/VMS floating point uses unbiased rounding. @xref{Round Function}.

VMS reports time values in GMT unless one of the @code{SYS$TIMEZONE_RULE}
or @code{TZ} logical names is set.  Older versions of VMS, such as VAX/VMS
7.3, do not set these logical names.

@c @cindex directory search
@c @cindex path, search
@cindex search paths
@cindex search paths, for source files
The default search path, when looking for @command{awk} program files specified
by the @option{-f} option, is @code{"SYS$DISK:[],AWK_LIBRARY:"}.  The logical
name @env{AWKPATH} can be used to override this default.  The format
of @env{AWKPATH} is a comma-separated list of directory specifications.
When defining it, the value should be quoted so that it retains a single
translation and not a multitranslation @code{RMS} searchlist.

@cindex redirection on VMS

This restriction also applies to running @command{gawk} under GNV,
as redirection is always to a DCL command.

If you are redirecting data to a VMS command or utility, the current
implementation requires that setting up a VMS foreign command that runs
a command file before invoking @command{gawk}.
(This restriction may be removed in a future release of @command{gawk} on VMS.)

Without this command file, the input data will also appear prepended
to the output data.

This also allows simulating POSIX commands that are not found on VMS or the
use of GNV utilities.

The example below is for @command{gawk} redirecting data to the VMS
@command{sort} command.

@example
$ sort = "@@device:[dir]vms_gawk_sort.com"
@end example

The command file needs to be of the format in the example below.

The first line inhibits the passed input data from also showing up in the
output.  It must be in the format in the example.

The next line creates a foreign command that overrides the outer foreign
command which prevents an infinite recursion of command files.

The next to the last command redirects @code{sys$input} to be
@code{sys$command}, in order to pick up the data that is being redirected
to the command.

The last line runs the actual command.  It must be the last command as the data
redirected from @command{gawk} will be read when the command file ends.

@example
$!'f$verify(0,0)'
$ sort := sort
$ define/user sys$input sys$command:
$ sort sys$input: sys$output:
@end example


@node VMS GNV
@appendixsubsubsec The VMS GNV Project

The VMS GNV package provides a build environment similar to POSIX with ports
of a collection of open source tools.  The @command{gawk} found in the GNV
base kit is an older port.  Currently, the GNV project is being reorganized
to supply individual PCSI packages for each component.
See @w{@uref{https://sourceforge.net/p/gnv/wiki/InstallingGNVPackages/}.}

The normal build procedure for @command{gawk} produces a program that
is suitable for use with GNV.

The file @file{vms/gawk_build_steps.txt} in the distribution documents
the procedure for building a VMS PCSI kit that is compatible with GNV.

@ignore
@c The VMS POSIX product, also known as POSIX for OpenVMS, is long defunct
@c and building gawk for it has not been tested in many years, but these
@c old instructions might still work if anyone is still using it.

@node VMS POSIX
@appendixsubsubsec Building and Using @command{gawk} on VMS POSIX

Ignore the instructions above, although @file{vms/gawk.hlp} should still
be made available in a help library.  The source tree should be unpacked
into a container file subsystem rather than into the ordinary VMS filesystem.
Make sure that the two scripts, @file{configure} and
@file{vms/posix-cc.sh}, are executable; use @samp{chmod +x} on them if
necessary.  Then execute the following two commands:

@example
psx> @kbd{CC=vms/posix-cc.sh configure}
psx> @kbd{make CC=c89 gawk}
@end example

@noindent
The first command constructs files @file{config.h} and @file{Makefile} out
of templates, using a script to make the C compiler fit @command{configure}'s
expectations.  The second command compiles and links @command{gawk} using
the C compiler directly; ignore any warnings from @command{make} about being
unable to redefine @code{CC}.  @command{configure} takes a very long
time to execute, but at least it provides incremental feedback as it runs.

This has been tested with VAX/VMS V6.2, VMS POSIX V2.0, and DEC C V5.2.

Once built, @command{gawk} works like any other shell utility.  Unlike
the normal VMS port of @command{gawk}, no special command-line manipulation is
needed in the VMS POSIX environment.
@end ignore

@node VMS Old Gawk
@appendixsubsubsec Some VMS Systems Have An Old Version of @command{gawk}

@c Thanks to "gerard labadie" <gerard.labadie@gmail.com>

Some versions of VMS have an old version of @command{gawk}.  To access it,
define a symbol, as follows:

@example
$ @kbd{gawk :== $sys$common:[syshlp.examples.tcpip.snmp]gawk.exe}
@end example

This is apparently @value{PVERSION} 2.15.6, which is extremely old. We
recommend compiling and using the current version.

@node Bugs
@appendixsec Reporting Problems and Bugs
@cindex archaeologists
@quotation
@i{There is nothing more dangerous than a bored archaeologist.}
@author Douglas Adams, @cite{The Hitchhiker's Guide to the Galaxy}
@end quotation
@c the radio show, not the book. :-)

@cindex debugging @command{gawk}, bug reports
@cindex troubleshooting, @command{gawk}, bug reports
If you have problems with @command{gawk} or think that you have found a bug,
report it to the developers; we cannot promise to do anything,
but we might well want to fix it.

@menu
* Bug address::                 Where to send reports to.
* Usenet::                      Where not to send reports to.
* Maintainers::                 Maintainers of non-*nix ports.
@end menu

@node Bug address
@appendixsubsec Submitting Bug Reports

Before reporting a bug, make sure you have really found a genuine bug.
First, verify that you have the latest version of @command{gawk}.
Many bugs (usually subtle ones) are fixed at each release, and if yours
is out of date, the problem may already have been solved.

Second, please see if setting the environment variable @env{LC_ALL}
to @code{LC_ALL=C} causes things to behave as you expect. If so, it's
a locale issue, and may or may not really be a bug.

Third, carefully reread the documentation and see if it says you can do
what you're trying to do.  If it's not clear whether you should be able
to do something or not, report that too; it's a bug in the documentation!

Finally, before reporting a bug or trying to fix it yourself, try to isolate it
to the smallest possible @command{awk} program and input @value{DF} that
reproduce the problem.  Then send us the program and @value{DF},
some idea of what kind of Unix system you're using,
the compiler you used to compile @command{gawk}, and the exact results
@command{gawk} gave you.  Also say what you expected to occur; this helps
us decide whether the problem is really in the documentation.

Make sure to include the version number of @command{gawk} you are using.
You can get this information with the command @samp{gawk --version}.

@cindex @code{bug-gawk@@gnu.org} bug reporting address
@cindex email address for bug reports, @code{bug-gawk@@gnu.org}
@cindex bug reports, email address, @code{bug-gawk@@gnu.org}
Once you have a precise problem description, send email to
@EMAIL{bug-gawk@@gnu.org,bug-gawk at gnu dot org}.

The @command{gawk} maintainers subscribe to this address, and
thus they will receive your bug report.
Although you can send mail to the maintainers directly,
the bug reporting address is preferred because the
email list is archived at the GNU Project.
@emph{All email must be in English. This is the only language
understood in common by all the maintainers.}
In addition, please be sure to send all mail in @emph{plain text},
not (or not exclusively) in HTML.

@quotation NOTE
Many distributions of GNU/Linux and the various BSD-based operating systems
have their own bug reporting systems.  If you report a bug using your distribution's
bug reporting system, you should also send a copy to
@EMAIL{bug-gawk@@gnu.org,bug-gawk at gnu dot org}.

This is for two reasons.  First, although some distributions forward
bug reports ``upstream'' to the GNU mailing list, many don't, so there is a good
chance that the @command{gawk}  maintainers won't even see the bug report!  Second,
mail to the GNU list is archived, and having everything at the GNU Project
keeps things self-contained and not dependent on other organizations.
@end quotation

Non-bug suggestions are always welcome as well.  If you have questions
about things that are unclear in the documentation or are just obscure
features, ask on the bug list; we will try to help you out if we can.

@node Usenet
@appendixsubsec Please Don't Post Bug Reports to USENET

@quotation
@c Date: Sun, 17 May 2015 19:50:14 -0400
@c From: Chet Ramey <chet.ramey@case.edu>
@c Reply-To: chet.ramey@case.edu
@c Organization: ITS, Case Western Reserve University
@c To: Aharon Robbins <arnold@skeeve.com>
@c CC: chet.ramey@case.edu
I gave up on Usenet a couple of years ago and haven't really looked back.
It's like sports talk radio---you feel smarter for not having read it.
@author Chet Ramey
@end quotation

@cindex @code{comp.lang.awk} newsgroup
Please do @emph{not} try to report bugs in @command{gawk} by posting to the
Usenet/Internet newsgroup @code{comp.lang.awk}.  Although some of the
@command{gawk} developers occasionally read this news group, the primary
@command{gawk} maintainer no longer does.  Thus it's virtually guaranteed
that he will @emph{not} see your posting.  The steps described here are
the only officially recognized way for reporting bugs.  Really.

@ignore
And another one:

Date: Thu, 11 Jun 2015 09:00:56 -0400
From: Chet Ramey <chet.ramey@case.edu>

My memory was imperfect.  Back in June 2009, I wrote:

"That's the nice thing about open source, right?  You can take your ball
and run to another section of the playground.  Then, if you like mixing
metaphors, you can throw rocks from there."
@end ignore

@node Maintainers
@appendixsubsec Reporting Problems with Non-Unix Ports

If you find bugs in one of the non-Unix ports of @command{gawk},
send an email to the bug list, with a copy to the
person who maintains that port.  The maintainers are named in the following list,
as well as in the @file{README} file in the @command{gawk} distribution.
Information in the @file{README} file should be considered authoritative
if it conflicts with this @value{DOCUMENT}.

The people maintaining the various @command{gawk} ports are:

@c put the index entries outside the table, for docbook
@cindex Buening, Andreas
@cindex Malmberg, John
@cindex Pitts, Dave
@cindex G., Daniel Richard
@cindex Robbins, Arnold
@cindex Zaretskii, Eli
@cindex Guerrero, Juan Manuel
@multitable {MS-Windows with MinGW} {123456789012345678901234567890123456789001234567890}
@item Unix and POSIX systems @tab Arnold Robbins, @EMAIL{arnold@@skeeve.com,arnold at skeeve dot com}

@item MS-DOS with DJGPP @tab Juan Manuel Guerrero, @EMAIL{juan.guerrero@@gmx.de, juan dot guerrero at gmx dot de}

@item MS-Windows with MinGW @tab Eli Zaretskii, @EMAIL{eliz@@gnu.org,eliz at gnu dot org}

@c Leave this in the document on purpose.
@c OS/2 is not mentioned anywhere else though.
@item OS/2 @tab Andreas Buening, @EMAIL{andreas.buening@@nexgo.de,andreas dot buening at nexgo dot de}

@item VMS @tab John Malmberg, @EMAIL{wb8tyw@@qsl.net,wb8tyw at qsl.net}

@item z/OS (OS/390) @tab Daniel Richard G.@: @EMAIL{skunk@@iSKUNK.ORG,skunk at iSKUNK.ORG}
@item  @tab Dave Pitts (Maintainer Emeritus), @EMAIL{dpitts@@cozx.com,dpitts at cozx dot com}
@end multitable

If your bug is also reproducible under Unix, send a copy of your
report to the @EMAIL{bug-gawk@@gnu.org,bug-gawk at gnu dot org} email list as well.

@node Other Versions
@appendixsec Other Freely Available @command{awk} Implementations
@cindex @command{awk}, implementations
@ignore
From: emory!amc.com!brennan (Michael Brennan)
Subject: C++ comments in awk programs
To: arnold@gnu.ai.mit.edu (Arnold Robbins)
Date: Wed, 4 Sep 1996 08:11:48 -0700 (PDT)

@end ignore
@cindex Brennan, Michael
@ifnotdocbook
@quotation
@i{It's kind of fun to put comments like this in your awk code:}@*
@ @ @ @ @ @ @code{// Do C++ comments work? answer: yes! of course}
@author Michael Brennan
@end quotation
@end ifnotdocbook

@docbook
<blockquote><attribution>Michael Brennan</attribution>
<literallayout><emphasis>It's kind of fun to put comments like this in your awk code.</emphasis>
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<literal>// Do C++ comments work? answer: yes! of course</literal></literallayout>
</blockquote>
@end docbook

There are a number of other freely available @command{awk} implementations.
This @value{SECTION} briefly describes where to get them:

@table @asis
@cindex Kernighan, Brian
@cindex source code, Brian Kernighan's @command{awk}
@cindex @command{awk}, versions of, See Also Brian Kernighan's @command{awk}
@cindex Brian Kernighan's @command{awk}, source code
@item Unix @command{awk}
Brian Kernighan, one of the original designers of Unix @command{awk},
has made his implementation of
@command{awk} freely available.
You can retrieve this version via
@uref{http://www.cs.princeton.edu/~bwk, his home page}.
It is available in several archive formats:

@table @asis
@item Shell archive
@uref{http://www.cs.princeton.edu/~bwk/btl.mirror/awk.shar}

@item Compressed @command{tar} file
@uref{http://www.cs.princeton.edu/~bwk/btl.mirror/awk.tar.gz}

@item Zip file
@uref{http://www.cs.princeton.edu/~bwk/btl.mirror/awk.zip}
@end table

@cindex @command{git} utility
You can also retrieve it from GitHub:

@example
git clone git://github.com/onetrueawk/awk bwkawk
@end example

@noindent
This command creates a copy of the @uref{https://git-scm.com, Git}
repository in a directory named @file{bwkawk}.  If you leave that argument
off the @command{git} command line, the repository copy is created in a
directory named @file{awk}.

This version requires an ISO C (1990 standard) compiler; the C compiler
from GCC (the GNU Compiler Collection) works quite nicely.

@xref{Common Extensions}
for a list of extensions in this @command{awk} that are not in POSIX @command{awk}.

As a side note, Dan Bornstein has created a Git repository tracking
all the versions of BWK @command{awk} that he could find. It's
available at @uref{git://github.com/danfuzz/one-true-awk}.

@cindex Brennan, Michael
@cindex @command{mawk} utility
@cindex source code, @command{mawk}
@item @command{mawk}
Michael Brennan wrote an independent implementation of @command{awk},
called @command{mawk}.  It is available under the
@ifclear FOR_PRINT
GPL (@pxref{Copying}),
@end ifclear
@ifset FOR_PRINT
GPL,
@end ifset
just as @command{gawk} is.

The original distribution site for the @command{mawk} source code
no longer has it.  A copy is available at
@uref{http://www.skeeve.com/gawk/mawk1.3.3.tar.gz}.

In 2009, Thomas Dickey took on @command{mawk} maintenance.
Basic information is available on
@uref{http://www.invisible-island.net/mawk, the project's web page}.
The download URL is
@url{http://invisible-island.net/datafiles/release/mawk.tar.gz}.

Once you have it,
@command{gunzip} may be used to decompress this file. Installation
is similar to @command{gawk}'s
(@pxref{Unix Installation}).

@xref{Common Extensions}
for a list of extensions in @command{mawk} that are not in POSIX @command{awk}.

@item @command{mawk} 2.0
In 2016, Michael Brennan resumed @command{mawk} development.
His development snapshots are available via Git from the project's
@uref{https://github.com/mikebrennan000/mawk-2, GitHub page}.

@cindex Sumner, Andrew
@cindex @command{awka} compiler for @command{awk}
@cindex source code, @command{awka}
@item @command{awka}
Written by Andrew Sumner,
@command{awka} translates @command{awk} programs into C, compiles them,
and links them with a library of functions that provide the core
@command{awk} functionality.
It also has a number of extensions.

The @command{awk} translator is released under the GPL, and the library
is under the LGPL.

To get @command{awka}, go to @url{https://sourceforge.net/projects/awka}.
@c You can reach Andrew Sumner at @email{andrew@@zbcom.net}.
@c andrewsumner@@yahoo.net

The project seems to be frozen; no new code changes have been made
since approximately 2001.

@cindex Beebe, Nelson H.F.@:
@cindex @command{pawk} (profiling version of Brian Kernighan's @command{awk})
@cindex source code, @command{pawk}
@item @command{pawk}
Nelson H.F.@: Beebe at the University of Utah has modified
BWK @command{awk} to provide timing and profiling information.
It is different from @command{gawk} with the @option{--profile} option
(@pxref{Profiling})
in that it uses CPU-based profiling, not line-count
profiling.  You may find it at either
@uref{ftp://ftp.math.utah.edu/pub/pawk/pawk-20030606.tar.gz}
or
@uref{http://www.math.utah.edu/pub/pawk/pawk-20030606.tar.gz}.

@item BusyBox @command{awk}
@cindex BusyBox Awk
@cindex source code, BusyBox Awk
BusyBox is a GPL-licensed program providing small versions of many
applications within a single executable. It is aimed at embedded systems.
It includes a full implementation of POSIX @command{awk}.  When building
it, be careful not to do @samp{make install} as it will overwrite
copies of other applications in your @file{/usr/local/bin}.  For more
information, see the @uref{https://busybox.net, project's home page}.

@cindex OpenSolaris
@cindex Solaris, POSIX-compliant @command{awk}
@cindex source code, Solaris @command{awk}
@item The OpenSolaris POSIX @command{awk}
The versions of @command{awk} in @file{/usr/xpg4/bin} and
@file{/usr/xpg6/bin} on Solaris are more or less POSIX-compliant.
They are based on the @command{awk} from Mortice Kern Systems for PCs.
We were able to make this code compile and work under GNU/Linux
with 1--2 hours of work.  Making it more generally portable (using
GNU Autoconf and/or Automake) would take more work, and this
has not been done, at least to our knowledge.

@cindex Illumos
@cindex Illumos, POSIX-compliant @command{awk}
@cindex source code, Illumos @command{awk}
The source code used to be available from the OpenSolaris website.
However, that project was ended and the website shut down.  Fortunately, the
@uref{https://wiki.illumos.org/display/illumos/illumos+Home, Illumos project}
makes this implementation available.  You can view the files one at a time from
@uref{https://github.com/joyent/illumos-joyent/blob/master/usr/src/cmd/awk_xpg4}.

@cindex @command{jawk}
@cindex Java implementation of @command{awk}
@cindex source code, @command{jawk}
@item @command{jawk}
This is an interpreter for @command{awk} written in Java. It claims
to be a full interpreter, although because it uses Java facilities
for I/O and for regexp matching, the language it supports is different
from POSIX @command{awk}.  More information is available on the
@uref{https://jawk.sourceforge.net, project's home page}.

@item Libmawk
@cindex libmawk
@cindex source code, libmawk
This is an embeddable @command{awk} interpreter derived from
@command{mawk}. For more information, see
@uref{http://repo.hu/projects/libmawk/}.

@item @code{pawk}
@cindex source code, @command{pawk} (Python version)
@cindex @code{pawk}, @command{awk}-like facilities for Python
This is a Python module that claims to bring @command{awk}-like
features to Python. See @uref{https://github.com/alecthomas/pawk}
for more information. (This is not related to Nelson Beebe's
modified version of BWK @command{awk}, described earlier.)

@item @w{QSE @command{awk}}
@cindex QSE @command{awk}
@cindex source code, QSE @command{awk}
This is an embeddable @command{awk} interpreter. For more information,
see @uref{https://code.google.com/p/qse/}. @c and @uref{http://awk.info/?tools/qse}.

@item @command{QTawk}
@cindex QuikTrim Awk
@cindex source code, QuikTrim Awk
This is an independent implementation of @command{awk} distributed
under the GPL. It has a large number of extensions over standard
@command{awk} and may not be 100% syntactically compatible with it.
See @uref{http://www.quiktrim.org/QTawk.html} for more information,
including the manual. The download link there is out of date; see
@uref{http://www.quiktrim.org/#AdditionalResources} for the latest
download link.

The project may also be frozen; no new code changes have been made
since approximately 2014.

@item Other versions
See also the ``Versions and implementations'' section of the
@uref{https://en.wikipedia.org/wiki/Awk_language#Versions_and_implementations,
Wikipedia article} on @command{awk} for information on additional versions.

@end table

@node Installation summary
@appendixsec Summary

@itemize @value{BULLET}
@item
The @command{gawk} distribution is available from the GNU Project's main
distribution site, @code{ftp.gnu.org}.  The canonical build recipe is:

@example
wget https://ftp.gnu.org/gnu/gawk/gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz
tar -xvpzf gawk-@value{VERSION}.@value{PATCHLEVEL}.tar.gz
cd gawk-@value{VERSION}.@value{PATCHLEVEL}
./configure && make && make check
@end example

@quotation NOTE
Because of the @samp{https://} URL, you may have to supply the
@option{--no-check-certificate} option to @command{wget} to download
the file.
@end quotation

@item
@command{gawk} may be built on non-POSIX systems as well. The currently
supported systems are MS-Windows using
MSYS, DJGPP, MinGW, and Cygwin,
@c OS/2,
and both Vax/VMS and OpenVMS.
Instructions for each system are included in this @value{APPENDIX}.

@item
Bug reports should be sent via email to @email{bug-gawk@@gnu.org}.
Bug reports should be in English and should include the version of @command{gawk},
how it was compiled, and a short program and @value{DF} that demonstrate
the problem.

@item
There are a number of other freely available @command{awk}
implementations.  Many are POSIX-compliant; others are less so.

@end itemize


@ifclear FOR_PRINT
@node Notes
@appendix Implementation Notes
@cindex @command{gawk}, implementation issues
@cindex implementation issues, @command{gawk}

This appendix contains information mainly of interest to implementers and
maintainers of @command{gawk}.  Everything in it applies specifically to
@command{gawk} and not to other implementations.

@menu
* Compatibility Mode::          How to disable certain @command{gawk}
                                extensions.
* Additions::                   Making Additions To @command{gawk}.
* Future Extensions::           New features that may be implemented one day.
* Implementation Limitations::  Some limitations of the implementation.
* Extension Design::            Design notes about the extension API.
* Old Extension Mechanism::     Some compatibility for old extensions.
* Notes summary::               Summary of implementation notes.
@end menu

@node Compatibility Mode
@appendixsec Downward Compatibility and Debugging
@cindex @command{gawk}, implementation issues, downward compatibility
@cindex @command{gawk}, implementation issues, debugging
@cindex troubleshooting, @command{gawk}
@cindex implementation issues@comma{} @command{gawk}, debugging

@xref{POSIX/GNU},
for a summary of the GNU extensions to the @command{awk} language and program.
All of these features can be turned off by invoking @command{gawk} with the
@option{--traditional} option or with the @option{--posix} option.

If @command{gawk} is compiled for debugging with @samp{-DDEBUG}, then there
is one more option available on the command line:

@table @code
@item -Y
@itemx --parsedebug
Print out the parse stack information as the program is being parsed.
@end table

This option is intended only for serious @command{gawk} developers
and not for the casual user.  It probably has not even been compiled into
your version of @command{gawk}, since it slows down execution.

@node Additions
@appendixsec Making Additions to @command{gawk}

If you find that you want to enhance @command{gawk} in a significant
fashion, you are perfectly free to do so.  That is the point of having
free software; the source code is available and you are free to change
it as you want (@pxref{Copying}).

This @value{SECTION} discusses the ways you might want to change @command{gawk}
as well as any considerations you should bear in mind.

@menu
* Accessing The Source::        Accessing the Git repository.
* Adding Code::                 Adding code to the main body of
                                @command{gawk}.
* New Ports::                   Porting @command{gawk} to a new operating
                                system.
* Derived Files::               Why derived files are kept in the Git
                                repository.
@end menu

@node Accessing The Source
@appendixsubsec Accessing The @command{gawk} Git Repository

As @command{gawk} is Free Software, the source code is always available.
@ref{Gawk Distribution} describes how to get and build the formal,
released versions of @command{gawk}.

@cindex @command{git} utility
However, if you want to modify @command{gawk} and contribute back your
changes, you will probably wish to work with the development version.
To do so, you will need to access the @command{gawk} source code
repository.  The code is maintained using the
@uref{https://git-scm.com, Git distributed version control system}.
You will need to install it if your system doesn't have it.
Once you have done so, use the command:

@example
git clone git://git.savannah.gnu.org/gawk.git
@end example

@noindent
This clones the @command{gawk} repository.  If you are behind a
firewall that does not allow you to use the Git native protocol, you
can still access the repository using:

@example
git clone https://git.savannah.gnu.org/r/gawk.git
@end example

Once you have made changes, you can use @samp{git diff} to produce a
patch, and send that to the @command{gawk} maintainer; see @ref{Bugs},
for how to do that.

Once upon a time there was Git--CVS gateway for use by people who could
not install Git. However, this gateway no longer works, so you may have
better luck using a more modern version control system like Bazaar,
that has a Git plug-in for working with Git repositories.

@node Adding Code
@appendixsubsec Adding New Features

@cindex adding, features to @command{gawk}
@cindex features, adding to @command{gawk}
@cindex @command{gawk}, features, adding
You are free to add any new features you like to @command{gawk}.
However, if you want your changes to be incorporated into the @command{gawk}
distribution, there are several steps that you need to take in order to
make it possible to include them:

@enumerate 1
@item
Before building the new feature into @command{gawk} itself,
consider writing it as an extension
(@pxref{Dynamic Extensions}).
If that's not possible, continue with the rest of the steps in this list.

@item
Be prepared to sign the appropriate paperwork.
In order for the FSF to distribute your changes, you must either place
those changes in the public domain and submit a signed statement to that
effect, or assign the copyright in your changes to the FSF.
Both of these actions are easy to do and @emph{many} people have done so
already. If you have questions, please contact me
(@pxref{Bugs}),
or @EMAIL{assign@@gnu.org,assign at gnu dot org}.

@item
Get the latest version.
It is much easier for me to integrate changes if they are relative to
the most recent distributed version of @command{gawk}, or better yet,
relative to the latest code in the Git repository.  If your version of
@command{gawk} is very old, I may not be able to integrate your changes at all.
(@xref{Getting},
for information on getting the latest version of @command{gawk}.)

@item
@ifnotinfo
Follow the @cite{GNU Coding Standards}.
@end ifnotinfo
@ifinfo
See @inforef{Top, , Version, standards, GNU Coding Standards}.
@end ifinfo
This document describes how GNU software should be written. If you haven't
read it, please do so, preferably @emph{before} starting to modify @command{gawk}.
(The @cite{GNU Coding Standards} are available from
the GNU Project's
@uref{https://www.gnu.org/prep/standards/, website}.
Texinfo, Info, and DVI versions are also available.)

@cindex @command{gawk}, coding style in
@item
Use the @command{gawk} coding style.
The C code for @command{gawk} follows the instructions in the
@cite{GNU Coding Standards}, with minor exceptions.  The code is formatted
using the traditional ``K&R'' style, particularly as regards to the placement
of braces and the use of TABs.  In brief, the coding rules for @command{gawk}
are as follows:

@itemize @value{BULLET}
@item
Use ANSI/ISO style (prototype) function headers when defining functions.

@item
Put the name of the function at the beginning of its own line.

@item
Use @samp{#elif} instead of nesting @samp{#if} inside @samp{#else}.

@item
Put the return type of the function, even if it is @code{int}, on the
line above the line with the name and arguments of the function.

@item
Put spaces around parentheses used in control structures
(@code{if}, @code{while}, @code{for}, @code{do}, @code{switch},
and @code{return}).

@item
Do not put spaces in front of parentheses used in function calls.

@item
Put spaces around all C operators and after commas in function calls.

@item
Do not use the comma operator to produce multiple side effects, except
in @code{for} loop initialization and increment parts, and in macro bodies.

@item
Use real TABs for indenting, not spaces.

@item
Use the ``K&R'' brace layout style.

@item
Use comparisons against @code{NULL} and @code{'\0'} in the conditions of
@code{if}, @code{while}, and @code{for} statements, as well as in the @code{case}s
of @code{switch} statements, instead of just the
plain pointer or character value.

@item
Use @code{true} and @code{false} for @code{bool} values,
the @code{NULL} symbolic constant for pointer values,
and the character constant @code{'\0'} where appropriate, instead of @code{1}
and @code{0}.

@item
Provide one-line descriptive comments for each function.

@item
Do not use the @code{alloca()} function for allocating memory off the
stack.  Its use causes more portability trouble than is worth the minor
benefit of not having to free the storage. Instead, use @code{malloc()}
and @code{free()}.

@item
Do not use comparisons of the form @samp{! strcmp(a, b)} or similar.
As Henry Spencer once said, ``@code{strcmp()} is not a boolean!''
Instead, use @samp{strcmp(a, b) == 0}.

@item
If adding new bit flag values, use explicit hexadecimal constants
(@code{0x001}, @code{0x002}, @code{0x004}, and so on) instead of
shifting one left by successive amounts (@samp{(1<<0)}, @samp{(1<<1)},
and so on).
@end itemize

@quotation NOTE
If I have to reformat your code to follow the coding style used in
@command{gawk}, I may not bother to integrate your changes at all.
@end quotation

@cindex Texinfo
@item
Update the documentation.
Along with your new code, please supply new sections and/or chapters
for this @value{DOCUMENT}.  If at all possible, please use real
Texinfo, instead of just supplying unformatted ASCII text (although
even that is better than no documentation at all).
Conventions to be followed in @cite{@value{TITLE}} are provided
after the @samp{@@bye} at the end of the Texinfo source file.
If possible, please update the @command{man} page as well.

You will also have to sign paperwork for your documentation changes.

@cindex @command{git} utility
@item
Submit changes as unified diffs.
Use @samp{diff -u -r -N} to compare
the original @command{gawk} source tree with your version.
I recommend using the GNU version of @command{diff}, or best of all,
@samp{git diff} or @samp{git format-patch}.
Send the output produced by @command{diff} to me when you
submit your changes.
(@xref{Bugs}, for the electronic mail
information.)

Using this format makes it easy for me to apply your changes to the
master version of the @command{gawk} source code (using @code{patch}).
If I have to apply the changes manually, using a text editor, I may
not do so, particularly if there are lots of changes.

@item
Include an entry for the @file{ChangeLog} file with your submission.
This helps further minimize the amount of work I have to do,
making it easier for me to accept patches.
It is simplest if you just make this part of your diff.
@end enumerate

Although this sounds like a lot of work, please remember that while you
may write the new code, I have to maintain it and support it. If it
isn't possible for me to do that with a minimum of extra work, then I
probably will not.

@node New Ports
@appendixsubsec Porting @command{gawk} to a New Operating System
@cindex portability, @command{gawk}
@cindex operating systems, porting @command{gawk} to

@cindex porting @command{gawk}
If you want to port @command{gawk} to a new operating system, there are
several steps:

@enumerate 1
@item
Follow the guidelines in
@ifinfo
@ref{Adding Code},
@end ifinfo
@ifnotinfo
the previous @value{SECTION}
@end ifnotinfo
concerning coding style, submission of diffs, and so on.

@item
Be prepared to sign the appropriate paperwork.
In order for the FSF to distribute your code, you must either place
your code in the public domain and submit a signed statement to that
effect, or assign the copyright in your code to the FSF.
Both of these actions are easy to do and @emph{many} people have done so
already. If you have questions, please contact me, or
@email{gnu@@gnu.org}.

@item
When doing a port, bear in mind that your code must coexist peacefully
with the rest of @command{gawk} and the other ports. Avoid gratuitous
changes to the system-independent parts of the code. If at all possible,
avoid sprinkling @samp{#ifdef}s just for your port throughout the
code.

If the changes needed for a particular system affect too much of the
code, I probably will not accept them.  In such a case, you can, of course,
distribute your changes on your own, as long as you comply
with the GPL
(@pxref{Copying}).

@item
A number of the files that come with @command{gawk} are maintained by other
people.  Thus, you should not change them
unless it is for a very good reason; i.e., changes are not out of the
question, but changes to these files are scrutinized extra carefully.
The files are
@file{dfa.c},
@file{dfa.h},
@file{getopt.c},
@file{getopt.h},
@file{getopt1.c},
@file{getopt_int.h},
@file{gettext.h},
@file{regcomp.c},
@file{regex.c},
@file{regex.h},
@file{regex_internal.c},
@file{regex_internal.h},
and
@file{regexec.c}.

@item
A number of other files are provided by the GNU
Autotools (Autoconf, Automake, and GNU @command{gettext}).
You should not change them either, unless it is for a very
good reason. The files are
@file{ABOUT-NLS},
@file{config.guess},
@file{config.rpath},
@file{config.sub},
@file{depcomp},
@file{INSTALL},
@file{install-sh},
@file{missing},
@file{mkinstalldirs},
@file{xalloc.h},
and
@file{ylwrap}.

@item
Be willing to continue to maintain the port.
Non-Unix operating systems are supported by volunteers who maintain
the code needed to compile and run @command{gawk} on their systems. If no-one
volunteers to maintain a port, it becomes unsupported and it may
be necessary to remove it from the distribution.

@item
Supply an appropriate @file{gawkmisc.???} file.
Each port has its own @file{gawkmisc.???} that implements certain
operating system specific functions. This is cleaner than a plethora of
@samp{#ifdef}s scattered throughout the code.  The @file{gawkmisc.c} in
the main source directory includes the appropriate
@file{gawkmisc.???} file from each subdirectory.
Be sure to update it as well.

Each port's @file{gawkmisc.???} file has a suffix reminiscent of the machine
or operating system for the port---for example, @file{pc/gawkmisc.pc} and
@file{vms/gawkmisc.vms}. The use of separate suffixes, instead of plain
@file{gawkmisc.c}, makes it possible to move files from a port's subdirectory
into the main subdirectory, without accidentally destroying the real
@file{gawkmisc.c} file.  (Currently, this is only an issue for the
PC operating system ports.)

@item
Supply a @file{Makefile} as well as any other C source and header files that are
necessary for your operating system.  All your code should be in a
separate subdirectory, with a name that is the same as, or reminiscent
of, either your operating system or the computer system.  If possible,
try to structure things so that it is not necessary to move files out
of the subdirectory into the main source directory.  If that is not
possible, then be sure to avoid using names for your files that
duplicate the names of files in the main source directory.

@item
Update the documentation.
Please write a section (or sections) for this @value{DOCUMENT} describing the
installation and compilation steps needed to compile and/or install
@command{gawk} for your system.
@end enumerate

Following these steps makes it much easier to integrate your changes
into @command{gawk} and have them coexist happily with other
operating systems' code that is already there.

In the code that you supply and maintain, feel free to use a
coding style and brace layout that suits your taste.

@node Derived Files
@appendixsubsec Why Generated Files Are Kept In Git

@cindex Git, use of for @command{gawk} source code
@c From emails written March 22, 2012, to the gawk developers list.

If you look at the @command{gawk} source in the Git
repository, you will notice that it includes files that are automatically
generated by GNU infrastructure tools, such as @file{Makefile.in} from
Automake and even @file{configure} from Autoconf.

This is different from many Free Software projects that do not store
the derived files, because that keeps the repository less cluttered,
and it is easier to see the substantive changes when comparing versions
and trying to understand what changed between commits.

However, there are several reasons why the @command{gawk} maintainer
likes to have everything in the repository.

First, because it is then easy to reproduce any given version completely,
without relying upon the availability of (older, likely obsolete, and
maybe even impossible to find) other tools.

As an extreme example, if you ever even think about trying to compile,
oh, say, the V7 @command{awk}, you will discover that not only do you
have to bootstrap the V7 @command{yacc} to do so, but you also need the
V7 @command{lex}.  And the latter is pretty much impossible to bring up
on a modern GNU/Linux system.@footnote{We tried. It was painful.}

(Or, let's say @command{gawk} 1.2 required @command{bison} whatever-it-was
in 1989 and that there was no @file{awkgram.c} file in the repository.  Is
there a guarantee that we could find that @command{bison} version? Or that
@emph{it} would build?)

If the repository has all the generated files, then it's easy to just check
them out and build. (Or @emph{easier}, depending upon how far back we go.)

And that brings us to the second (and stronger) reason why all the files
really need to be in Git.  It boils down to who do you cater
to---the @command{gawk} developer(s), or the user who just wants to check
out a version and try it out?

The @command{gawk} maintainer
wants it to be possible for any interested @command{awk} user in the
world to just clone the repository, check out the branch of interest and
build it. Without their having to have the correct version(s) of the
autotools.@footnote{There is one GNU program that is (in our opinion)
severely difficult to bootstrap from the Git repository. For
example, on the author's old (but still working) PowerPC Macintosh with
Mac OS X 10.5, it was necessary to bootstrap a ton of software, starting
with Git itself, in order to try to work with the latest code.
It's not pleasant, and especially on older systems, it's a big waste
of time.

Starting with the latest tarball was no picnic either. The maintainers
had dropped @file{.gz} and @file{.bz2} files and only distribute
@file{.tar.xz} files.  It was necessary to bootstrap @command{xz} first!}
That is the point of the @file{bootstrap.sh} file.  It touches the
various other files in the right order such that

@example
# The canonical incantation for building GNU software:
./bootstrap.sh && ./configure && make
@end example

@noindent
will @emph{just work}.

This is extremely important for the @code{master} and
@code{gawk-@var{X}.@var{Y}-stable} branches.

Further, the @command{gawk} maintainer would argue that it's also
important for the @command{gawk} developers. When he tried to check out
the @code{xgawk} branch@footnote{A branch (since removed) created by one of the other
developers that did not include the generated files.} to build it, he
couldn't. (No @file{ltmain.sh} file, and he had no idea how to create it,
and that was not the only problem.)

He felt @emph{extremely} frustrated.  With respect to that branch,
the maintainer is no different than Jane User who wants to try to build
@code{gawk-4.1-stable} or @code{master} from the repository.

Thus, the maintainer thinks that it's not just important, but critical,
that for any given branch, the above incantation @emph{just works}.

@c Added 9/2014:
A third reason to have all the files is that without them, using @samp{git
bisect} to try to find the commit that introduced a bug is exceedingly
difficult. The maintainer tried to do that on another project that
requires running bootstrapping scripts just to create @command{configure}
and so on; it was really painful. When the repository is self-contained,
using @command{git bisect} in it is very easy.

@c So - that's my reasoning and philosophy.

What are some of the consequences and/or actions to take?

@enumerate 1
@item
We don't mind that there are differing files in the different branches
as a result of different versions of the autotools.

@enumerate A
@item
It's the maintainer's job to merge them and he will deal with it.

@item
He is really good at @samp{git diff x y > /tmp/diff1 ; gvim /tmp/diff1} to
remove the diffs that aren't of interest in order to review code.
@end enumerate

@item
It would certainly help if everyone used the same versions of the GNU tools
as he does, which in general are the latest released versions of
Automake,
Autoconf,
@command{bison},
and
GNU @command{gettext}.

@ignore
If it would help if I sent out an ``I just upgraded to version x.y
of tool Z'' kind of message to this list, I can do that.  Up until
now it hasn't been a real issue since I'm the only one who's been
dorking with the configuration machinery.
@end ignore

@c @enumerate A
@c @item
Installing from source is quite easy. It's how the maintainer worked for years
(and still works).
He had @file{/usr/local/bin} at the front of his @env{PATH} and just did:

@example
wget https://ftp.gnu.org/gnu/@var{package}/@var{package}-@var{x}.@var{y}.@var{z}.tar.gz
tar -xpzvf @var{package}-@var{x}.@var{y}.@var{z}.tar.gz
cd @var{package}-@var{x}.@var{y}.@var{z}
./configure && make && make check
make install    # as root
@end example

@quotation NOTE
Because of the @samp{https://} URL, you may have to supply the
@option{--no-check-certificate} option to @command{wget} to download
the file.
@end quotation

@c @item
@ignore
These days the maintainer uses Ubuntu 12.04 which is medium current, but
he is already doing the above for Automake, Autoconf, and @command{bison}.
@end ignore

@ignore
(C. Rant: Recent Linux versions with GNOME 3 really suck. What
    are all those people thinking?  Fedora 15 was such a bust it drove
    me to Ubuntu, but Ubuntu 11.04 and 11.10 are totally unusable from
    a UI perspective. Bleah.)
@end ignore
@c @end enumerate

@ignore
@item
If someone still feels really strongly about all this, then perhaps they
can have two branches, one for their development with just the clean
changes, and one that is buildable (xgawk and xgawk-buildable, maybe).
Or, as I suggested in another mail, make commits in pairs, the first with
the "real" changes and the second with "everything else needed for
 building".
@end ignore
@end enumerate

Most of the above was originally written by the maintainer to other
@command{gawk} developers.  It raised the objection from one of
the developers ``@dots{} that anybody pulling down the source from
Git is not an end user.''

However, this is not true. There are ``power @command{awk} users''
who can build @command{gawk} (using the magic incantation shown previously)
but who can't program in C.  Thus, the major branches should be
kept buildable all the time.

It was then suggested that there be a @command{cron} job to create
nightly tarballs of ``the source.''  Here, the problem is that there
are source trees, corresponding to the various branches! So,
nightly tarballs aren't the answer, especially as the repository can go
for weeks without significant change being introduced.

Fortunately, the Git server can meet this need. For any given
branch named @var{branchname}, use:

@example
wget https://git.savannah.gnu.org/cgit/gawk.git/snapshot/gawk-@var{branchname}.tar.gz
@end example

@noindent
to retrieve a snapshot of the given branch.

@node Future Extensions
@appendixsec Probable Future Extensions
@ignore
From emory!scalpel.netlabs.com!lwall Tue Oct 31 12:43:17 1995
Return-Path: <emory!scalpel.netlabs.com!lwall>
Message-Id: <9510311732.AA28472@scalpel.netlabs.com>
To: arnold@skeeve.atl.ga.us (Arnold D. Robbins)
Subject: Re: May I quote you?
In-Reply-To: Your message of "Tue, 31 Oct 95 09:11:00 EST."
             <m0tAHPQ-00014MC@skeeve.atl.ga.us>
Date: Tue, 31 Oct 95 09:32:46 -0800
From: Larry Wall <emory!scalpel.netlabs.com!lwall>

: Greetings. I am working on the release of gawk 3.0. Part of it will be a
: thoroughly updated manual. One of the sections deals with planned future
: extensions and enhancements.  I have the following at the beginning
: of it:
:
: @cindex PERL
: @cindex Wall, Larry
: @display
: @i{AWK is a language similar to PERL, only considerably more elegant.} @*
: Arnold Robbins
: @sp 1
: @i{Hey!} @*
: Larry Wall
: @end display
:
: Before I actually release this for publication, I wanted to get your
: permission to quote you.  (Hopefully, in the spirit of much of GNU, the
: implied humor is visible... :-)

I think that would be fine.

Larry
@end ignore
@cindex Perl
@cindex Wall, Larry
@cindex Robbins, Arnold
@quotation
@i{AWK is a language similar to PERL, only considerably more elegant.}
@author Arnold Robbins
@end quotation

@quotation
@i{Hey!}
@author Larry Wall
@end quotation

The @file{TODO} file in the @code{master} branch of the @command{gawk}
Git repository lists possible future enhancements.  Some of these relate
to the source code, and others to possible new features.  Please see
that file for the list.
@xref{Additions},
if you are interested in tackling any of the projects listed there.

@node Implementation Limitations
@appendixsec Some Limitations of the Implementation

This following table describes limits of @command{gawk} on a Unix-like
system (although it is variable even then). Other systems may have
different limits.

@multitable @columnfractions .40 .60
@headitem Item @tab Limit
@item Characters in a character class @tab 2^(number of bits per byte)
@item Length of input record @tab @code{MAX_INT}
@item Length of output record @tab Unlimited
@item Length of source line @tab Unlimited
@item Number of fields in a record @tab @code{MAX_LONG}
@item Number of file redirections @tab Unlimited
@item Number of input records in one file @tab @code{MAX_LONG}
@item Number of input records total @tab @code{MAX_LONG}
@item Number of pipe redirections @tab min(number of processes per user, number of open files)
@item Numeric values @tab Double-precision floating point (if not using MPFR)
@item Size of a field @tab @code{MAX_INT}
@item Size of a literal string @tab @code{MAX_INT}
@item Size of a printf string @tab @code{MAX_INT}
@end multitable

@node Extension Design
@appendixsec Extension API Design

This @value{SECTION} documents the design of the extension API,
including a discussion of some of the history and problems that needed
to be solved.

The first version of extensions for @command{gawk} was developed in
the mid-1990s and released with @command{gawk} 3.1 in the late 1990s.
The basic mechanisms and design remained unchanged for close to 15 years,
until 2012.

The old extension mechanism used data types and functions from
@command{gawk} itself, with a ``clever hack'' to install extension
functions.

@command{gawk} included some sample extensions, of which a few were
really useful.  However, it was clear from the outset that the extension
mechanism was bolted onto the side and was not really well thought out.

@menu
* Old Extension Problems::           Problems with the old mechanism.
* Extension New Mechanism Goals::    Goals for the new mechanism.
* Extension Other Design Decisions:: Some other design decisions.
* Extension Future Growth::          Some room for future growth.
@end menu

@node Old Extension Problems
@appendixsubsec Problems With The Old Mechanism

The old extension mechanism had several problems:

@itemize @value{BULLET}
@item
It depended heavily upon @command{gawk} internals.  Any time the
@code{NODE} structure@footnote{A critical central data structure
inside @command{gawk}.} changed, an extension would have to be
recompiled. Furthermore, to really write extensions required understanding
something about @command{gawk}'s internal functions.  There was some
documentation in this @value{DOCUMENT}, but it was quite minimal.

@item
Being able to call into @command{gawk} from an extension required linker
facilities that are common on Unix-derived systems but that did
not work on MS-Windows systems; users wanting extensions on MS-Windows
had to statically link them into @command{gawk}, even though MS-Windows supports
dynamic loading of shared objects.

@item
The API would change occasionally as @command{gawk} changed; no compatibility
between versions was ever offered or planned for.
@end itemize

Despite the drawbacks, the @command{xgawk} project developers forked
@command{gawk} and developed several significant extensions. They also
enhanced @command{gawk}'s facilities relating to file inclusion and
shared object access.

A new API was desired for a long time, but only in 2012 did the
@command{gawk} maintainer and the @command{xgawk} developers finally
start working on it together.  More information about the @command{xgawk}
project is provided in @ref{gawkextlib}.

@node Extension New Mechanism Goals
@appendixsubsec Goals For A New Mechanism

Some goals for the new API were:

@itemize @value{BULLET}
@item
The API should be independent of @command{gawk} internals.  Changes in
@command{gawk} internals should not be visible to the writer of an
extension function.

@item
The API should provide @emph{binary} compatibility across @command{gawk}
releases as long as the API itself does not change.

@item
The API should enable extensions written in C or C++ to have roughly the
same ``appearance'' to @command{awk}-level code as @command{awk}
functions do. This means that extensions should have:

@itemize @value{MINUS}
@item
The ability to access function parameters.

@item
The ability to turn an undefined parameter into an array (call by reference).

@item
The ability to create, access and update global variables.

@item
Easy access to all the elements of an array at once (``array flattening'')
in order to loop over all the element in an easy fashion for C code.

@item
The ability to create arrays (including @command{gawk}'s true
arrays of arrays).
@end itemize
@end itemize

Some additional important goals were:

@itemize @value{BULLET}
@item
The API should use only features in ISO C 90, so that extensions
can be written using the widest range of C and C++ compilers. The header
should include the appropriate @samp{#ifdef __cplusplus} and @samp{extern "C"}
magic so that a C++ compiler could be used.  (If using C++, the runtime
system has to be smart enough to call any constructors and destructors,
as @command{gawk} is a C program. As of this writing, this has not been
tested.)

@item
The API mechanism should not require access to @command{gawk}'s
symbols@footnote{The @dfn{symbols} are the variables and functions
defined inside @command{gawk}.  Access to these symbols by code
external to @command{gawk} loaded dynamically at runtime is
problematic on MS-Windows.} by the compile-time or dynamic linker,
in order to enable creation of extensions that also work on MS-Windows.
@end itemize

During development, it became clear that there were other features
that should be available to extensions, which were also subsequently
provided:

@itemize @value{BULLET}
@item
Extensions should have the ability to hook into @command{gawk}'s
I/O redirection mechanism.  In particular, the @command{xgawk}
developers provided a so-called ``open hook'' to take over reading
records.  During development, this was generalized to allow
extensions to hook into input processing, output processing, and
two-way I/O.

@item
An extension should be able to provide a ``call back'' function
to perform cleanup actions when @command{gawk} exits.

@item
An extension should be able to provide a version string so that
@command{gawk}'s @option{--version} option can provide information
about extensions as well.
@end itemize

The requirement to avoid access to @command{gawk}'s symbols is, at first
glance, a difficult one to meet.

One design, apparently used by Perl and Ruby and maybe others, would
be to make the mainline @command{gawk} code into a library, with the
@command{gawk} utility a small C @code{main()} function linked against
the library.

This seemed like the tail wagging the dog, complicating build and
installation and making a simple copy of the @command{gawk} executable
from one system to another (or one place to another on the same
system!) into a chancy operation.

Pat Rankin suggested the solution that was adopted.
@xref{Extension Mechanism Outline}, for the details.

@node Extension Other Design Decisions
@appendixsubsec Other Design Decisions

As an arbitrary design decision, extensions can read the values of
predefined variables and arrays (such as @code{ARGV} and @code{FS}), but cannot
change them, with the exception of @code{PROCINFO}.

The reason for this is to prevent an extension function from affecting
the flow of an @command{awk} program outside its control.  While a real
@command{awk} function can do what it likes, that is at the discretion
of the programmer.  An extension function should provide a service or
make a C API available for use within @command{awk}, and not mess with
@code{FS} or @code{ARGC} and @code{ARGV}.

In addition, it becomes easy to start down a slippery slope. How
much access to @command{gawk} facilities do extensions need?
Do they need @code{getline}?  What about calling @code{gsub()} or
compiling regular expressions?  What about calling into @command{awk}
functions? (@emph{That} would be messy.)

In order to avoid these issues, the @command{gawk} developers chose
to start with the simplest, most basic features that are still truly useful.

Another decision is that although @command{gawk} provides nice things like
MPFR, and arrays indexed internally by integers, these features are not
being brought out to the API in order to keep things simple and close to
traditional @command{awk} semantics.  (In fact, arrays indexed internally
by integers are so transparent that they aren't even documented!)

Additionally, all functions in the API check that their pointer
input parameters are not @code{NULL}. If they are, they return an error.
(It is a good idea for extension code to verify that
pointers received from @command{gawk} are not @code{NULL}.
Such a thing should not happen, but the @command{gawk} developers
are only human, and they have been known to occasionally make
mistakes.)

With time, the API will undoubtedly evolve; the @command{gawk} developers
expect this to be driven by user needs. For now, the current API seems
to provide a minimal yet powerful set of features for creating extensions.

@node Extension Future Growth
@appendixsubsec Room For Future Growth

The API can later be expanded, in two ways:

@itemize @value{BULLET}
@item
@command{gawk} passes an ``extension id'' into the extension when it
first loads the extension.  The extension then passes this id back
to @command{gawk} with each function call.  This mechanism allows
@command{gawk} to identify the extension calling into it, should it need
to know.

@item
Similarly, the extension passes a ``name space'' into @command{gawk}
when it registers each extension function.  This accommodates a possible future
mechanism for grouping extension functions and possibly avoiding name
conflicts.
@end itemize

Of course, as of this writing, no decisions have been made with respect
to any of the above.

@node Old Extension Mechanism
@appendixsec Compatibility For Old Extensions

@ref{Dynamic Extensions}, describes the supported API and mechanisms
for writing extensions for @command{gawk}.  This API was introduced
in @value{PVERSION} 4.1.  However, for many years @command{gawk}
provided an extension mechanism that required knowledge of @command{gawk}
internals and that was not as well designed.

In order to provide a transition period, @command{gawk} @value{PVERSION} 4.1
continues to support the original extension mechanism.
This will be true for the life of exactly one major release.  This support
will be withdrawn, and removed from the source code, at the next major
release.

Briefly, original-style extensions should be compiled by including the
@file{awk.h} header file in the extension source code. Additionally,
you must define the identifier @samp{GAWK} when building (use
@samp{-DGAWK} with Unix-style compilers).  Otherwise, the definitions
in @file{gawkapi.h} will cause conflicts with those in @file{awk.h}
and your extension will not compile.

Just as in previous versions, you load an old-style extension with the
@code{extension()} built-in function (which is not otherwise documented).
This function in turn finds and loads the shared object file containing
the extension and calls its @code{dl_load()} C routine.

Because original-style and new-style extensions use different initialization
routines (@code{dl_load()} versus @code{dlload()}), they may safely
be installed in the same directory (to be found by @env{AWKLIBPATH})
without conflict.

The @command{gawk} development team strongly recommends that you
convert any old extensions that you may have to use the new API
described in @ref{Dynamic Extensions}.

@node Notes summary
@appendixsec Summary

@itemize @value{BULLET}
@item
@command{gawk}'s extensions can be disabled with either the
@option{--traditional} option or with the @option{--posix} option.
The @option{--parsedebug} option is available if @command{gawk} is
compiled with @samp{-DDEBUG}.

@item
The source code for @command{gawk} is maintained in a publicly
accessible Git repository. Anyone may check it out and view the source.

@item
Contributions to @command{gawk} are welcome. Following the steps
outlined in this @value{CHAPTER} will make it easier to integrate
your contributions into the code base.
This applies both to new feature contributions and to ports to
additional operating systems.

@item
@command{gawk} has some limits---generally those that are imposed by
the machine architecture.

@item
The extension API design was intended to solve a number of problems
with the previous extension mechanism, enable features needed by
the @code{xgawk} project, and provide binary compatibility going forward.

@item
The previous extension mechanism is still supported in @value{PVERSION} 4.1
of @command{gawk}, but it @emph{will} be removed in the next major release.

@end itemize


@node Basic Concepts
@appendix Basic Programming Concepts
@cindex programming, concepts
@cindex programming, concepts

This @value{APPENDIX} attempts to define some of the basic concepts
and terms that are used throughout the rest of this @value{DOCUMENT}.
As this @value{DOCUMENT} is specifically about @command{awk},
and not about computer programming in general, the coverage here
is by necessity fairly cursory and simplistic.
(If you need more background, there are many
other introductory texts that you should refer to instead.)

@menu
* Basic High Level::            The high level view.
* Basic Data Typing::           A very quick intro to data types.
@end menu

@node Basic High Level
@appendixsec What a Program Does

@cindex processing data
At the most basic level, the job of a program is to process
some input data and produce results.
@ifnotdocbook
See @ref{figure-general-flow}.
@end ifnotdocbook
@ifdocbook
See @inlineraw{docbook, <xref linkend="figure-general-flow"/>}.
@end ifdocbook

@ifnotdocbook
@float Figure,figure-general-flow
@caption{General Program Flow}
@center @image{general-program, , , General program flow}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-general-flow" float="0">
<title>General Program Flow</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="general-program.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

@cindex compiled programs
@cindex interpreted programs
The ``program'' in the figure can be either a compiled
program@footnote{Compiled programs are typically written
in lower-level languages such as C, C++, or Ada,
and then translated, or @dfn{compiled}, into a form that
the computer can execute directly.}
(such as @command{ls}),
or it may be @dfn{interpreted}.  In the latter case, a machine-executable
program such as @command{awk} reads your program, and then uses the
instructions in your program to process the data.

@cindex programming, basic steps
When you write a program, it usually consists
of the following, very basic set of steps,
@ifnotdocbook
as shown in @ref{figure-process-flow}:
@end ifnotdocbook
@ifdocbook
as shown in @inlineraw{docbook, <xref linkend="figure-process-flow"/>}:
@end ifdocbook

@ifnotdocbook
@float Figure,figure-process-flow
@caption{Basic Program Steps}
@center @image{process-flow, , , Basic Program Stages}
@end float
@end ifnotdocbook

@docbook
<figure id="figure-process-flow" float="0">
<title>Basic Program Stages</title>
<mediaobject>
<imageobject role="web"><imagedata fileref="process-flow.png" format="PNG"/></imageobject>
</mediaobject>
</figure>
@end docbook

@table @asis
@item Initialization
These are the things you do before actually starting to process
data, such as checking arguments, initializing any data you need
to work with, and so on.
This step corresponds to @command{awk}'s @code{BEGIN} rule
(@pxref{BEGIN/END}).

If you were baking a cake, this might consist of laying out all the
mixing bowls and the baking pan, and making sure you have all the
ingredients that you need.

@item Processing
This is where the actual work is done.  Your program reads data,
one logical chunk at a time, and processes it as appropriate.

In most programming languages, you have to manually manage the reading
of data, checking to see if there is more each time you read a chunk.
@command{awk}'s pattern-action paradigm
(@pxref{Getting Started})
handles the mechanics of this for you.

In baking a cake, the processing corresponds to the actual labor:
breaking eggs, mixing the flour, water, and other ingredients, and then putting the cake
into the oven.

@item Clean Up
Once you've processed all the data, you may have things you need to
do before exiting.
This step corresponds to @command{awk}'s @code{END} rule
(@pxref{BEGIN/END}).

After the cake comes out of the oven, you still have to wrap it in
plastic wrap to keep anyone from tasting it, as well as wash
the mixing bowls and utensils.
@end table

@cindex algorithms
An @dfn{algorithm} is a detailed set of instructions necessary to accomplish
a task, or process data.  It is much the same as a recipe for baking
a cake.  Programs implement algorithms.  Often, it is up to you to design
the algorithm and implement it, simultaneously.

@cindex records
@cindex fields
The ``logical chunks'' we talked about previously are called @dfn{records},
similar to the records a company keeps on employees, a school keeps for
students, or a doctor keeps for patients.
Each record has many component parts, such as first and last names,
date of birth, address, and so on.  The component parts are referred
to as the @dfn{fields} of the record.

The act of reading data is termed @dfn{input}, and that of
generating results, not too surprisingly, is termed @dfn{output}.
They are often referred to together as ``input/output,''
and even more often, as ``I/O'' for short.
(You will also see ``input'' and ``output'' used as verbs.)

@cindex data-driven languages
@cindex languages@comma{} data-driven
@command{awk} manages the reading of data for you, as well as the
breaking it up into records and fields.  Your program's job is to
tell @command{awk} what to do with the data.  You do this by describing
@dfn{patterns} in the data to look for, and @dfn{actions} to execute
when those patterns are seen.  This @dfn{data-driven} nature of
@command{awk} programs usually makes them both easier to write
and easier to read.

@node Basic Data Typing
@appendixsec Data Values in a Computer

@cindex variables
In a program,
you keep track of information and values in things called @dfn{variables}.
A variable is just a name for a given value, such as @code{first_name},
@code{last_name}, @code{address}, and so on.
@command{awk} has several predefined variables, and it has
special names to refer to the current input record
and the fields of the record.
You may also group multiple
associated values under one name, as an array.

@cindex values, numeric
@cindex values, string
@cindex scalar values
Data, particularly in @command{awk}, consists of either numeric
values, such as 42 or 3.1415927, or string values.
String values are essentially anything that's not a number, such as a name.
Strings are sometimes referred to as @dfn{character data}, since they
store the individual characters that comprise them.
Individual variables, as well as numeric and string variables, are
referred to as @dfn{scalar} values.
Groups of values, such as arrays, are not scalars.

@ref{Computer Arithmetic}, provided a basic introduction to numeric
types (integer and floating-point) and how they are used in a computer.
Please review that information, including a number of caveats that
were presented.

@cindex null strings
While you are probably used to the idea of a number without a value (i.e., zero),
it takes a bit more getting used to the idea of zero-length character data.
Nevertheless, such a thing exists.
It is called the @dfn{null string}.
The null string is character data that has no value.
In other words, it is empty.  It is written in @command{awk} programs
like this: @code{""}.

Humans are used to working in decimal; i.e., base 10.  In base 10,
numbers go from 0 to 9, and then ``roll over'' into the next
@iftex
column.  (Remember grade school? @math{42 = 4\times 10 + 2}.)
@end iftex
@ifnottex
column.  (Remember grade school? 42 = 4 x 10 + 2.)
@end ifnottex

There are other number bases though.  Computers commonly use base 2
or @dfn{binary}, base 8 or @dfn{octal}, and base 16 or @dfn{hexadecimal}.
In binary, each column represents two times the value in the column to
its right. Each column may contain either a 0 or a 1.
@iftex
Thus, binary 1010 represents @math{(1\times 8) + (0\times 4) + (1\times 2) + (0\times 1)}, or decimal 10.
@end iftex
@ifnottex
Thus, binary 1010 represents (1 x 8) + (0 x 4) + (1 x 2)
+ (0 x 1), or decimal 10.
@end ifnottex
Octal and hexadecimal are discussed more in
@ref{Nondecimal-numbers}.

At the very lowest level, computers store values as groups of binary digits,
or @dfn{bits}.  Modern computers group bits into groups of eight, called @dfn{bytes}.
Advanced applications sometimes have to manipulate bits directly,
and @command{gawk} provides functions for doing so.

Programs are written in programming languages.
Hundreds, if not thousands, of programming languages exist.
One of the most popular is the C programming language.
The C language had a very strong influence on the design of
the @command{awk} language.

@cindex Kernighan, Brian
@cindex Ritchie, Dennis
There have been several versions of C.  The first is often referred to
as ``K&R'' C, after the initials of Brian Kernighan and Dennis Ritchie,
the authors of the first book on C.  (Dennis Ritchie created the language,
and Brian Kernighan was one of the creators of @command{awk}.)

In the mid-1980s, an effort began to produce an international standard
for C.  This work culminated in 1989, with the production of the ANSI
standard for C.  This standard became an ISO standard in 1990.
In 1999, a revised ISO C standard was approved and released.
Where it makes sense, POSIX @command{awk} is compatible with 1999 ISO C.


@node Glossary
@unnumbered Glossary

@table @asis
@item Action
A series of @command{awk} statements attached to a rule.  If the rule's
pattern matches an input record, @command{awk} executes the
rule's action.  Actions are always enclosed in braces.
(@xref{Action Overview}.)

@cindex Ada programming language
@cindex programming languages, Ada
@item Ada
A programming language originally defined by the U.S.@: Department of
Defense for embedded programming. It was designed to enforce good
Software Engineering practices.

@cindex Spencer, Henry
@cindex @command{sed} utility
@cindex amazing @command{awk} assembler (@command{aaa})
@item Amazing @command{awk} Assembler
Henry Spencer at the University of Toronto wrote a retargetable assembler
completely as @command{sed} and @command{awk} scripts.  It is thousands
of lines long, including machine descriptions for several eight-bit
microcomputers.  It is a good example of a program that would have been
better written in another language.
@c You can get it from @uref{http://awk.info/?awk100/aaa}.

@cindex amazingly workable formatter (@command{awf})
@cindex @command{awf} (amazingly workable formatter) program
@item Amazingly Workable Formatter (@command{awf})
Henry Spencer at the University of Toronto wrote a formatter that accepts
a large subset of the @samp{nroff -ms} and @samp{nroff -man} formatting
commands, using @command{awk} and @command{sh}.
@c It is available
@c from @uref{http://awk.info/?tools/awf}.

@item Anchor
The regexp metacharacters @samp{^} and @samp{$}, which force the match
to the beginning or end of the string, respectively.

@cindex ANSI
@item ANSI
The American National Standards Institute.  This organization produces
many standards, among them the standards for the C and C++ programming
languages.
These standards often become international standards as well. See also
``ISO.''

@item Argument
An argument can be two different things.  It can be an option or a
@value{FN} passed to a command while invoking it from the command line, or
it can be something passed to a @dfn{function} inside a program, e.g.
inside @command{awk}.

In the latter case, an argument can be passed to a function in two ways.
Either it is given to the called function by value, i.e., a copy of the
value of the variable is made available to the called function, but the
original variable cannot be modified by the function itself; or it is
given by reference, i.e., a pointer to the interested variable is passed to
the function, which can then directly modify it. In @command{awk}
scalars are passed by value, and arrays are passed by reference.
See ``Pass By Value/Reference.''

@item Array
A grouping of multiple values under the same name.
Most languages just provide sequential arrays.
@command{awk} provides associative arrays.

@item Assertion
A statement in a program that a condition is true at this point in the program.
Useful for reasoning about how a program is supposed to behave.

@item Assignment
An @command{awk} expression that changes the value of some @command{awk}
variable or data object.  An object that you can assign to is called an
@dfn{lvalue}.  The assigned values are called @dfn{rvalues}.
@xref{Assignment Ops}.

@item Associative Array
Arrays in which the indices may be numbers or strings, not just
sequential integers in a fixed range.

@item @command{awk} Language
The language in which @command{awk} programs are written.

@item @command{awk} Program
An @command{awk} program consists of a series of @dfn{patterns} and
@dfn{actions}, collectively known as @dfn{rules}.  For each input record
given to the program, the program's rules are all processed in turn.
@command{awk} programs may also contain function definitions.

@item @command{awk} Script
Another name for an @command{awk} program.

@item Bash
The GNU version of the standard shell
@ifnotinfo
(the @b{B}ourne-@b{A}gain @b{SH}ell).
@end ifnotinfo
@ifinfo
(the Bourne-Again SHell).
@end ifinfo
See also ``Bourne Shell.''

@item Binary
Base-two notation, where the digits are @code{0}--@code{1}. Since
electronic circuitry works ``naturally'' in base 2 (just think of Off/On),
everything inside a computer is calculated using base 2. Each digit
represents the presence (or absence) of a power of 2 and is called a
@dfn{bit}. So, for example, the base-two number @code{10101} is
@iftex
the same as decimal 21, (@math{(1\times 16) + (1\times 4) + (1\times 1)}).
@end iftex
@ifnottex
the same as decimal 21, ((1 x 16) + (1 x 4) + (1 x 1)).
@end ifnottex

Since base-two numbers quickly become
very long to read and write, they are usually grouped by 3 (i.e., they are
read as octal numbers), or by 4 (i.e., they are read as hexadecimal
numbers). There is no direct way to insert base 2 numbers in a C program.
If need arises, such numbers are usually inserted as octal or hexadecimal
numbers. The number of base-two digits that fit into registers used for
representing integer numbers in computers is a rough indication of the
computing power of the computer itself.  Most computers nowadays use 64
bits for representing integer numbers in their registers, but 32-bit,
16-bit and 8-bit registers have been widely used in the past.
@xref{Nondecimal-numbers}.
@item Bit
Short for ``Binary Digit.''
All values in computer memory ultimately reduce to binary digits: values
that are either zero or one.
Groups of bits may be interpreted differently---as integers,
floating-point numbers, character data, addresses of other
memory objects, or other data.
@command{awk} lets you work with floating-point numbers and strings.
@command{gawk} lets you manipulate bit values with the built-in
functions described in
@ref{Bitwise Functions}.

Computers are often defined by how many bits they use to represent integer
values.  Typical systems are 32-bit systems, but 64-bit systems are
becoming increasingly popular, and 16-bit systems have essentially
disappeared.

@item Boolean Expression
Named after the English mathematician Boole. See also ``Logical Expression.''

@item Bourne Shell
The standard shell (@file{/bin/sh}) on Unix and Unix-like systems,
originally written by Steven R.@: Bourne at Bell Laboratories.
Many shells (Bash, @command{ksh}, @command{pdksh}, @command{zsh}) are
generally upwardly compatible with the Bourne shell.

@item Braces
The characters @samp{@{} and @samp{@}}.  Braces are used in
@command{awk} for delimiting actions, compound statements, and function
bodies.

@item Bracket Expression
Inside a @dfn{regular expression}, an expression included in square
brackets, meant to designate a single character as belonging to a
specified character class. A bracket expression can contain a list of one
or more characters, like @samp{[abc]}, a range of characters, like
@samp{[A-Z]}, or a name, delimited by @samp{:}, that designates a known set
of characters, like @samp{[:digit:]}. The form of bracket expression
enclosed between @samp{:} is independent of the underlying representation
of the character themselves, which could utilize the ASCII, EBCDIC, or
Unicode codesets, depending on the architecture of the computer system, and on
localization.
See also ``Regular Expression.''

@item Built-in Function
The @command{awk} language provides built-in functions that perform various
numerical, I/O-related, and string computations.  Examples are
@code{sqrt()} (for the square root of a number) and @code{substr()} (for a
substring of a string).
@command{gawk} provides functions for timestamp management, bit manipulation,
array sorting, type checking,
and runtime string translation.
(@xref{Built-in}.)

@item Built-in Variable
@code{ARGC},
@code{ARGV},
@code{CONVFMT},
@code{ENVIRON},
@code{FILENAME},
@code{FNR},
@code{FS},
@code{NF},
@code{NR},
@code{OFMT},
@code{OFS},
@code{ORS},
@code{RLENGTH},
@code{RSTART},
@code{RS},
and
@code{SUBSEP}
are the variables that have special meaning to @command{awk}.
In addition,
@code{ARGIND},
@code{BINMODE},
@code{ERRNO},
@code{FIELDWIDTHS},
@code{FPAT},
@code{IGNORECASE},
@code{LINT},
@code{PROCINFO},
@code{RT},
and
@code{TEXTDOMAIN}
are the variables that have special meaning to @command{gawk}.
Changing some of them affects @command{awk}'s running environment.
(@xref{Built-in Variables}.)

@item C
The system programming language that most GNU software is written in.  The
@command{awk} programming language has C-like syntax, and this @value{DOCUMENT}
points out similarities between @command{awk} and C when appropriate.

In general, @command{gawk} attempts to be as similar to the 1990 version
of ISO C as makes sense.

@item C Shell
The C Shell (@command{csh} or its improved version, @command{tcsh}) is a Unix shell that was
created by Bill Joy in the late 1970s. The C shell was differentiated from
other shells by its interactive features and overall style, which
looks more like C. The C Shell is not backward compatible with the Bourne
Shell, so special attention is required when converting scripts
written for other Unix shells to the C shell, especially with regard to the management of
shell variables.
See also ``Bourne Shell.''

@item C++
A popular object-oriented programming language derived from C.

@item Character Class
See ``Bracket Expression.''

@item Character List
See ``Bracket Expression.''

@cindex ASCII
@cindex ISO 8859-1
@cindex ISO Latin-1
@cindex character sets (machine character encodings)
@cindex Unicode
@item Character Set
The set of numeric codes used by a computer system to represent the
characters (letters, numbers, punctuation, etc.) of a particular country
or place. The most common character set in use today is ASCII (American
Standard Code for Information Interchange).  Many European
countries use an extension of ASCII known as ISO-8859-1 (ISO Latin-1).
The @uref{http://www.unicode.org, Unicode character set} is
increasingly popular and standard, and is particularly
widely used on GNU/Linux systems.

@cindex Kernighan, Brian
@cindex Bentley, Jon
@cindex @command{chem} utility
@item CHEM
A preprocessor for @command{pic} that reads descriptions of molecules
and produces @command{pic} input for drawing them.
It was written in @command{awk}
by Brian Kernighan and Jon Bentley, and is available from
@uref{http://netlib.org/typesetting/chem}.

@item Comparison Expression
A relation that is either true or false, such as @samp{a < b}.
Comparison expressions are used in @code{if}, @code{while}, @code{do},
and @code{for}
statements, and in patterns to select which input records to process.
(@xref{Typing and Comparison}.)

@cindex compiled programs
@item Compiler
A program that translates human-readable source code into
machine-executable object code.  The object code is then executed
directly by the computer.
See also ``Interpreter.''

@item Complemented Bracket Expression
The negation of a @dfn{bracket expression}.  All that is @emph{not}
described by a given bracket expression. The symbol @samp{^} precedes
the negated bracket expression.  E.g.: @samp{[^[:digit:]]}
designates whatever character is not a digit. @samp{[^bad]}
designates whatever character is not one of the letters @samp{b}, @samp{a},
or @samp{d}.
See ``Bracket Expression.''

@item Compound Statement
A series of @command{awk} statements, enclosed in curly braces.  Compound
statements may be nested.
(@xref{Statements}.)

@item Computed Regexps
See ``Dynamic Regular Expressions.''

@item Concatenation
Concatenating two strings means sticking them together, one after another,
producing a new string.  For example, the string @samp{foo} concatenated with
the string @samp{bar} gives the string @samp{foobar}.
(@xref{Concatenation}.)

@item Conditional Expression
An expression using the @samp{?:} ternary operator, such as
@samp{@var{expr1} ? @var{expr2} : @var{expr3}}.  The expression
@var{expr1} is evaluated; if the result is true, the value of the whole
expression is the value of @var{expr2}; otherwise the value is
@var{expr3}.  In either case, only one of @var{expr2} and @var{expr3}
is evaluated. (@xref{Conditional Exp}.)

@item Control Statement
A control statement is an instruction to perform a given operation or a set
of operations inside an @command{awk} program, if a given condition is
true. Control statements are: @code{if}, @code{for}, @code{while}, and
@code{do}
(@pxref{Statements}).

@cindex McIlroy, Doug
@cindex cookie
@item Cookie
A peculiar goodie, token, saying or remembrance
produced by or presented to a program. (With thanks to Professor Doug McIlroy.)
@ignore
From: Doug McIlroy <doug@cs.dartmouth.edu>
Date: Sat, 13 Oct 2012 19:55:25 -0400
To: arnold@skeeve.com
Subject: Re: origin of the term "cookie"?

I believe the term "cookie", for a more or less inscrutable
saying or crumb of information, was injected into Unix
jargon by Bob Morris, who used the word quite frequently.
It had no fixed meaning as it now does in browsers.

The word had been around long before it was recognized in
the 8th edition glossary (earlier editions had no glossary):

cookie   a peculiar goodie, token, saying or remembrance
returned by or presented to a program. [I would say that
"returned by" would better read "produced by", and assume
responsibility for the inexactitude.]

Doug McIlroy

From: Doug McIlroy <doug@cs.dartmouth.edu>
Date: Sun, 14 Oct 2012 10:08:43 -0400
To: arnold@skeeve.com
Subject: Re: origin of the term "cookie"?

> Can I forward your email to Eric Raymond, for possible addition to the
> Jargon File?

Sure. I might add that I don't know how "cookie" entered Morris's
vocabulary. Certainly "values of beta give rise to dom!" (see google)
was an early, if not the earliest Unix cookie.  The fact that it was
found lying around on a model 37 teletype (which had Greek beta in
its type box) suggests that maybe it was seen to be like milk and
cookies laid out for Santa Claus. Morris was wont to make such
connections.

Doug
@end ignore

@item Coprocess
A subordinate program with which two-way communications is possible.

@item Curly Braces
See ``Braces.''

@cindex dark corner
@item Dark Corner
An area in the language where specifications often were (or still
are) not clear, leading to unexpected or undesirable behavior.
Such areas are marked in this @value{DOCUMENT} with
@iftex
the picture of a flashlight in the margin
@end iftex
@ifnottex
``(d.c.)'' in the text
@end ifnottex
and are indexed under the heading ``dark corner.''

@item Data Driven
A description of @command{awk} programs, where you specify the data you
are interested in processing, and what to do when that data is seen.

@item Data Objects
These are numbers and strings of characters.  Numbers are converted into
strings and vice versa, as needed.
(@xref{Conversion}.)

@item Deadlock
The situation in which two communicating processes are each waiting
for the other to perform an action.

@item Debugger
A program used to help developers remove ``bugs'' from (de-bug)
their programs.

@item Double Precision
An internal representation of numbers that can have fractional parts.
Double precision numbers keep track of more digits than do single precision
numbers, but operations on them are sometimes more expensive.  This is the way
@command{awk} stores numeric values.  It is the C type @code{double}.

@item Dynamic Regular Expression
A dynamic regular expression is a regular expression written as an
ordinary expression.  It could be a string constant, such as
@code{"foo"}, but it may also be an expression whose value can vary.
(@xref{Computed Regexps}.)

@item Empty String
See ``Null String.''

@item Environment
A collection of strings, of the form @samp{@var{name}=@var{val}}, that each
program has available to it. Users generally place values into the
environment in order to provide information to various programs. Typical
examples are the environment variables @env{HOME} and @env{PATH}.

@cindex epoch, definition of
@item Epoch
The date used as the ``beginning of time'' for timestamps.
Time values in most systems are represented as seconds since the epoch,
with library functions available for converting these values into
standard date and time formats.

The epoch on Unix and POSIX systems is 1970-01-01 00:00:00 UTC.
See also ``GMT'' and ``UTC.''

@item Escape Sequences
A special sequence of characters used for describing nonprinting
characters, such as @samp{\n} for newline or @samp{\033} for the ASCII
ESC (Escape) character. (@xref{Escape Sequences}.)

@item Extension
An additional feature or change to a programming language or
utility not defined by that language's or utility's standard.
@command{gawk} has (too) many extensions over POSIX @command{awk}.

@item FDL
See ``Free Documentation License.''

@item Field
When @command{awk} reads an input record, it splits the record into pieces
separated by whitespace (or by a separator regexp that you can
change by setting the predefined variable @code{FS}).  Such pieces are
called fields.  If the pieces are of fixed length, you can use the built-in
variable @code{FIELDWIDTHS} to describe their lengths.
If you wish to specify the contents of fields instead of the field
separator, you can use the predefined variable @code{FPAT} to do so.
(@xref{Field Separators},
@ref{Constant Size},
and
@ref{Splitting By Content}.)

@item Flag
A variable whose truth value indicates the existence or nonexistence
of some condition.

@item Floating-Point Number
Often referred to in mathematical terms as a ``rational'' or real number,
this is just a number that can have a fractional part.
See also ``Double Precision'' and ``Single Precision.''

@item Format
Format strings control the appearance of output in the
@code{strftime()} and @code{sprintf()} functions, and in the
@code{printf} statement as well.  Also, data conversions from numbers to strings
are controlled by the format strings contained in the predefined variables
@code{CONVFMT} and @code{OFMT}. (@xref{Control Letters}.)

@item Fortran
Shorthand for FORmula TRANslator, one of the first programming languages
available for scientific calculations. It was created by John Backus,
and has been available since 1957. It is still in use today.

@item Free Documentation License
This document describes the terms under which this @value{DOCUMENT}
is published and may be copied. (@xref{GNU Free Documentation License}.)

@cindex FSF (Free Software Foundation)
@cindex Free Software Foundation (FSF)
@cindex Stallman, Richard
@item Free Software Foundation
A nonprofit organization dedicated
to the production and distribution of freely distributable software.
It was founded by Richard M.@: Stallman, the author of the original
Emacs editor.  GNU Emacs is the most widely used version of Emacs today.

@item FSF
See ``Free Software Foundation.''

@item Function
A part of an @command{awk} program that can be invoked from every point of
the program, to perform a task.  @command{awk} has several built-in
functions.
Users can define their own functions in every part of the program.
Function can be recursive, i.e., they may invoke themselves.
@xref{Functions}.
In @command{gawk} it is also possible to have functions shared
among different programs, and included where required using the
@code{@@include} directive
(@pxref{Include Files}).
In @command{gawk} the name of the function that should be invoked
can be generated at run time, i.e., dynamically.
The @command{gawk} extension API provides constructor functions
(@pxref{Constructor Functions}).


@item @command{gawk}
The GNU implementation of @command{awk}.

@cindex GPL (General Public License)
@cindex General Public License (GPL)
@cindex GNU General Public License
@item General Public License
This document describes the terms under which @command{gawk} and its source
code may be distributed. (@xref{Copying}.)

@item GMT
``Greenwich Mean Time.''
This is the old term for UTC.
It is the time of day used internally for Unix and POSIX systems.
See also ``Epoch'' and ``UTC.''

@cindex FSF (Free Software Foundation)
@cindex Free Software Foundation (FSF)
@cindex GNU Project
@item GNU
``GNU's not Unix''.  An on-going project of the Free Software Foundation
to create a complete, freely distributable, POSIX-compliant computing
environment.

@item GNU/Linux
A variant of the GNU system using the Linux kernel, instead of the
Free Software Foundation's Hurd kernel.
The Linux kernel is a stable, efficient, full-featured clone of Unix that has
been ported to a variety of architectures.
It is most popular on PC-class systems, but runs well on a variety of
other systems too.
The Linux kernel source code is available under the terms of the GNU General
Public License, which is perhaps its most important aspect.

@item GPL
See ``General Public License.''

@item Hexadecimal
Base 16 notation, where the digits are @code{0}--@code{9} and
@code{A}--@code{F}, with @samp{A}
representing 10, @samp{B} representing 11, and so on, up to @samp{F} for 15.
Hexadecimal numbers are written in C using a leading @samp{0x},
@iftex
to indicate their base.  Thus, @code{0x12} is 18 (@math{(1\times 16) + 2}).
@end iftex
@ifnottex
to indicate their base.  Thus, @code{0x12} is 18 ((1 x 16) + 2).
@end ifnottex
@xref{Nondecimal-numbers}.

@item I/O
Abbreviation for ``Input/Output,'' the act of moving data into and/or
out of a running program.

@item Input Record
A single chunk of data that is read in by @command{awk}.  Usually, an @command{awk} input
record consists of one line of text.
(@xref{Records}.)

@item Integer
A whole number, i.e., a number that does not have a fractional part.

@item Internationalization
The process of writing or modifying a program so
that it can use multiple languages without requiring
further source code changes.

@cindex interpreted programs
@item Interpreter
A program that reads human-readable source code directly, and uses
the instructions in it to process data and produce results.
@command{awk} is typically (but not always) implemented as an interpreter.
See also ``Compiler.''

@item Interval Expression
A component of a regular expression that lets you specify repeated matches of
some part of the regexp.  Interval expressions were not originally available
in @command{awk} programs.

@cindex ISO
@item ISO
The International Organization for Standardization.
This organization produces international standards for many things, including
programming languages, such as C and C++.
In the computer arena, important standards like those for C, C++, and POSIX
become both American national and ISO international standards simultaneously.
This @value{DOCUMENT} refers to Standard C as ``ISO C'' throughout.
See @uref{https://www.iso.org/iso/home/about.htm, the ISO website} for more
information about the name of the organization and its language-independent
three-letter acronym.

@cindex Java programming language
@cindex programming languages, Java
@item Java
A modern programming language originally developed by Sun Microsystems
(now Oracle) supporting Object-Oriented programming.  Although usually
implemented by compiling to the instructions for a standard virtual
machine (the JVM), the language can be compiled to native code.

@item Keyword
In the @command{awk} language, a keyword is a word that has special
meaning.  Keywords are reserved and may not be used as variable names.

@command{gawk}'s keywords are:
@code{BEGIN},
@code{BEGINFILE},
@code{END},
@code{ENDFILE},
@code{break},
@code{case},
@code{continue},
@code{default},
@code{delete},
@code{do@dots{}while},
@code{else},
@code{exit},
@code{for@dots{}in},
@code{for},
@code{function},
@code{func},
@code{if},
@code{next},
@code{nextfile},
@code{switch},
and
@code{while}.

@item Korn Shell
The Korn Shell (@command{ksh}) is a Unix shell which was developed by David Korn at Bell
Laboratories in the early 1980s. The Korn Shell is backward-compatible with the Bourne
shell and includes many features of the C shell.
See also ``Bourne Shell.''

@cindex LGPL (Lesser General Public License)
@cindex Lesser General Public License (LGPL)
@cindex GNU Lesser General Public License
@item Lesser General Public License
This document describes the terms under which binary library archives
or shared objects,
and their source code may be distributed.

@item LGPL
See ``Lesser General Public License.''

@item Linux
See ``GNU/Linux.''

@item Localization
The process of providing the data necessary for an
internationalized program to work in a particular language.

@item Logical Expression
An expression using the operators for logic, AND, OR, and NOT, written
@samp{&&}, @samp{||}, and @samp{!} in @command{awk}. Often called Boolean
expressions, after the mathematician who pioneered this kind of
mathematical logic.

@item Lvalue
An expression that can appear on the left side of an assignment
operator.  In most languages, lvalues can be variables or array
elements.  In @command{awk}, a field designator can also be used as an
lvalue.

@item Matching
The act of testing a string against a regular expression.  If the
regexp describes the contents of the string, it is said to @dfn{match} it.

@item Metacharacters
Characters used within a regexp that do not stand for themselves.
Instead, they denote regular expression operations, such as repetition,
grouping, or alternation.

@item Nesting
Nesting is where information is organized in layers, or where objects
contain other similar objects.
In @command{gawk} the @code{@@include}
directive can be nested. The ``natural'' nesting of arithmetic and
logical operations can be changed using parentheses
(@pxref{Precedence}).

@item No-op
An operation that does nothing.

@item Null String
A string with no characters in it.  It is represented explicitly in
@command{awk} programs by placing two double quote characters next to
each other (@code{""}).  It can appear in input data by having two successive
occurrences of the field separator appear next to each other.

@item Number
A numeric-valued data object.  Modern @command{awk} implementations use
double precision floating-point to represent numbers.
Ancient @command{awk} implementations used single precision floating-point.

@item Octal
Base-eight notation, where the digits are @code{0}--@code{7}.
Octal numbers are written in C using a leading @samp{0},
@iftex
to indicate their base.  Thus, @code{013} is 11 (@math{(1\times 8) + 3}).
@end iftex
@ifnottex
to indicate their base.  Thus, @code{013} is 11 ((1 x 8) + 3).
@end ifnottex
@xref{Nondecimal-numbers}.

@item Output Record
A single chunk of data that is written out by @command{awk}.  Usually, an
@command{awk} output record consists of one or more lines of text.
@xref{Records}.

@item Pattern
Patterns tell @command{awk} which input records are interesting to which
rules.

A pattern is an arbitrary conditional expression against which input is
tested.  If the condition is satisfied, the pattern is said to @dfn{match}
the input record.  A typical pattern might compare the input record against
a regular expression. (@xref{Pattern Overview}.)

@item PEBKAC
An acronym describing what is possibly the most frequent
source of computer usage problems. (Problem Exists Between
Keyboard And Chair.)

@item Plug-in
See ``Extensions.''

@item POSIX
The name for a series of standards
that specify a Portable Operating System interface.  The ``IX'' denotes
the Unix heritage of these standards.  The main standard of interest for
@command{awk} users is
@cite{IEEE Standard for Information Technology, Standard 1003.1-2008}.
The 2008 POSIX standard can be found online at
@url{http://pubs.opengroup.org/onlinepubs/9699919799/}.

@item Precedence
The order in which operations are performed when operators are used
without explicit parentheses.

@item Private
Variables and/or functions that are meant for use exclusively by library
functions and not for the main @command{awk} program. Special care must be
taken when naming such variables and functions.
(@xref{Library Names}.)

@item Range (of input lines)
A sequence of consecutive lines from the input file(s).  A pattern
can specify ranges of input lines for @command{awk} to process or it can
specify single lines. (@xref{Pattern Overview}.)

@item Record
See ``Input record'' and ``Output record.''

@item Recursion
When a function calls itself, either directly or indirectly.
If this is clear, stop, and proceed to the next entry.
Otherwise, refer to the entry for ``recursion.''

@item Redirection
Redirection means performing input from something other than the standard input
stream, or performing output to something other than the standard output stream.

You can redirect input to the @code{getline} statement using
the @samp{<}, @samp{|}, and @samp{|&} operators.
You can redirect the output of the @code{print} and @code{printf} statements
to a file or a system command, using the @samp{>}, @samp{>>}, @samp{|}, and @samp{|&}
operators.
(@xref{Getline},
and @ref{Redirection}.)

@item Reference Counts
An internal mechanism in @command{gawk} to minimize the amount of memory
needed to store the value of string variables. If the value assumed by
a variable is used in more than one place, only one copy of the value
itself is kept, and the associated reference count is increased when the
same value is used by an additional variable, and decreased when the related
variable is no longer in use. When the reference count goes to zero,
the memory space used to store the value of the variable is freed.

@item Regexp
See ``Regular Expression.''

@item Regular Expression
A regular expression (``regexp'' for short) is a pattern that denotes a
set of strings, possibly an infinite set.  For example, the regular expression
@samp{R.*xp} matches any string starting with the letter @samp{R}
and ending with the letters @samp{xp}.  In @command{awk}, regular expressions are
used in patterns and in conditional expressions.  Regular expressions may contain
escape sequences. (@xref{Regexp}.)

@item Regular Expression Constant
A regular expression constant is a regular expression written within
slashes, such as @code{/foo/}.  This regular expression is chosen
when you write the @command{awk} program and cannot be changed during
its execution. (@xref{Regexp Usage}.)

@item Regular Expression Operators
See ``Metacharacters.''

@item Rounding
Rounding the result of an arithmetic operation can be tricky.
More than one way of rounding exists, and in @command{gawk}
it is possible to choose which method should be used in a program.
@xref{Setting the rounding mode}.

@item Rule
A segment of an @command{awk} program that specifies how to process single
input records.  A rule consists of a @dfn{pattern} and an @dfn{action}.
@command{awk} reads an input record; then, for each rule, if the input record
satisfies the rule's pattern, @command{awk} executes the rule's action.
Otherwise, the rule does nothing for that input record.

@item Rvalue
A value that can appear on the right side of an assignment operator.
In @command{awk}, essentially every expression has a value. These values
are rvalues.

@item Scalar
A single value, be it a number or a string.
Regular variables are scalars; arrays and functions are not.

@item Search Path
In @command{gawk}, a list of directories to search for @command{awk} program source files.
In the shell, a list of directories to search for executable programs.

@item @command{sed}
See ``Stream Editor.''

@item Seed
The initial value, or starting point, for a sequence of random numbers.

@item Shell
The command interpreter for Unix and POSIX-compliant systems.
The shell works both interactively, and as a programming language
for batch files, or shell scripts.

@item Short-Circuit
The nature of the @command{awk} logical operators @samp{&&} and @samp{||}.
If the value of the entire expression is determinable from evaluating just
the lefthand side of these operators, the righthand side is not
evaluated.
(@xref{Boolean Ops}.)

@item Side Effect
A side effect occurs when an expression has an effect aside from merely
producing a value.  Assignment expressions, increment and decrement
expressions, and function calls have side effects.
(@xref{Assignment Ops}.)

@item Single Precision
An internal representation of numbers that can have fractional parts.
Single precision numbers keep track of fewer digits than do double precision
numbers, but operations on them are sometimes less expensive in terms of CPU time.
This is the type used by some ancient versions of @command{awk} to store
numeric values.  It is the C type @code{float}.

@item Space
The character generated by hitting the space bar on the keyboard.

@item Special File
A @value{FN} interpreted internally by @command{gawk}, instead of being handed
directly to the underlying operating system---for example, @file{/dev/stderr}.
(@xref{Special Files}.)

@item Statement
An expression inside an @command{awk} program in the action part
of a pattern--action rule, or inside an
@command{awk} function. A statement can be a variable assignment,
an array operation, a loop, etc.

@item Stream Editor
A program that reads records from an input stream and processes them one
or more at a time.  This is in contrast with batch programs, which may
expect to read their input files in entirety before starting to do
anything, as well as with interactive programs which require input from the
user.

@item String
A datum consisting of a sequence of characters, such as @samp{I am a
string}.  Constant strings are written with double quotes in the
@command{awk} language and may contain escape sequences.
(@xref{Escape Sequences}.)

@item Tab
The character generated by hitting the @kbd{TAB} key on the keyboard.
It usually expands to up to eight spaces upon output.

@item Text Domain
A unique name that identifies an application.
Used for grouping messages that are translated at runtime
into the local language.

@item Timestamp
A value in the ``seconds since the epoch'' format used by Unix
and POSIX systems.  Used for the @command{gawk} functions
@code{mktime()}, @code{strftime()}, and @code{systime()}.
See also ``Epoch,'' ``GMT,'' and ``UTC.''

@cindex Linux
@cindex GNU/Linux
@cindex Unix
@cindex BSD-based operating systems
@cindex NetBSD
@cindex FreeBSD
@cindex OpenBSD
@item Unix
A computer operating system originally developed in the early 1970's at
AT&T Bell Laboratories.  It initially became popular in universities around
the world and later moved into commercial environments as a software
development system and network server system. There are many commercial
versions of Unix, as well as several work-alike systems whose source code
is freely available (such as GNU/Linux, @uref{http://www.netbsd.org, NetBSD},
@uref{https://www.freebsd.org, FreeBSD}, and @uref{http://www.openbsd.org, OpenBSD}).

@item UTC
The accepted abbreviation for ``Universal Coordinated Time.''
This is standard time in Greenwich, England, which is used as a
reference time for day and date calculations.
See also ``Epoch'' and ``GMT.''

@item Variable
A name for a value. In @command{awk}, variables may be either scalars
or arrays.

@item Whitespace
A sequence of space, TAB, or newline characters occurring inside an input
record or a string.

@end table

@end ifclear

@c The GNU General Public License.
@node Copying
@unnumbered GNU General Public License
@ifnotdocbook
@center Version 3, 29 June 2007
@end ifnotdocbook
@docbook
<subtitle>Version 3, 29 June 2007</subtitle>
@end docbook

@c This file is intended to be included within another document,
@c hence no sectioning command or @node.

@display
Copyright @copyright{} 2007 Free Software Foundation, Inc. @url{https://fsf.org/}

Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.
@end display

@c fakenode --- for prepinfo
@heading Preamble

The GNU General Public License is a free, copyleft license for
software and other kinds of works.

The licenses for most software and other practical works are designed
to take away your freedom to share and change the works.  By contrast,
the GNU General Public License is intended to guarantee your freedom
to share and change all versions of a program---to make sure it remains
free software for all its users.  We, the Free Software Foundation,
use the GNU General Public License for most of our software; it
applies also to any other work released this way by its authors.  You
can apply it to your programs, too.

When we speak of free software, we are referring to freedom, not
price.  Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
them if you wish), that you receive source code or can get it if you
want it, that you can change the software or use pieces of it in new
free programs, and that you know you can do these things.

To protect your rights, we need to prevent others from denying you
these rights or asking you to surrender the rights.  Therefore, you
have certain responsibilities if you distribute copies of the
software, or if you modify it: responsibilities to respect the freedom
of others.

For example, if you distribute copies of such a program, whether
gratis or for a fee, you must pass on to the recipients the same
freedoms that you received.  You must make sure that they, too,
receive or can get the source code.  And you must show them these
terms so they know their rights.

Developers that use the GNU GPL protect your rights with two steps:
(1) assert copyright on the software, and (2) offer you this License
giving you legal permission to copy, distribute and/or modify it.

For the developers' and authors' protection, the GPL clearly explains
that there is no warranty for this free software.  For both users' and
authors' sake, the GPL requires that modified versions be marked as
changed, so that their problems will not be attributed erroneously to
authors of previous versions.

Some devices are designed to deny users access to install or run
modified versions of the software inside them, although the
manufacturer can do so.  This is fundamentally incompatible with the
aim of protecting users' freedom to change the software.  The
systematic pattern of such abuse occurs in the area of products for
individuals to use, which is precisely where it is most unacceptable.
Therefore, we have designed this version of the GPL to prohibit the
practice for those products.  If such problems arise substantially in
other domains, we stand ready to extend this provision to those
domains in future versions of the GPL, as needed to protect the
freedom of users.

Finally, every program is threatened constantly by software patents.
States should not allow patents to restrict development and use of
software on general-purpose computers, but in those that do, we wish
to avoid the special danger that patents applied to a free program
could make it effectively proprietary.  To prevent this, the GPL
assures that patents cannot be used to render the program non-free.

The precise terms and conditions for copying, distribution and
modification follow.

@c fakenode --- for prepinfo
@heading TERMS AND CONDITIONS

@enumerate 0
@item Definitions.

``This License'' refers to version 3 of the GNU General Public License.

``Copyright'' also means copyright-like laws that apply to other kinds
of works, such as semiconductor masks.

``The Program'' refers to any copyrightable work licensed under this
License.  Each licensee is addressed as ``you''.  ``Licensees'' and
``recipients'' may be individuals or organizations.

To ``modify'' a work means to copy from or adapt all or part of the work
in a fashion requiring copyright permission, other than the making of
an exact copy.  The resulting work is called a ``modified version'' of
the earlier work or a work ``based on'' the earlier work.

A ``covered work'' means either the unmodified Program or a work based
on the Program.

To ``propagate'' a work means to do anything with it that, without
permission, would make you directly or secondarily liable for
infringement under applicable copyright law, except executing it on a
computer or modifying a private copy.  Propagation includes copying,
distribution (with or without modification), making available to the
public, and in some countries other activities as well.

To ``convey'' a work means any kind of propagation that enables other
parties to make or receive copies.  Mere interaction with a user
through a computer network, with no transfer of a copy, is not
conveying.

An interactive user interface displays ``Appropriate Legal Notices'' to
the extent that it includes a convenient and prominently visible
feature that (1) displays an appropriate copyright notice, and (2)
tells the user that there is no warranty for the work (except to the
extent that warranties are provided), that licensees may convey the
work under this License, and how to view a copy of this License.  If
the interface presents a list of user commands or options, such as a
menu, a prominent item in the list meets this criterion.

@item Source Code.

The ``source code'' for a work means the preferred form of the work for
making modifications to it.  ``Object code'' means any non-source form
of a work.

A ``Standard Interface'' means an interface that either is an official
standard defined by a recognized standards body, or, in the case of
interfaces specified for a particular programming language, one that
is widely used among developers working in that language.

The ``System Libraries'' of an executable work include anything, other
than the work as a whole, that (a) is included in the normal form of
packaging a Major Component, but which is not part of that Major
Component, and (b) serves only to enable use of the work with that
Major Component, or to implement a Standard Interface for which an
implementation is available to the public in source code form.  A
``Major Component'', in this context, means a major essential component
(kernel, window system, and so on) of the specific operating system
(if any) on which the executable work runs, or a compiler used to
produce the work, or an object code interpreter used to run it.

The ``Corresponding Source'' for a work in object code form means all
the source code needed to generate, install, and (for an executable
work) run the object code and to modify the work, including scripts to
control those activities.  However, it does not include the work's
System Libraries, or general-purpose tools or generally available free
programs which are used unmodified in performing those activities but
which are not part of the work.  For example, Corresponding Source
includes interface definition files associated with source files for
the work, and the source code for shared libraries and dynamically
linked subprograms that the work is specifically designed to require,
such as by intimate data communication or control flow between those
subprograms and other parts of the work.

The Corresponding Source need not include anything that users can
regenerate automatically from other parts of the Corresponding Source.

The Corresponding Source for a work in source code form is that same
work.

@item Basic Permissions.

All rights granted under this License are granted for the term of
copyright on the Program, and are irrevocable provided the stated
conditions are met.  This License explicitly affirms your unlimited
permission to run the unmodified Program.  The output from running a
covered work is covered by this License only if the output, given its
content, constitutes a covered work.  This License acknowledges your
rights of fair use or other equivalent, as provided by copyright law.

You may make, run and propagate covered works that you do not convey,
without conditions so long as your license otherwise remains in force.
You may convey covered works to others for the sole purpose of having
them make modifications exclusively for you, or provide you with
facilities for running those works, provided that you comply with the
terms of this License in conveying all material for which you do not
control copyright.  Those thus making or running the covered works for
you must do so exclusively on your behalf, under your direction and
control, on terms that prohibit them from making any copies of your
copyrighted material outside their relationship with you.

Conveying under any other circumstances is permitted solely under the
conditions stated below.  Sublicensing is not allowed; section 10
makes it unnecessary.

@item Protecting Users' Legal Rights From Anti-Circumvention Law.

No covered work shall be deemed part of an effective technological
measure under any applicable law fulfilling obligations under article
11 of the WIPO copyright treaty adopted on 20 December 1996, or
similar laws prohibiting or restricting circumvention of such
measures.

When you convey a covered work, you waive any legal power to forbid
circumvention of technological measures to the extent such
circumvention is effected by exercising rights under this License with
respect to the covered work, and you disclaim any intention to limit
operation or modification of the work as a means of enforcing, against
the work's users, your or third parties' legal rights to forbid
circumvention of technological measures.

@item Conveying Verbatim Copies.

You may convey verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice;
keep intact all notices stating that this License and any
non-permissive terms added in accord with section 7 apply to the code;
keep intact all notices of the absence of any warranty; and give all
recipients a copy of this License along with the Program.

You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee.

@item Conveying Modified Source Versions.

You may convey a work based on the Program, or the modifications to
produce it from the Program, in the form of source code under the
terms of section 4, provided that you also meet all of these
conditions:

@enumerate a
@item
The work must carry prominent notices stating that you modified it,
and giving a relevant date.

@item
The work must carry prominent notices stating that it is released
under this License and any conditions added under section 7.  This
requirement modifies the requirement in section 4 to ``keep intact all
notices''.

@item
You must license the entire work, as a whole, under this License to
anyone who comes into possession of a copy.  This License will
therefore apply, along with any applicable section 7 additional terms,
to the whole of the work, and all its parts, regardless of how they
are packaged.  This License gives no permission to license the work in
any other way, but it does not invalidate such permission if you have
separately received it.

@item
If the work has interactive user interfaces, each must display
Appropriate Legal Notices; however, if the Program has interactive
interfaces that do not display Appropriate Legal Notices, your work
need not make them do so.
@end enumerate

A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work,
and which are not combined with it such as to form a larger program,
in or on a volume of a storage or distribution medium, is called an
``aggregate'' if the compilation and its resulting copyright are not
used to limit the access or legal rights of the compilation's users
beyond what the individual works permit.  Inclusion of a covered work
in an aggregate does not cause this License to apply to the other
parts of the aggregate.

@item  Conveying Non-Source Forms.

You may convey a covered work in object code form under the terms of
sections 4 and 5, provided that you also convey the machine-readable
Corresponding Source under the terms of this License, in one of these
ways:

@enumerate a
@item
Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by the
Corresponding Source fixed on a durable physical medium customarily
used for software interchange.

@item
Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by a written
offer, valid for at least three years and valid for as long as you
offer spare parts or customer support for that product model, to give
anyone who possesses the object code either (1) a copy of the
Corresponding Source for all the software in the product that is
covered by this License, on a durable physical medium customarily used
for software interchange, for a price no more than your reasonable
cost of physically performing this conveying of source, or (2) access
to copy the Corresponding Source from a network server at no charge.

@item
Convey individual copies of the object code with a copy of the written
offer to provide the Corresponding Source.  This alternative is
allowed only occasionally and noncommercially, and only if you
received the object code with such an offer, in accord with subsection
6b.

@item
Convey the object code by offering access from a designated place
(gratis or for a charge), and offer equivalent access to the
Corresponding Source in the same way through the same place at no
further charge.  You need not require recipients to copy the
Corresponding Source along with the object code.  If the place to copy
the object code is a network server, the Corresponding Source may be
on a different server (operated by you or a third party) that supports
equivalent copying facilities, provided you maintain clear directions
next to the object code saying where to find the Corresponding Source.
Regardless of what server hosts the Corresponding Source, you remain
obligated to ensure that it is available for as long as needed to
satisfy these requirements.

@item
Convey the object code using peer-to-peer transmission, provided you
inform other peers where the object code and Corresponding Source of
the work are being offered to the general public at no charge under
subsection 6d.

@end enumerate

A separable portion of the object code, whose source code is excluded
from the Corresponding Source as a System Library, need not be
included in conveying the object code work.

A ``User Product'' is either (1) a ``consumer product'', which means any
tangible personal property which is normally used for personal,
family, or household purposes, or (2) anything designed or sold for
incorporation into a dwelling.  In determining whether a product is a
consumer product, doubtful cases shall be resolved in favor of
coverage.  For a particular product received by a particular user,
``normally used'' refers to a typical or common use of that class of
product, regardless of the status of the particular user or of the way
in which the particular user actually uses, or expects or is expected
to use, the product.  A product is a consumer product regardless of
whether the product has substantial commercial, industrial or
non-consumer uses, unless such uses represent the only significant
mode of use of the product.

``Installation Information'' for a User Product means any methods,
procedures, authorization keys, or other information required to
install and execute modified versions of a covered work in that User
Product from a modified version of its Corresponding Source.  The
information must suffice to ensure that the continued functioning of
the modified object code is in no case prevented or interfered with
solely because modification has been made.

If you convey an object code work under this section in, or with, or
specifically for use in, a User Product, and the conveying occurs as
part of a transaction in which the right of possession and use of the
User Product is transferred to the recipient in perpetuity or for a
fixed term (regardless of how the transaction is characterized), the
Corresponding Source conveyed under this section must be accompanied
by the Installation Information.  But this requirement does not apply
if neither you nor any third party retains the ability to install
modified object code on the User Product (for example, the work has
been installed in ROM).

The requirement to provide Installation Information does not include a
requirement to continue to provide support service, warranty, or
updates for a work that has been modified or installed by the
recipient, or for the User Product in which it has been modified or
installed.  Access to a network may be denied when the modification
itself materially and adversely affects the operation of the network
or violates the rules and protocols for communication across the
network.

Corresponding Source conveyed, and Installation Information provided,
in accord with this section must be in a format that is publicly
documented (and with an implementation available to the public in
source code form), and must require no special password or key for
unpacking, reading or copying.

@item Additional Terms.

``Additional permissions'' are terms that supplement the terms of this
License by making exceptions from one or more of its conditions.
Additional permissions that are applicable to the entire Program shall
be treated as though they were included in this License, to the extent
that they are valid under applicable law.  If additional permissions
apply only to part of the Program, that part may be used separately
under those permissions, but the entire Program remains governed by
this License without regard to the additional permissions.

When you convey a copy of a covered work, you may at your option
remove any additional permissions from that copy, or from any part of
it.  (Additional permissions may be written to require their own
removal in certain cases when you modify the work.)  You may place
additional permissions on material, added by you to a covered work,
for which you have or can give appropriate copyright permission.

Notwithstanding any other provision of this License, for material you
add to a covered work, you may (if authorized by the copyright holders
of that material) supplement the terms of this License with terms:

@enumerate a
@item
Disclaiming warranty or limiting liability differently from the terms
of sections 15 and 16 of this License; or

@item
Requiring preservation of specified reasonable legal notices or author
attributions in that material or in the Appropriate Legal Notices
displayed by works containing it; or

@item
Prohibiting misrepresentation of the origin of that material, or
requiring that modified versions of such material be marked in
reasonable ways as different from the original version; or

@item
Limiting the use for publicity purposes of names of licensors or
authors of the material; or

@item
Declining to grant rights under trademark law for use of some trade
names, trademarks, or service marks; or

@item
Requiring indemnification of licensors and authors of that material by
anyone who conveys the material (or modified versions of it) with
contractual assumptions of liability to the recipient, for any
liability that these contractual assumptions directly impose on those
licensors and authors.
@end enumerate

All other non-permissive additional terms are considered ``further
restrictions'' within the meaning of section 10.  If the Program as you
received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further
restriction, you may remove that term.  If a license document contains
a further restriction but permits relicensing or conveying under this
License, you may add to a covered work material governed by the terms
of that license document, provided that the further restriction does
not survive such relicensing or conveying.

If you add terms to a covered work in accord with this section, you
must place, in the relevant source files, a statement of the
additional terms that apply to those files, or a notice indicating
where to find the applicable terms.

Additional terms, permissive or non-permissive, may be stated in the
form of a separately written license, or stated as exceptions; the
above requirements apply either way.

@item Termination.

You may not propagate or modify a covered work except as expressly
provided under this License.  Any attempt otherwise to propagate or
modify it is void, and will automatically terminate your rights under
this License (including any patent licenses granted under the third
paragraph of section 11).

However, if you cease all violation of this License, then your license
from a particular copyright holder is reinstated (a) provisionally,
unless and until the copyright holder explicitly and finally
terminates your license, and (b) permanently, if the copyright holder
fails to notify you of the violation by some reasonable means prior to
60 days after the cessation.

Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.

Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License.  If your rights have been terminated and not permanently
reinstated, you do not qualify to receive new licenses for the same
material under section 10.

@item Acceptance Not Required for Having Copies.

You are not required to accept this License in order to receive or run
a copy of the Program.  Ancillary propagation of a covered work
occurring solely as a consequence of using peer-to-peer transmission
to receive a copy likewise does not require acceptance.  However,
nothing other than this License grants you permission to propagate or
modify any covered work.  These actions infringe copyright if you do
not accept this License.  Therefore, by modifying or propagating a
covered work, you indicate your acceptance of this License to do so.

@item Automatic Licensing of Downstream Recipients.

Each time you convey a covered work, the recipient automatically
receives a license from the original licensors, to run, modify and
propagate that work, subject to this License.  You are not responsible
for enforcing compliance by third parties with this License.

An ``entity transaction'' is a transaction transferring control of an
organization, or substantially all assets of one, or subdividing an
organization, or merging organizations.  If propagation of a covered
work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.

You may not impose any further restrictions on the exercise of the
rights granted or affirmed under this License.  For example, you may
not impose a license fee, royalty, or other charge for exercise of
rights granted under this License, and you may not initiate litigation
(including a cross-claim or counterclaim in a lawsuit) alleging that
any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program or any portion of it.

@item Patents.

A ``contributor'' is a copyright holder who authorizes use under this
License of the Program or a work on which the Program is based.  The
work thus licensed is called the contributor's ``contributor version''.

A contributor's ``essential patent claims'' are all patent claims owned
or controlled by the contributor, whether already acquired or
hereafter acquired, that would be infringed by some manner, permitted
by this License, of making, using, or selling its contributor version,
but do not include claims that would be infringed only as a
consequence of further modification of the contributor version.  For
purposes of this definition, ``control'' includes the right to grant
patent sublicenses in a manner consistent with the requirements of
this License.

Each contributor grants you a non-exclusive, worldwide, royalty-free
patent license under the contributor's essential patent claims, to
make, use, sell, offer for sale, import and otherwise run, modify and
propagate the contents of its contributor version.

In the following three paragraphs, a ``patent license'' is any express
agreement or commitment, however denominated, not to enforce a patent
(such as an express permission to practice a patent or covenant not to
sue for patent infringement).  To ``grant'' such a patent license to a
party means to make such an agreement or commitment not to enforce a
patent against the party.

If you convey a covered work, knowingly relying on a patent license,
and the Corresponding Source of the work is not available for anyone
to copy, free of charge and under the terms of this License, through a
publicly available network server or other readily accessible means,
then you must either (1) cause the Corresponding Source to be so
available, or (2) arrange to deprive yourself of the benefit of the
patent license for this particular work, or (3) arrange, in a manner
consistent with the requirements of this License, to extend the patent
license to downstream recipients.  ``Knowingly relying'' means you have
actual knowledge that, but for the patent license, your conveying the
covered work in a country, or your recipient's use of the covered work
in a country, would infringe one or more identifiable patents in that
country that you have reason to believe are valid.

If, pursuant to or in connection with a single transaction or
arrangement, you convey, or propagate by procuring conveyance of, a
covered work, and grant a patent license to some of the parties
receiving the covered work authorizing them to use, propagate, modify
or convey a specific copy of the covered work, then the patent license
you grant is automatically extended to all recipients of the covered
work and works based on it.

A patent license is ``discriminatory'' if it does not include within the
scope of its coverage, prohibits the exercise of, or is conditioned on
the non-exercise of one or more of the rights that are specifically
granted under this License.  You may not convey a covered work if you
are a party to an arrangement with a third party that is in the
business of distributing software, under which you make payment to the
third party based on the extent of your activity of conveying the
work, and under which the third party grants, to any of the parties
who would receive the covered work from you, a discriminatory patent
license (a) in connection with copies of the covered work conveyed by
you (or copies made from those copies), or (b) primarily for and in
connection with specific products or compilations that contain the
covered work, unless you entered into that arrangement, or that patent
license was granted, prior to 28 March 2007.

Nothing in this License shall be construed as excluding or limiting
any implied license or other defenses to infringement that may
otherwise be available to you under applicable patent law.

@item No Surrender of Others' Freedom.

If conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License.  If you cannot convey
a covered work so as to satisfy simultaneously your obligations under
this License and any other pertinent obligations, then as a
consequence you may not convey it at all.  For example, if you agree
to terms that obligate you to collect a royalty for further conveying
from those to whom you convey the Program, the only way you could
satisfy both those terms and this License would be to refrain entirely
from conveying the Program.

@item Use with the GNU Affero General Public License.

Notwithstanding any other provision of this License, you have
permission to link or combine any covered work with a work licensed
under version 3 of the GNU Affero General Public License into a single
combined work, and to convey the resulting work.  The terms of this
License will continue to apply to the part which is the covered work,
but the special requirements of the GNU Affero General Public License,
section 13, concerning interaction through a network will apply to the
combination as such.

@item Revised Versions of this License.

The Free Software Foundation may publish revised and/or new versions
of the GNU General Public License from time to time.  Such new
versions will be similar in spirit to the present version, but may
differ in detail to address new problems or concerns.

Each version is given a distinguishing version number.  If the Program
specifies that a certain numbered version of the GNU General Public
License ``or any later version'' applies to it, you have the option of
following the terms and conditions either of that numbered version or
of any later version published by the Free Software Foundation.  If
the Program does not specify a version number of the GNU General
Public License, you may choose any version ever published by the Free
Software Foundation.

If the Program specifies that a proxy can decide which future versions
of the GNU General Public License can be used, that proxy's public
statement of acceptance of a version permanently authorizes you to
choose that version for the Program.

Later license versions may give you additional or different
permissions.  However, no additional obligations are imposed on any
author or copyright holder as a result of your choosing to follow a
later version.

@item Disclaimer of Warranty.

THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM ``AS IS'' WITHOUT
WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND
PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE PROGRAM PROVE
DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR
CORRECTION.

@item Limitation of Liability.

IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR
CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES
ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT
NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR
LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM
TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER
PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

@item Interpretation of Sections 15 and 16.

If the disclaimer of warranty and limitation of liability provided
above cannot be given local legal effect according to their terms,
reviewing courts shall apply local law that most closely approximates
an absolute waiver of all civil liability in connection with the
Program, unless a warranty or assumption of liability accompanies a
copy of the Program in return for a fee.

@end enumerate

@c fakenode --- for prepinfo
@heading END OF TERMS AND CONDITIONS

@c fakenode --- for prepinfo
@heading How to Apply These Terms to Your New Programs

If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these
terms.

To do so, attach the following notices to the program.  It is safest
to attach them to the start of each source file to most effectively
state the exclusion of warranty; and each file should have at least
the ``copyright'' line and a pointer to where the full notice is found.

@smallexample
@var{one line to give the program's name and a brief idea of what it does.}
Copyright (C) @var{year} @var{name of author}

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or (at
your option) any later version.

This program is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program.  If not, see @url{https://www.gnu.org/licenses/}.
@end smallexample

Also add information on how to contact you by electronic and paper mail.

If the program does terminal interaction, make it output a short
notice like this when it starts in an interactive mode:

@smallexample
@var{program} Copyright (C) @var{year} @var{name of author}
This program comes with ABSOLUTELY NO WARRANTY; for details type @samp{show w}.
This is free software, and you are welcome to redistribute it
under certain conditions; type @samp{show c} for details.
@end smallexample

The hypothetical commands @samp{show w} and @samp{show c} should show
the appropriate parts of the General Public License.  Of course, your
program's commands might be different; for a GUI interface, you would
use an ``about box''.

You should also get your employer (if you work as a programmer) or school,
if any, to sign a ``copyright disclaimer'' for the program, if necessary.
For more information on this, and how to apply and follow the GNU GPL, see
@url{https://www.gnu.org/licenses/}.

The GNU General Public License does not permit incorporating your
program into proprietary programs.  If your program is a subroutine
library, you may consider it more useful to permit linking proprietary
applications with the library.  If this is what you want to do, use
the GNU Lesser General Public License instead of this License.  But
first, please read @url{https://www.gnu.org/philosophy/why-not-lgpl.html}.

@ifclear FOR_PRINT
@c The GNU Free Documentation License.
@node GNU Free Documentation License
@unnumbered GNU Free Documentation License
@ifnotdocbook
@center Version 1.3, 3 November 2008
@end ifnotdocbook

@docbook
<subtitle>Version 1.3, 3 November 2008</subtitle>
@end docbook

@cindex FDL (Free Documentation License)
@cindex Free Documentation License (FDL)
@cindex GNU Free Documentation License

@c This file is intended to be included within another document,
@c hence no sectioning command or @node.

@display
Copyright @copyright{} 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
@uref{https://fsf.org/}

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
@end display

@enumerate 0
@item
PREAMBLE

The purpose of this License is to make a manual, textbook, or other
functional and useful document @dfn{free} in the sense of freedom: to
assure everyone the effective freedom to copy and redistribute it,
with or without modifying it, either commercially or noncommercially.
Secondarily, this License preserves for the author and publisher a way
to get credit for their work, while not being considered responsible
for modifications made by others.

This License is a kind of ``copyleft'', which means that derivative
works of the document must themselves be free in the same sense.  It
complements the GNU General Public License, which is a copyleft
license designed for free software.

We have designed this License in order to use it for manuals for free
software, because free software needs free documentation: a free
program should come with manuals providing the same freedoms that the
software does.  But this License is not limited to software manuals;
it can be used for any textual work, regardless of subject matter or
whether it is published as a printed book.  We recommend this License
principally for works whose purpose is instruction or reference.

@item
APPLICABILITY AND DEFINITIONS

This License applies to any manual or other work, in any medium, that
contains a notice placed by the copyright holder saying it can be
distributed under the terms of this License.  Such a notice grants a
world-wide, royalty-free license, unlimited in duration, to use that
work under the conditions stated herein.  The ``Document'', below,
refers to any such manual or work.  Any member of the public is a
licensee, and is addressed as ``you''.  You accept the license if you
copy, modify or distribute the work in a way requiring permission
under copyright law.

A ``Modified Version'' of the Document means any work containing the
Document or a portion of it, either copied verbatim, or with
modifications and/or translated into another language.

A ``Secondary Section'' is a named appendix or a front-matter section
of the Document that deals exclusively with the relationship of the
publishers or authors of the Document to the Document's overall
subject (or to related matters) and contains nothing that could fall
directly within that overall subject.  (Thus, if the Document is in
part a textbook of mathematics, a Secondary Section may not explain
any mathematics.)  The relationship could be a matter of historical
connection with the subject or with related matters, or of legal,
commercial, philosophical, ethical or political position regarding
them.

The ``Invariant Sections'' are certain Secondary Sections whose titles
are designated, as being those of Invariant Sections, in the notice
that says that the Document is released under this License.  If a
section does not fit the above definition of Secondary then it is not
allowed to be designated as Invariant.  The Document may contain zero
Invariant Sections.  If the Document does not identify any Invariant
Sections then there are none.

The ``Cover Texts'' are certain short passages of text that are listed,
as Front-Cover Texts or Back-Cover Texts, in the notice that says that
the Document is released under this License.  A Front-Cover Text may
be at most 5 words, and a Back-Cover Text may be at most 25 words.

A ``Transparent'' copy of the Document means a machine-readable copy,
represented in a format whose specification is available to the
general public, that is suitable for revising the document
straightforwardly with generic text editors or (for images composed of
pixels) generic paint programs or (for drawings) some widely available
drawing editor, and that is suitable for input to text formatters or
for automatic translation to a variety of formats suitable for input
to text formatters.  A copy made in an otherwise Transparent file
format whose markup, or absence of markup, has been arranged to thwart
or discourage subsequent modification by readers is not Transparent.
An image format is not Transparent if used for any substantial amount
of text.  A copy that is not ``Transparent'' is called ``Opaque''.

Examples of suitable formats for Transparent copies include plain
@sc{ascii} without markup, Texinfo input format, La@TeX{} input
format, @acronym{SGML} or @acronym{XML} using a publicly available
@acronym{DTD}, and standard-conforming simple @acronym{HTML},
PostScript or @acronym{PDF} designed for human modification.  Examples
of transparent image formats include @acronym{PNG}, @acronym{XCF} and
@acronym{JPG}.  Opaque formats include proprietary formats that can be
read and edited only by proprietary word processors, @acronym{SGML} or
@acronym{XML} for which the @acronym{DTD} and/or processing tools are
not generally available, and the machine-generated @acronym{HTML},
PostScript or @acronym{PDF} produced by some word processors for
output purposes only.

The ``Title Page'' means, for a printed book, the title page itself,
plus such following pages as are needed to hold, legibly, the material
this License requires to appear in the title page.  For works in
formats which do not have any title page as such, ``Title Page'' means
the text near the most prominent appearance of the work's title,
preceding the beginning of the body of the text.

The ``publisher'' means any person or entity that distributes copies
of the Document to the public.

A section ``Entitled XYZ'' means a named subunit of the Document whose
title either is precisely XYZ or contains XYZ in parentheses following
text that translates XYZ in another language.  (Here XYZ stands for a
specific section name mentioned below, such as ``Acknowledgements'',
``Dedications'', ``Endorsements'', or ``History''.)  To ``Preserve the Title''
of such a section when you modify the Document means that it remains a
section ``Entitled XYZ'' according to this definition.

The Document may include Warranty Disclaimers next to the notice which
states that this License applies to the Document.  These Warranty
Disclaimers are considered to be included by reference in this
License, but only as regards disclaiming warranties: any other
implication that these Warranty Disclaimers may have is void and has
no effect on the meaning of this License.

@item
VERBATIM COPYING

You may copy and distribute the Document in any medium, either
commercially or noncommercially, provided that this License, the
copyright notices, and the license notice saying this License applies
to the Document are reproduced in all copies, and that you add no other
conditions whatsoever to those of this License.  You may not use
technical measures to obstruct or control the reading or further
copying of the copies you make or distribute.  However, you may accept
compensation in exchange for copies.  If you distribute a large enough
number of copies you must also follow the conditions in section 3.

You may also lend copies, under the same conditions stated above, and
you may publicly display copies.

@item
COPYING IN QUANTITY

If you publish printed copies (or copies in media that commonly have
printed covers) of the Document, numbering more than 100, and the
Document's license notice requires Cover Texts, you must enclose the
copies in covers that carry, clearly and legibly, all these Cover
Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
the back cover.  Both covers must also clearly and legibly identify
you as the publisher of these copies.  The front cover must present
the full title with all words of the title equally prominent and
visible.  You may add other material on the covers in addition.
Copying with changes limited to the covers, as long as they preserve
the title of the Document and satisfy these conditions, can be treated
as verbatim copying in other respects.

If the required texts for either cover are too voluminous to fit
legibly, you should put the first ones listed (as many as fit
reasonably) on the actual cover, and continue the rest onto adjacent
pages.

If you publish or distribute Opaque copies of the Document numbering
more than 100, you must either include a machine-readable Transparent
copy along with each Opaque copy, or state in or with each Opaque copy
a computer-network location from which the general network-using
public has access to download using public-standard network protocols
a complete Transparent copy of the Document, free of added material.
If you use the latter option, you must take reasonably prudent steps,
when you begin distribution of Opaque copies in quantity, to ensure
that this Transparent copy will remain thus accessible at the stated
location until at least one year after the last time you distribute an
Opaque copy (directly or through your agents or retailers) of that
edition to the public.

It is requested, but not required, that you contact the authors of the
Document well before redistributing any large number of copies, to give
them a chance to provide you with an updated version of the Document.

@item
MODIFICATIONS

You may copy and distribute a Modified Version of the Document under
the conditions of sections 2 and 3 above, provided that you release
the Modified Version under precisely this License, with the Modified
Version filling the role of the Document, thus licensing distribution
and modification of the Modified Version to whoever possesses a copy
of it.  In addition, you must do these things in the Modified Version:

@enumerate A
@item
Use in the Title Page (and on the covers, if any) a title distinct
from that of the Document, and from those of previous versions
(which should, if there were any, be listed in the History section
of the Document).  You may use the same title as a previous version
if the original publisher of that version gives permission.

@item
List on the Title Page, as authors, one or more persons or entities
responsible for authorship of the modifications in the Modified
Version, together with at least five of the principal authors of the
Document (all of its principal authors, if it has fewer than five),
unless they release you from this requirement.

@item
State on the Title page the name of the publisher of the
Modified Version, as the publisher.

@item
Preserve all the copyright notices of the Document.

@item
Add an appropriate copyright notice for your modifications
adjacent to the other copyright notices.

@item
Include, immediately after the copyright notices, a license notice
giving the public permission to use the Modified Version under the
terms of this License, in the form shown in the Addendum below.

@item
Preserve in that license notice the full lists of Invariant Sections
and required Cover Texts given in the Document's license notice.

@item
Include an unaltered copy of this License.

@item
Preserve the section Entitled ``History'', Preserve its Title, and add
to it an item stating at least the title, year, new authors, and
publisher of the Modified Version as given on the Title Page.  If
there is no section Entitled ``History'' in the Document, create one
stating the title, year, authors, and publisher of the Document as
given on its Title Page, then add an item describing the Modified
Version as stated in the previous sentence.

@item
Preserve the network location, if any, given in the Document for
public access to a Transparent copy of the Document, and likewise
the network locations given in the Document for previous versions
it was based on.  These may be placed in the ``History'' section.
You may omit a network location for a work that was published at
least four years before the Document itself, or if the original
publisher of the version it refers to gives permission.

@item
For any section Entitled ``Acknowledgements'' or ``Dedications'', Preserve
the Title of the section, and preserve in the section all the
substance and tone of each of the contributor acknowledgements and/or
dedications given therein.

@item
Preserve all the Invariant Sections of the Document,
unaltered in their text and in their titles.  Section numbers
or the equivalent are not considered part of the section titles.

@item
Delete any section Entitled ``Endorsements''.  Such a section
may not be included in the Modified Version.

@item
Do not retitle any existing section to be Entitled ``Endorsements'' or
to conflict in title with any Invariant Section.

@item
Preserve any Warranty Disclaimers.
@end enumerate

If the Modified Version includes new front-matter sections or
appendices that qualify as Secondary Sections and contain no material
copied from the Document, you may at your option designate some or all
of these sections as invariant.  To do this, add their titles to the
list of Invariant Sections in the Modified Version's license notice.
These titles must be distinct from any other section titles.

You may add a section Entitled ``Endorsements'', provided it contains
nothing but endorsements of your Modified Version by various
parties---for example, statements of peer review or that the text has
been approved by an organization as the authoritative definition of a
standard.

You may add a passage of up to five words as a Front-Cover Text, and a
passage of up to 25 words as a Back-Cover Text, to the end of the list
of Cover Texts in the Modified Version.  Only one passage of
Front-Cover Text and one of Back-Cover Text may be added by (or
through arrangements made by) any one entity.  If the Document already
includes a cover text for the same cover, previously added by you or
by arrangement made by the same entity you are acting on behalf of,
you may not add another; but you may replace the old one, on explicit
permission from the previous publisher that added the old one.

The author(s) and publisher(s) of the Document do not by this License
give permission to use their names for publicity for or to assert or
imply endorsement of any Modified Version.

@item
COMBINING DOCUMENTS

You may combine the Document with other documents released under this
License, under the terms defined in section 4 above for modified
versions, provided that you include in the combination all of the
Invariant Sections of all of the original documents, unmodified, and
list them all as Invariant Sections of your combined work in its
license notice, and that you preserve all their Warranty Disclaimers.

The combined work need only contain one copy of this License, and
multiple identical Invariant Sections may be replaced with a single
copy.  If there are multiple Invariant Sections with the same name but
different contents, make the title of each such section unique by
adding at the end of it, in parentheses, the name of the original
author or publisher of that section if known, or else a unique number.
Make the same adjustment to the section titles in the list of
Invariant Sections in the license notice of the combined work.

In the combination, you must combine any sections Entitled ``History''
in the various original documents, forming one section Entitled
``History''; likewise combine any sections Entitled ``Acknowledgements'',
and any sections Entitled ``Dedications''.  You must delete all
sections Entitled ``Endorsements.''

@item
COLLECTIONS OF DOCUMENTS

You may make a collection consisting of the Document and other documents
released under this License, and replace the individual copies of this
License in the various documents with a single copy that is included in
the collection, provided that you follow the rules of this License for
verbatim copying of each of the documents in all other respects.

You may extract a single document from such a collection, and distribute
it individually under this License, provided you insert a copy of this
License into the extracted document, and follow this License in all
other respects regarding verbatim copying of that document.

@item
AGGREGATION WITH INDEPENDENT WORKS

A compilation of the Document or its derivatives with other separate
and independent documents or works, in or on a volume of a storage or
distribution medium, is called an ``aggregate'' if the copyright
resulting from the compilation is not used to limit the legal rights
of the compilation's users beyond what the individual works permit.
When the Document is included in an aggregate, this License does not
apply to the other works in the aggregate which are not themselves
derivative works of the Document.

If the Cover Text requirement of section 3 is applicable to these
copies of the Document, then if the Document is less than one half of
the entire aggregate, the Document's Cover Texts may be placed on
covers that bracket the Document within the aggregate, or the
electronic equivalent of covers if the Document is in electronic form.
Otherwise they must appear on printed covers that bracket the whole
aggregate.

@item
TRANSLATION

Translation is considered a kind of modification, so you may
distribute translations of the Document under the terms of section 4.
Replacing Invariant Sections with translations requires special
permission from their copyright holders, but you may include
translations of some or all Invariant Sections in addition to the
original versions of these Invariant Sections.  You may include a
translation of this License, and all the license notices in the
Document, and any Warranty Disclaimers, provided that you also include
the original English version of this License and the original versions
of those notices and disclaimers.  In case of a disagreement between
the translation and the original version of this License or a notice
or disclaimer, the original version will prevail.

If a section in the Document is Entitled ``Acknowledgements'',
``Dedications'', or ``History'', the requirement (section 4) to Preserve
its Title (section 1) will typically require changing the actual
title.

@item
TERMINATION

You may not copy, modify, sublicense, or distribute the Document
except as expressly provided under this License.  Any attempt
otherwise to copy, modify, sublicense, or distribute it is void, and
will automatically terminate your rights under this License.

However, if you cease all violation of this License, then your license
from a particular copyright holder is reinstated (a) provisionally,
unless and until the copyright holder explicitly and finally
terminates your license, and (b) permanently, if the copyright holder
fails to notify you of the violation by some reasonable means prior to
60 days after the cessation.

Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.

Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License.  If your rights have been terminated and not permanently
reinstated, receipt of a copy of some or all of the same material does
not give you any rights to use it.

@item
FUTURE REVISIONS OF THIS LICENSE

The Free Software Foundation may publish new, revised versions
of the GNU Free Documentation License from time to time.  Such new
versions will be similar in spirit to the present version, but may
differ in detail to address new problems or concerns.  See
@uref{https://www.gnu.org/copyleft/}.

Each version of the License is given a distinguishing version number.
If the Document specifies that a particular numbered version of this
License ``or any later version'' applies to it, you have the option of
following the terms and conditions either of that specified version or
of any later version that has been published (not as a draft) by the
Free Software Foundation.  If the Document does not specify a version
number of this License, you may choose any version ever published (not
as a draft) by the Free Software Foundation.  If the Document
specifies that a proxy can decide which future versions of this
License can be used, that proxy's public statement of acceptance of a
version permanently authorizes you to choose that version for the
Document.

@item
RELICENSING

``Massive Multiauthor Collaboration Site'' (or ``MMC Site'') means any
World Wide Web server that publishes copyrightable works and also
provides prominent facilities for anybody to edit those works.  A
public wiki that anybody can edit is an example of such a server.  A
``Massive Multiauthor Collaboration'' (or ``MMC'') contained in the
site means any set of copyrightable works thus published on the MMC
site.

``CC-BY-SA'' means the Creative Commons Attribution-Share Alike 3.0
license published by Creative Commons Corporation, a not-for-profit
corporation with a principal place of business in San Francisco,
California, as well as future copyleft versions of that license
published by that same organization.

``Incorporate'' means to publish or republish a Document, in whole or
in part, as part of another Document.

An MMC is ``eligible for relicensing'' if it is licensed under this
License, and if all works that were first published under this License
somewhere other than this MMC, and subsequently incorporated in whole
or in part into the MMC, (1) had no cover texts or invariant sections,
and (2) were thus incorporated prior to November 1, 2008.

The operator of an MMC Site may republish an MMC contained in the site
under CC-BY-SA on the same site at any time before August 1, 2009,
provided the MMC is eligible for relicensing.

@end enumerate

@c fakenode --- for prepinfo
@unnumberedsec ADDENDUM: How to use this License for your documents

To use this License in a document you have written, include a copy of
the License in the document and put the following copyright and
license notices just after the title page:

@smallexample
@group
  Copyright (C)  @var{year}  @var{your name}.
  Permission is granted to copy, distribute and/or modify this document
  under the terms of the GNU Free Documentation License, Version 1.3
  or any later version published by the Free Software Foundation;
  with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
  Texts.  A copy of the license is included in the section entitled ``GNU
  Free Documentation License''.
@end group
@end smallexample

If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
replace the ``with@dots{}Texts.'' line with this:

@smallexample
@group
    with the Invariant Sections being @var{list their titles}, with
    the Front-Cover Texts being @var{list}, and with the Back-Cover Texts
    being @var{list}.
@end group
@end smallexample

If you have Invariant Sections without Cover Texts, or some other
combination of the three, merge those two alternatives to suit the
situation.

If your document contains nontrivial examples of program code, we
recommend releasing these examples in parallel under your choice of
free software license, such as the GNU General Public License,
to permit their use in free software.

@end ifclear

@ifnotdocbook
@node Index
@unnumbered Index
@end ifnotdocbook
@printindex cp

@bye

Unresolved Issues:
------------------
1. From ADR.

   Robert J. Chassell points out that awk programs should have some indication
   of how to use them.  It would be useful to perhaps have a "programming
   style" section of the manual that would include this and other tips.

Consistency issues:
	/.../ regexps are in @code, not @samp
	".." strings are in @code, not @samp
	no @print before @dots
	values of expressions in the text (@code{x} has the value 15),
		should be in roman, not @code
	Use   TAB   and not   tab
	Use   ESC   and not   ESCAPE
	Use   space and not   blank	to describe the space bar's character
	The term "blank" is thus basically reserved for "blank lines" etc.
	To make dark corners work, the @value{DARKCORNER} has to be outside
		closing `.' of a sentence and after (pxref{...}).
	Make sure that each @value{DARKCORNER} has an index entry, and
		also that each `@cindex dark corner' has an @value{DARKCORNER}.
	" " should have an @w{} around it
	Use "non-" only with language names or acronyms, or the words bug and option and null
	Use @command{ftp} when talking about anonymous ftp
	Use uppercase and lowercase, not "upper-case" and "lower-case"
		or "upper case" and "lower case"
	Use "single precision" and "double precision", not "single-precision" or "double-precision"
	Use alphanumeric, not alpha-numeric
	Use POSIX-compliant, not POSIX compliant
	Use --foo, not -Wfoo when describing long options
	Use "Bell Laboratories", but not "Bell Labs".
	Use "behavior" instead of "behaviour".
	Use "coprocess" instead of "co-process".
	Use "zeros" instead of "zeroes".
	Use "nonzero" not "non-zero".
	Use "runtime" not "run time" or "run-time".
	Use "command-line" as an adjective and "command line" as a noun.
	Use "online" not "on-line".
	Use "whitespace" not "white space".
	Use "Input/Output", not "input/output". Also "I/O", not "i/o".
	Use "lefthand"/"righthand", not "left-hand"/"right-hand".
	Use "workaround", not "work-around".
	Use "startup"/"cleanup", not "start-up"/"clean-up"
	Use "filesystem", not "file system"
	Use @code{do}, and not @code{do}-@code{while}, except where
		actually discussing the do-while.
	Use "versus" in text and "vs." in index entries
	Use @code{"C"} for the C locale, not ``C'' or @samp{C}.
	The words "a", "and", "as", "between", "for", "from", "in", "of",
		"on", "that", "the", "to", "with", and "without",
		should not be capitalized in @chapter, @section etc.
		"Into" and "How" should.
	Search for @dfn; make sure important items are also indexed.
	"e.g." should always be followed by a comma.
	"i.e." should always be followed by a comma.
	The numbers zero through ten should be spelled out, except when
		talking about file descriptor numbers. > 10 and < 0, it's
		ok to use numbers.
	For most cases, do NOT put a comma before "and", "or" or "but".
		But exercise taste with this rule.
	Don't show the awk command with a program in quotes when it's
		just the program.  I.e.

			{
				....
			}

		not
			awk '{
				...
			}'

	Do show it when showing command-line arguments, data files, etc, even
		if there is no output shown.

	Use numbered lists only to show a sequential series of steps.

	Use @code{xxx} for the xxx operator in indexing statements, not @samp.
	Use MS-Windows not MS Windows
	Use MS-DOS not MS DOS
	Use an empty set of parentheses after built-in and awk function names.
	Use "multiFOO" without a hyphen.
	Use "time zone" as two words, not "timezone".

Date: Wed, 13 Apr 94 15:20:52 -0400
From: rms@gnu.org (Richard Stallman)
To: gnu-prog@gnu.org
Subject: A reminder: no pathnames in GNU

It's a GNU convention to use the term "file name" for the name of a
file, never "pathname".  We use the term "path" for search paths,
which are lists of file names.  Using it for a single file name as
well is potentially confusing to users.

So please check any documentation you maintain, if you think you might
have used "pathname".

Note that "file name" should be two words when it appears as ordinary
text.  It's ok as one word when it's a metasyntactic variable, though.

------------------------
ORA uses filename, thus the macro.

Suggestions:
------------

Better sidebars can almost sort of be done with:

	@ifdocbook
	@macro @sidebar{title, content}
	@inlinefmt{docbook, <sidebar><title>}
	\title\
	@inlinefmt{docbook, </title>}
	\content\
	@inlinefmt{docbook, </sidebar>}
	@end macro
	@end ifdocbook


	@ifnotdocbook
	@macro @sidebar{title, content}
	@cartouche
	@center @b{\title\}

	\content\
	@end cartouche
	@end macro
	@end ifnotdocbook

But to use it you have to say

	@sidebar{Title Here,
	@include file-with-content
	}

which sorta sucks.

TODO:
Add a section explaining recursion from ground zero. Probably
easiest to do it with factorial as the example. Explain that
recursion needs a stopping condition. Thanks to
Bill Duncan <bduncan@beachnet.org> for the suggestion.