path: root/doc/gawk.1

.ds PX \s-1POSIX\s+1
.ds UX \s-1UNIX\s+1
.ds GN \s-1GNU\s+1
.ds AK \s-1AWK\s+1
.ds EP \fIGAWK: Effective AWK Programming\fP
.if !\n(.g \{\
.	if !\w|\*(lq| \{\
.		ds lq ``
.		if \w'\(lq' .ds lq "\(lq
.	\}
.	if !\w|\*(rq| \{\
.		ds rq ''
.		if \w'\(rq' .ds rq "\(rq
.	\}
.\}
.TH GAWK 1 "Jan 28 2014" "Free Software Foundation" "Utility Commands"
.SH NAME
gawk \- pattern scanning and processing language
.SH SYNOPSIS
.B gawk
[ \*(PX or \*(GN style options ]
.B \-f
.I program-file
[
.B \-\^\-
] file .\|.\|.
.br
.B gawk
[ \*(PX or \*(GN style options ]
[
.B \-\^\-
]
.I program-text
file .\|.\|.
.SH DESCRIPTION
.I Gawk
is the \*(GN Project's implementation of the \*(AK programming language.
It conforms to the definition of the language in
the \*(PX 1003.1 Standard.
This version in turn is based on the description in
.IR "The AWK Programming Language" ,
by Aho, Kernighan, and Weinberger.
.I Gawk
provides the additional features found in the current version
of Brian Kernighan's
.I awk
and a number of \*(GN-specific extensions.
.PP
The command line consists of options to
.I gawk
itself, the \*(AK program text (if not supplied via the
.B \-f
or
.B \-\^\-file
options), and values to be made
available in the
.B ARGC
and
.B ARGV
pre-defined \*(AK variables.
.PP
When
.I gawk
is invoked with the
.B \-\^\-profile
option, it starts gathering profiling statistics
from the execution of the program.
.I Gawk
runs more slowly in this mode, and automatically produces an execution
profile in the file
.B awkprof.out
when done.
See the
.B \-\^\-profile
option, below.
.PP
.I Gawk
also has an integrated debugger. An interactive debugging session can
be started by supplying the
.B \-\^\-debug
option to the command line. In this mode of execution,
.I gawk
loads the
AWK source code and then prompts for debugging commands.
.I Gawk
can only debug AWK program source provided with the
.B \-f
option.
The debugger is documented in \*(EP.
.SH OPTION FORMAT
.PP
.I Gawk
options may be either traditional \*(PX-style one letter options,
or \*(GN-style long options.  \*(PX options start with a single \*(lq\-\*(rq,
while long options start with \*(lq\-\^\-\*(rq.
Long options are provided for both \*(GN-specific features and
for \*(PX-mandated features.
.PP
.IR Gawk -specific
options are typically used in long-option form.
Arguments to long options are either joined with the option
by an
.B =
sign, with no intervening spaces, or they may be provided in the
next command line argument.
Long options may be abbreviated, as long as the abbreviation
remains unique.
.PP
Additionally, every long option has a corresponding short
option, so that the option's functionality may be used from
within
.B #!
executable scripts.
.SH OPTIONS
.PP
.I Gawk
accepts the following options.
Standard options are listed first, followed by options for
.I gawk
extensions, listed alphabetically by short option.
.TP
.PD 0
.BI \-f " program-file"
.TP
.PD
.BI \-\^\-file " program-file"
Read the \*(AK program source from the file
.IR program-file ,
instead of from the first command line argument.
Multiple
.B \-f
(or
.BR \-\^\-file )
options may be used.
.TP
.PD 0
.BI \-F " fs"
.TP
.PD
.BI \-\^\-field-separator " fs"
Use
.I fs
for the input field separator (the value of the
.B FS
predefined
variable).
.TP
.PD 0
\fB\-v\fI var\fB\^=\^\fIval\fR
.TP
.PD
\fB\-\^\-assign \fIvar\fB\^=\^\fIval\fR
Assign the value
.I val
to the variable
.IR var ,
before execution of the program begins.
Such variable values are available to the
.B BEGIN
rule of an \*(AK program.
.TP
.PD 0
.B \-b
.TP
.PD
.B \-\^\-characters\-as\-bytes
Treat all input data as single-byte characters. In other words,
don't pay any attention to the locale information when attempting to
process strings as multibyte characters.
The
.B "\-\^\-posix"
option overrides this one.
.bp
.TP
.PD 0
.B \-c
.TP
.PD
.B \-\^\-traditional
Run in
.I compatibility
mode.  In compatibility mode,
.I gawk
behaves identically to Brian Kernighan's
.IR awk ;
none of the \*(GN-specific extensions are recognized.
.\" The use of
.\" .B \-\^\-traditional
.\" is preferred over the other forms of this option.
See
.BR "GNU EXTENSIONS" ,
below, for more information.
.TP
.PD 0
.B \-C
.TP
.PD
.B \-\^\-copyright
Print the short version of the \*(GN copyright information message on
the standard output and exit successfully.
.TP
.PD 0
\fB\-d\fR[\fIfile\fR]
.TP
.PD
\fB\-\^\-dump-variables\fR[\fB=\fIfile\fR]
Print a sorted list of global variables, their types and final values to
.IR file .
If no
.I file
is provided,
.I gawk
uses a file named
.B awkvars.out
in the current directory.
.sp .5
Having a list of all the 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
.BR i ,
.BR j ,
and so on.)
.TP
.PD 0
\fB\-D\fR[\fIfile\fR]
.TP
.PD
\fB\-\^\-debug\fR[\fB=\fIfile\fR]
Enable debugging of \*(AK programs.
By default, the debugger reads commands interactively from the terminal.
The optional
.IR file
argument specifies a file with a list
of commands for the debugger to execute non-interactively.
.TP
.PD 0
.BI "\-e " program-text
.TP
.PD
.BI \-\^\-source " program-text"
Use
.I program-text
as \*(AK program source code.
This option allows the easy intermixing of library functions (used via the
.B \-f
and
.B \-\^\-file
options) with source code entered on the command line.
It is intended primarily for medium to large \*(AK programs used
in shell scripts.
.TP
.PD 0
.BI "\-E " file
.TP
.PD
.BI \-\^\-exec " file"
Similar to
.BR \-f ,
however, this is option is the last one processed.
This should be used with
.B #!
scripts, particularly for CGI applications, to avoid
passing in options or source code (!) on the command line
from a URL.
This option disables command-line variable assignments.
.TP
.PD 0
.B \-g
.TP
.PD
.B \-\^\-gen\-pot
Scan and parse the \*(AK program, and generate a \*(GN
.B \&.pot
(Portable Object Template)
format file on standard output with entries for all localizable
strings in the program.  The program itself is not executed.
See the \*(GN
.I gettext
distribution for more information on
.B \&.pot
files.
.TP
.PD 0
.B \-h
.TP
.PD
.B \-\^\-help
Print a relatively short summary of the available options on
the standard output.
(Per the
.IR "GNU Coding Standards" ,
these options cause an immediate, successful exit.)
.TP
.PD 0
.BI "\-i " include-file
.TP
.PD
.BI \-\^\-include " include-file"
Load an awk source library.
This searches for the library using the
.B AWKPATH
environment variable.  If the initial search fails, another attempt will
be made after appending the
.B \&.awk
suffix.  The file will be loaded only
once (i.e., duplicates are eliminated), and the code does not constitute
the main program source.
.TP
.PD 0
.BI "\-l " lib
.TP
.PD
.BI \-\^\-load " lib"
Load a shared library
.IR lib .
This searches for the library using the
.B AWKLIBPATH
environment variable.  If the initial search fails, another attempt will
be made after appending the default shared library suffix for the platform.
The library initialization routine is expected to be named
.BR dl_load() .
.TP
.PD 0
.BR "\-L " [ \fIvalue\fR ]
.TP
.PD
.BR \-\^\-lint [ =\fIvalue\fR ]
Provide warnings about constructs that are
dubious or non-portable to other \*(AK implementations.
With an optional argument of
.BR fatal ,
lint warnings become fatal errors.
This may be drastic, but its use will certainly encourage the
development of cleaner \*(AK programs.
With an optional argument of
.BR invalid ,
only warnings about things that are
actually invalid are issued. (This is not fully implemented yet.)
.TP
.PD 0
.B \-M
.TP
.PD
.B \-\^\-bignum
Force arbitrary precision arithmetic on numbers. This option has
no effect if
.I gawk
is not compiled to use the GNU MPFR and MP libraries.
.TP
.PD 0
.B \-n
.TP
.PD
.B "\-\^\-non\-decimal\-data"
Recognize octal and hexadecimal values in input data.
.I "Use this option with great caution!"
.TP
.PD 0
.B \-N
.TP
.PD
.B \-\^\-use\-lc\-numeric
This forces
.I gawk
to use the locale's decimal point character when parsing input data.
Although the POSIX standard requires this behavior, and
.I gawk
does so when
.B \-\^\-posix
is in effect, the default is to follow traditional behavior and use a
period as the decimal point, even in locales where the period is not the
decimal point character.  This option overrides the default behavior,
without the full draconian strictness of the
.B \-\^\-posix
option.
.ig
.\" This option is left undocumented, on purpose.
.TP
.PD 0
.B "\-W nostalgia"
.TP
.PD
.B \-\^\-nostalgia
Provide a moment of nostalgia for long time
.I awk
users.
..
.TP
.PD 0
\fB\-o\fR[\fIfile\fR]
.TP
.PD
\fB\-\^\-pretty-print\fR[\fB=\fIfile\fR]
Output a pretty printed version of the program to
.IR file .
If no
.I file
is provided,
.I gawk
uses a file named
.B awkprof.out
in the current directory.
.TP
.PD 0
.B \-O
.TP
.PD
.B \-\^\-optimize
Enable optimizations upon the internal representation of the program.
Currently, this includes simple constant-folding, and tail call
elimination for recursive functions. The
.I gawk
maintainer hopes to add additional optimizations over time.
.TP
.PD 0
\fB\-p\fR[\fIprof-file\fR]
.TP
.PD
\fB\-\^\-profile\fR[\fB=\fIprof-file\fR]
Start a profiling session, and send the profiling data to
.IR prof-file .
The default is
.BR awkprof.out .
The profile contains execution counts of each statement in the program
in the left margin and function call counts for each user-defined function.
.TP
.PD 0
.B \-P
.TP
.PD
.B \-\^\-posix
This turns on
.I compatibility
mode, with the following additional restrictions:
.RS
.TP "\w'\(bu'u+1n"
\(bu
.B \ex
escape sequences are not recognized.
.TP
\(bu
Only space and tab act as field separators when
.B FS
is set to a single space, newline does not.
.TP
\(bu
You cannot continue lines after
.B ?
and
.BR : .
.TP
\(bu
The synonym
.B func
for the keyword
.B function
is not recognized.
.TP
\(bu
The operators
.B **
and
.B **=
cannot be used in place of
.B ^
and
.BR ^= .
.RE
.TP
.PD 0
.B \-r
.TP
.PD
.B \-\^\-re\-interval
Enable the use of
.I "interval expressions"
in regular expression matching
(see
.BR "Regular Expressions" ,
below).
Interval expressions were not traditionally available in the
\*(AK language.  The \*(PX standard added them, to make
.I awk
and
.I egrep
consistent with each other.
They are enabled by default, but this option remains for use with
.BR \-\^\-traditional .
.TP
.PD 0
.BI \-S
.TP
.PD
.BI \-\^\-sandbox
Runs
.I gawk
in sandbox mode, disabling the
.B system()
function, input redirection with
.BR getline ,
output redirection with
.BR print " and " printf ,
and loading dynamic extensions.
Command execution (through pipelines) is also disabled.
This effectively blocks a script from accessing local resources
(except for the files specified on the command line).
.TP
.PD 0
.B \-t
.TP
.PD
.B \-\^\-lint\-old
Provide warnings about constructs that are
not portable to the original version of \*(UX
.IR awk .
.TP
.PD 0
.B \-V
.TP
.PD
.B \-\^\-version
Print version information for this particular copy of
.I gawk
on the standard output.
This is useful mainly for knowing if the current copy of
.I gawk
on your system
is up to date with respect to whatever the Free Software Foundation
is distributing.
This is also useful when reporting bugs.
(Per the
.IR "GNU Coding Standards" ,
these options cause an immediate, successful exit.)
.TP
.B \-\^\-
Signal the end of options. This is useful to allow further arguments to the
\*(AK program itself to start with a \*(lq\-\*(rq.
This provides consistency with the argument parsing convention used
by most other \*(PX programs.
.PP
In compatibility mode,
any other options are flagged as invalid, but are otherwise ignored.
In normal operation, as long as program text has been supplied, unknown
options are passed on to the \*(AK program in the
.B ARGV
array for processing.  This is particularly useful for running \*(AK
programs via the \*(lq#!\*(rq executable interpreter mechanism.
.PP
For \*(PX compatibility, the
.B \-W
option may be used, followed by the name of a long option.
.SH AWK PROGRAM EXECUTION
.PP
An \*(AK program consists of a sequence of pattern-action statements
and optional function definitions.
.RS
.PP
\fB@include "\fIfilename\fB"
.br
\fB@load "\fIfilename\fB"
.br
\fIpattern\fB	{ \fIaction statements\fB }\fR
.br
\fBfunction \fIname\fB(\fIparameter list\fB) { \fIstatements\fB }\fR
.RE
.PP
.I Gawk
first reads the program source from the
.IR program-file (s)
if specified,
from arguments to
.BR \-\^\-source ,
or from the first non-option argument on the command line.
The
.B \-f
and
.B \-\^\-source
options may be used multiple times on the command line.
.I Gawk
reads the program text as if all the
.IR program-file s
and command line source texts
had been concatenated together.  This is useful for building libraries
of \*(AK functions, without having to include them in each new \*(AK
program that uses them.  It also provides the ability to mix library
functions with command line programs.
.PP
In addition, lines beginning with
.B @include
may be used to include other source files into your program,
making library use even easier.  This is equivalent
to using the
.B \-i
option.
.PP
Lines beginning with
.B @load
may be used to load shared libraries into your program.  This is equivalent
to using the
.B \-l
option.
.PP
The environment variable
.B AWKPATH
specifies a search path to use when finding source files named with
the
.B \-f
and
.B \-i
options.  If this variable does not exist, the default path is
\fB".:/usr/local/share/awk"\fR.
(The actual directory may vary, depending upon how
.I gawk
was built and installed.)
If a file name given to the
.B \-f
option contains a \*(lq/\*(rq character, no path search is performed.
.PP
The environment variable
.B AWKLIBPATH
specifies a search path to use when finding source files named with
the
.B \-l
option.  If this variable does not exist, the default path is
\fB".:/usr/local/lib/gawk"\fR.
(The actual directory may vary, depending upon how
.I gawk
was built and installed.)
.PP
.I Gawk
executes \*(AK programs in the following order.
First,
all variable assignments specified via the
.B \-v
option are performed.
Next,
.I gawk
compiles the program into an internal form.
Then,
.I gawk
executes the code in the
.B BEGIN
rule(s) (if any),
and then proceeds to read
each file named in the
.B ARGV
array (up to
.BR ARGV[ARGC] ).
If there are no files named on the command line,
.I gawk
reads the standard input.
.PP
If a filename on the command line has the form
.IB var = val
it is treated as a variable assignment.  The variable
.I var
will be assigned the value
.IR val .
(This happens after any
.B BEGIN
rule(s) have been run.)
Command line variable assignment
is most useful for dynamically assigning values to the variables
\*(AK uses to control how input is broken into fields and records.
It is also useful for controlling state if multiple passes are needed over
a single data file.
.PP
If the value of a particular element of
.B ARGV
is empty (\fB""\fR),
.I gawk
skips over it.
.PP
For each input file,
if a
.B BEGINFILE
rule exists,
.I gawk
executes the associated code
before processing the contents of the file. Similarly,
.I gawk
executes
the code associated with
.B ENDFILE
after processing the file.
.PP
For each record in the input,
.I gawk
tests to see if it matches any
.I pattern
in the \*(AK program.
For each pattern that the record matches,
.I gawk
executes the associated
.IR action .
The patterns are tested in the order they occur in the program.
.PP
Finally, after all the input is exhausted,
.I gawk
executes the code in the
.B END
rule(s) (if any).
.SS Command Line Directories
.PP
According to POSIX, files named on the
.I awk
command line must be
text files.  The behavior is ``undefined'' if they are not.  Most versions
of
.I awk
treat a directory on the command line as a fatal error.
.PP
Starting with version 4.0 of
.IR gawk ,
a directory on the command line
produces a warning, but is otherwise skipped.  If either of the
.B \-\^\-posix
or
.B \-\^\-traditional
options is given, then
.I gawk
reverts to
treating directories on the command line as a fatal error.
.SH VARIABLES, RECORDS AND FIELDS
\*(AK variables are dynamic; they come into existence when they are
first used.  Their values are either floating-point numbers or strings,
or both,
depending upon how they are used.  \*(AK also has one dimensional
arrays; arrays with multiple dimensions may be simulated.
.I Gawk
provides true arrays of arrays; see
.BR Arrays ,
below.
Several pre-defined variables are set as a program
runs; these are described as needed and summarized below.
.SS Records
Normally, records are separated by newline characters.  You can control how
records are separated by assigning values to the built-in variable
.BR RS .
If
.B RS
is any single character, that character separates records.
Otherwise,
.B RS
is a regular expression.  Text in the input that matches this
regular expression separates the record.
However, in compatibility mode,
only the first character of its string
value is used for separating records.
If
.B RS
is set to the null string, then records are separated by
blank lines.
When
.B RS
is set to the null string, the newline character always acts as
a field separator, in addition to whatever value
.B FS
may have.
.SS Fields
.PP
As each input record is read,
.I gawk
splits the record into
.IR fields ,
using the value of the
.B FS
variable as the field separator.
If
.B FS
is a single character, fields are separated by that character.
If
.B FS
is the null string, then each individual character becomes a
separate field.
Otherwise,
.B FS
is expected to be a full regular expression.
In the special case that
.B FS
is a single space, fields are separated
by runs of spaces and/or tabs and/or newlines.
(But see the section
.BR "POSIX COMPATIBILITY" ,
below).
.BR NOTE :
The value of
.B IGNORECASE
(see below) also affects how fields are split when
.B FS
is a regular expression, and how records are separated when
.B RS
is a regular expression.
.PP
If the
.B FIELDWIDTHS
variable is set to a space separated list of numbers, each field is
expected to have fixed width, and
.I gawk
splits up the record using the specified widths.  The value of
.B FS
is ignored.
Assigning a new value to
.B FS
or
.B FPAT
overrides the use of
.BR FIELDWIDTHS .
.PP
Similarly, if the
.B FPAT
variable is set to a string representing a regular expression,
each field is made up of text that matches that regular expression. In
this case, the regular expression describes the fields themselves,
instead of the text that separates the fields.
Assigning a new value to
.B FS
or
.B FIELDWIDTHS
overrides the use of
.BR FPAT .
.PP
Each field in the input record may be referenced by its position:
.BR $1 ,
.BR $2 ,
and so on.
.B $0
is the whole record.
Fields need not be referenced by constants:
.RS
.PP
.ft B
n = 5
.br
print $n
.ft R
.RE
.PP
prints the fifth field in the input record.
.PP
The variable
.B NF
is set to the total number of fields in the input record.
.PP
References to non-existent fields (i.e., fields after
.BR $NF )
produce the null-string.  However, assigning to a non-existent field
(e.g.,
.BR "$(NF+2) = 5" )
increases the value of
.BR NF ,
creates any intervening fields with the null string as their values, and
causes the value of
.B $0
to be recomputed, with the fields being separated by the value of
.BR OFS .
References to negative numbered fields cause a fatal error.
Decrementing
.B NF
causes the values of fields past the new value to be lost, and the value of
.B $0
to be recomputed, with the fields being separated by the value of
.BR OFS .
.PP
Assigning a value to an existing field
causes the whole record to be rebuilt when
.B $0
is referenced.
Similarly, assigning a value to
.B $0
causes the record to be resplit, creating new
values for the fields.
.SS Built-in Variables
.PP
.IR Gawk\^ "'s"
built-in variables are:
.PP
.TP "\w'\fBFIELDWIDTHS\fR'u+1n"
.B ARGC
The number of command line arguments (does not include options to
.IR gawk ,
or the program source).
.TP
.B ARGIND
The index in
.B ARGV
of the current file being processed.
.TP
.B ARGV
Array of command line arguments.  The array is indexed from
0 to
.B ARGC
\- 1.
Dynamically changing the contents of
.B ARGV
can control the files used for data.
.TP
.B BINMODE
On non-POSIX systems, specifies use of \*(lqbinary\*(rq mode for all file I/O.
Numeric values of 1, 2, or 3, specify that input files, output files, or
all files, respectively, should use binary I/O.
String values of \fB"r"\fR, or \fB"w"\fR specify that input files, or output files,
respectively, should use binary I/O.
String values of \fB"rw"\fR or \fB"wr"\fR specify that all files
should use binary I/O.
Any other string value is treated as \fB"rw"\fR, but generates a warning message.
.TP
.B CONVFMT
The conversion format for numbers, \fB"%.6g"\fR, by default.
.TP
.B ENVIRON
An array containing the values of the current environment.
The array is indexed by the environment variables, each element being
the value of that variable (e.g., \fBENVIRON["HOME"]\fP might be
\fB"/home/arnold"\fR).
Changing this array does not affect the environment seen by programs which
.I gawk
spawns via redirection or the
.B system()
function.
.TP
.B ERRNO
If a system error occurs either doing a redirection for
.BR getline ,
during a read for
.BR getline ,
or during a
.BR close() ,
then
.B ERRNO
will contain
a string describing the error.
The value is subject to translation in non-English locales.
.TP
.B FIELDWIDTHS
A whitespace separated list of field widths.  When set,
.I gawk
parses the input into fields of fixed width, instead of using the
value of the
.B FS
variable as the field separator.
See
.BR Fields ,
above.
.TP
.B FILENAME
The name of the current input file.
If no files are specified on the command line, the value of
.B FILENAME
is \*(lq\-\*(rq.
However,
.B FILENAME
is undefined inside the
.B BEGIN
rule
(unless set by
.BR getline ).
.TP
.B FNR
The input record number in the current input file.
.TP
.B FPAT
A regular expression describing the contents of the
fields in a record.
When set,
.I gawk
parses the input into fields, where the fields match the
regular expression, instead of using the
value of the
.B FS
variable as the field separator.
See
.BR Fields ,
above.
.TP
.B FS
The input field separator, a space by default.  See
.BR Fields ,
above.
.TP
.B FUNCTAB
An array whose indices and corresponding values
are the names of all the user-defined
or extension functions in the program.
.BR NOTE :
You may not use the
.B delete
statement with the
.B FUNCTAB
array.
.TP
.B IGNORECASE
Controls the case-sensitivity of all regular expression
and string operations.  If
.B IGNORECASE
has a non-zero value, then string comparisons and
pattern matching in rules,
field splitting with
.B FS
and
.BR FPAT ,
record separating with
.BR RS ,
regular expression
matching with
.B ~
and
.BR !~ ,
and the
.BR gensub() ,
.BR gsub() ,
.BR index() ,
.BR match() ,
.BR patsplit() ,
.BR split() ,
and
.B sub()
built-in functions all ignore case when doing regular expression
operations.
.BR NOTE :
Array subscripting is
.I not
affected.
However, the
.B asort()
and
.B asorti()
functions are affected.
.sp .5
Thus, if
.B IGNORECASE
is not equal to zero,
.B /aB/
matches all of the strings \fB"ab"\fP, \fB"aB"\fP, \fB"Ab"\fP,
and \fB"AB"\fP.
As with all \*(AK variables, the initial value of
.B IGNORECASE
is zero, so all regular expression and string
operations are normally case-sensitive.
.TP
.B LINT
Provides dynamic control of the
.B \-\^\-lint
option from within an \*(AK program.
When true,
.I gawk
prints lint warnings. When false, it does not.
When assigned the string value \fB"fatal"\fP,
lint warnings become fatal errors, exactly like
.BR \-\^\-lint=fatal .
Any other true value just prints warnings.
.TP
.B NF
The number of fields in the current input record.
.TP
.B NR
The total number of input records seen so far.
.TP
.B OFMT
The output format for numbers, \fB"%.6g"\fR, by default.
.TP
.B OFS
The output field separator, a space by default.
.TP
.B ORS
The output record separator, by default a newline.
.TP
.B PREC
The working precision of arbitrary precision floating-point
numbers, 53 by default.
.TP
.B PROCINFO
The elements of this array provide access to information about the
running \*(AK program.
On some systems,
there may be elements in the array, \fB"group1"\fP through
\fB"group\fIn\fB"\fR for some
.IR n ,
which is the number of supplementary groups that the process has.
Use the
.B in
operator to test for these elements.
The following elements are guaranteed to be available:
.RS
.TP \w'\fBPROCINFO["version"]\fR'u+1n
\fBPROCINFO["egid"]\fP
The value of the
.IR getegid (2)
system call.
.TP
\fBPROCINFO["strftime"]\fP
The default time format string for
.BR strftime() .
.TP
\fBPROCINFO["euid"]\fP
The value of the
.IR geteuid (2)
system call.
.TP
\fBPROCINFO["FS"]\fP
\fB"FS"\fP if field splitting with
.B FS
is in effect,
\fB"FPAT"\fP if field splitting with
.B FPAT
is in effect,
or \fB"FIELDWIDTHS"\fP if field splitting with
.B FIELDWIDTHS
is in effect.
.TP
\fBPROCINFO["identifiers"]\fP
A subarray, indexed by the names of all identifiers used in the
text of the AWK program.
The values indicate what
.I gawk
knows about the identifiers after it has finished parsing the program; they are
.I not
updated while the program runs.
For each identifier, the value of the element is one of the following:
.RS
.TP
\fB"array"\fR
The identifier is an array.
.TP
\fB"extension"\fR
The identifier is an extension function loaded via
.BR @load .
.TP
\fB"scalar"\fR
The identifier is a scalar.
.TP
\fB"untyped"\fR
The identifier is untyped (could be used as a scalar or array,
.I gawk
doesn't know yet).
.TP
\fB"user"\fR
The identifier is a user-defined function.
.RE
.TP
\fBPROCINFO["gid"]\fP
The value of the
.IR getgid (2)
system call.
.TP
\fBPROCINFO["pgrpid"]\fP
The process group ID of the current process.
.TP
\fBPROCINFO["pid"]\fP
The process ID of the current process.
.TP
\fBPROCINFO["ppid"]\fP
The parent process ID of the current process.
.TP
\fBPROCINFO["uid"]\fP
The value of the
.IR getuid (2)
system call.
.TP
\fBPROCINFO["sorted_in"]\fP
If this element exists in
.BR PROCINFO ,
then its value controls the order in which array elements
are traversed in
.B for
loops.
Supported values are
\fB"@ind_str_asc"\fR,
\fB"@ind_num_asc"\fR,
\fB"@val_type_asc"\fR,
\fB"@val_str_asc"\fR,
\fB"@val_num_asc"\fR,
\fB"@ind_str_desc"\fR,
\fB"@ind_num_desc"\fR,
\fB"@val_type_desc"\fR,
\fB"@val_str_desc"\fR,
\fB"@val_num_desc"\fR,
and
\fB"@unsorted"\fR.
The value can also be the name of any comparison function defined
as follows:
.sp
.in +5m
\fBfunction cmp_func(i1, v1, i2, v2)\fR
.in -5m
.sp
where
.I i1
and
.I i2
are the indices, and
.I v1
and
.I v2
are the
corresponding values of the two elements being compared.
It should return a number less than, equal to, or greater than 0,
depending on how the elements of the array are to be ordered.
.TP
\fBPROCINFO["input", "READ_TIMEOUT"]\fP
The timeout in milliseconds for reading data from
.IR input ,
where
.I input
is a redirection string or a filename. A value of zero or
less than zero means no timeout.
.TP
\fBPROCINFO["mpfr_version"]\fP
The version of the GNU MPFR library used for arbitrary precision
number support in
.IR gawk .
This entry is not present if MPFR support is not compiled into
.IR gawk .
.TP
\fBPROCINFO["gmp_version"]\fP
The version of the GNU MP library used for arbitrary precision
number support in
.IR gawk .
This entry is not present if MPFR support is not compiled into
.IR gawk .
.TP
\fBPROCINFO["prec_max"]\fP
The maximum precision supported by the GNU MPFR library for
arbitrary precision floating-point numbers.
This entry is not present if MPFR support is not compiled into
.IR gawk .
.TP
\fBPROCINFO["prec_min"]\fP
The minimum precision allowed by the GNU MPFR library for
arbitrary precision floating-point numbers.
This entry is not present if MPFR support is not compiled into
.IR gawk .
.TP
\fBPROCINFO["api_major"]\fP
The major version of the extension API.
This entry is not present if loading dynamic extensions is not available.
.TP
\fBPROCINFO["api_minor"]\fP
The minor version of the extension API.
This entry is not present if loading dynamic extensions is not available.
.TP
\fBPROCINFO["version"]\fP
the version of
.IR gawk .
.RE
.TP
.B ROUNDMODE
The rounding mode to use for arbitrary precision arithmetic on
numbers, by default \fB"N"\fR (IEEE-754 roundTiesToEven mode).
The accepted values are
\fB"N"\fR or \fB"n"\fR for roundTiesToEven,
\fB"U"\fR or \fB"u"\fR for roundTowardPositive,
\fB"D"\fR or \fB"d"\fR for roundTowardNegative,
\fB"Z"\fR or \fB"z"\fR for roundTowardZero,
and if your version of GNU MPFR library supports it,
\fB"A"\fR or \fB"a"\fR for roundTiesToAway.
.TP
.B RS
The input record separator, by default a newline.
.TP
.B RT
The record terminator.
.I Gawk
sets
.B RT
to the input text that matched the character or regular expression
specified by
.BR RS .
.TP
.B RSTART
The index of the first character matched by
.BR match() ;
0 if no match.
(This implies that character indices start at one.)
.TP
.B RLENGTH
The length of the string matched by
.BR match() ;
\-1 if no match.
.TP
.B SUBSEP
The character used to separate multiple subscripts in array
elements, by default \fB"\e034"\fR.
.TP
.B SYMTAB
An array whose indices are the names of all currently defined
global variables and arrays in the program.  The array may be used
for indirect access to read or write the value of a variable:
.sp
.ft B
.nf
.in +5m
foo = 5
SYMTAB["foo"] = 4
print foo    # prints 4
.fi
.ft R
.in -5m
.sp
The
.B isarray()
function may be used to test if an element in
.B SYMTAB
is an array.
You may not use the
.B delete
statement with the
.B SYMTAB
array.
.TP
.B TEXTDOMAIN
The text domain of the \*(AK program; used to find the localized
translations for the program's strings.
.SS Arrays
.PP
Arrays are subscripted with an expression between square brackets
.RB ( [ " and " ] ).
If the expression is an expression list
.RI ( expr ", " expr " .\|.\|.)"
then the array subscript is a string consisting of the
concatenation of the (string) value of each expression,
separated by the value of the
.B SUBSEP
variable.
This facility is used to simulate multiply dimensioned
arrays.  For example:
.PP
.RS
.ft B
i = "A";\^ j = "B";\^ k = "C"
.br
x[i, j, k] = "hello, world\en"
.ft R
.RE
.PP
assigns the string \fB"hello, world\en"\fR to the element of the array
.B x
which is indexed by the string \fB"A\e034B\e034C"\fR.  All arrays in \*(AK
are associative, i.e., indexed by string values.
.PP
The special operator
.B in
may be used to test if an array has an index consisting of a particular
value:
.PP
.RS
.ft B
.nf
if (val in array)
	print array[val]
.fi
.ft
.RE
.PP
If the array has multiple subscripts, use
.BR "(i, j) in array" .
.PP
The
.B in
construct may also be used in a
.B for
loop to iterate over all the elements of an array.
However, the
.B "(i, j) in array"
construct only works in tests, not in
.B for
loops.
.PP
An element may be deleted from an array using the
.B delete
statement.
The
.B delete
statement may also be used to delete the entire contents of an array,
just by specifying the array name without a subscript.
.PP
.I gawk
supports true multidimensional arrays. It does not require that
such arrays be ``rectangular'' as in C or C++.
For example:
.sp
.RS
.ft B
.nf
a[1] = 5
a[2][1] = 6
a[2][2] = 7
.fi
.ft
.RE
.PP
.BR NOTE :
You may need to tell
.I gawk
that an array element is really a subarray in order to use it where
.I gawk
expects an array (such as in the second argument to
.BR split() ).
You can do this by creating an element in the subarray and then
deleting it with the
.B delete
statement.
.SS Variable Typing And Conversion
.PP
Variables and fields
may be (floating point) numbers, or strings, or both.  How the
value of a variable is interpreted depends upon its context.  If used in
a numeric expression, it will be treated as a number; if used as a string
it will be treated as a string.
.PP
To force a variable to be treated as a number, add 0 to it; to force it
to be treated as a string, concatenate it with the null string.
.PP
Uninitialized variables have the numeric value 0 and the string value ""
(the null, or empty, string).
.PP
When a string must be converted to a number, the conversion is accomplished
using
.IR strtod (3).
A number is converted to a string by using the value of
.B CONVFMT
as a format string for
.IR sprintf (3),
with the numeric value of the variable as the argument.
However, even though all numbers in \*(AK are floating-point,
integral values are
.I always
converted as integers.  Thus, given
.PP
.RS
.ft B
.nf
CONVFMT = "%2.2f"
a = 12
b = a ""
.fi
.ft R
.RE
.PP
the variable
.B b
has a string value of \fB"12"\fR and not \fB"12.00"\fR.
.PP
.BR NOTE :
When operating in POSIX mode (such as with the
.B \-\^\-posix
option),
beware that locale settings may interfere with the way
decimal numbers are treated: the decimal separator of the numbers you
are feeding to
.I gawk
must conform to what your locale would expect, be it
a comma (,) or a period (.).
.PP
.I Gawk
performs comparisons as follows:
If two variables are numeric, they are compared numerically.
If one value is numeric and the other has a string value that is a
\*(lqnumeric string,\*(rq then comparisons are also done numerically.
Otherwise, the numeric value is converted to a string and a string
comparison is performed.
Two strings are compared, of course, as strings.
.PP
Note that string constants, such as \fB"57"\fP, are
.I not
numeric strings, they are string constants.
The idea of \*(lqnumeric string\*(rq
only applies to fields,
.B getline
input,
.BR FILENAME ,
.B ARGV
elements,
.B ENVIRON
elements and the elements of an array created by
.B split()
or
.B patsplit()
that are numeric strings.
The basic idea is that
.IR "user input" ,
and only user input, that looks numeric,
should be treated that way.
.SS Octal and Hexadecimal Constants
You may use C-style octal and hexadecimal constants in your AWK
program source code.
For example, the octal value
.B 011
is equal to decimal
.BR 9 ,
and the hexadecimal value
.B 0x11
is equal to decimal 17.
.SS String Constants
.PP
String constants in \*(AK are sequences of characters enclosed
between double quotes (like \fB"value"\fR).  Within strings, certain
.I "escape sequences"
are recognized, as in C.  These are:
.PP
.TP "\w'\fB\e\^\fIddd\fR'u+1n"
.B \e\e
A literal backslash.
.TP
.B \ea
The \*(lqalert\*(rq character; usually the \s-1ASCII\s+1 \s-1BEL\s+1 character.
.TP
.B \eb
Backspace.
.TP
.B \ef
Form-feed.
.TP
.B \en
Newline.
.TP
.B \er
Carriage return.
.TP
.B \et
Horizontal tab.
.TP
.B \ev
Vertical tab.
.TP
.BI \ex "\^hex digits"
The character represented by the string of hexadecimal digits following
the
.BR \ex .
As in ISO C, all following hexadecimal digits are considered part of
the escape sequence.
(This feature should tell us something about language design by committee.)
E.g., \fB"\ex1B"\fR is the \s-1ASCII\s+1 \s-1ESC\s+1 (escape) character.
.TP
.BI \e ddd
The character represented by the 1-, 2-, or 3-digit sequence of octal
digits.
E.g., \fB"\e033"\fR is the \s-1ASCII\s+1 \s-1ESC\s+1 (escape) character.
.TP
.BI \e c
The literal character
.IR c\^ .
.PP
The escape sequences may also be used inside constant regular expressions
(e.g.,
.B "/[\ \et\ef\en\er\ev]/"
matches whitespace characters).
.PP
In compatibility mode, the characters represented by octal and
hexadecimal escape sequences are treated literally when used in
regular expression constants.  Thus,
.B /a\e52b/
is equivalent to
.BR /a\e*b/ .
.SH PATTERNS AND ACTIONS
\*(AK is a line-oriented language.  The pattern comes first, and then the
action.  Action statements are enclosed in
.B {
and
.BR } .
Either the pattern may be missing, or the action may be missing, but,
of course, not both.  If the pattern is missing, the action is
executed for every single record of input.
A missing action is equivalent to
.RS
.PP
.B "{ print }"
.RE
.PP
which prints the entire record.
.PP
Comments begin with the
.B #
character, and continue until the
end of the line.
Blank lines may be used to separate statements.
Normally, a statement ends with a newline, however, this is not the
case for lines ending in
a comma,
.BR { ,
.BR ? ,
.BR : ,
.BR && ,
or
.BR || .
Lines ending in
.B do
or
.B else
also have their statements automatically continued on the following line.
In other cases, a line can be continued by ending it with a \*(lq\e\*(rq,
in which case the newline is ignored.
.PP
Multiple statements may
be put on one line by separating them with a \*(lq;\*(rq.
This applies to both the statements within the action part of a
pattern-action pair (the usual case),
and to the pattern-action statements themselves.
.SS Patterns
\*(AK patterns may be one of the following:
.PP
.RS
.nf
.B BEGIN
.B END
.B BEGINFILE
.B ENDFILE
.BI / "regular expression" /
.I "relational expression"
.IB pattern " && " pattern
.IB pattern " || " pattern
.IB pattern " ? " pattern " : " pattern
.BI ( pattern )
.BI ! " pattern"
.IB pattern1 ", " pattern2
.fi
.RE
.PP
.B BEGIN
and
.B END
are two special kinds of patterns which are not tested against
the input.
The action parts of all
.B BEGIN
patterns are merged as if all the statements had
been written in a single
.B BEGIN
rule.  They are executed before any
of the input is read.  Similarly, all the
.B END
rules are merged,
and executed when all the input is exhausted (or when an
.B exit
statement is executed).
.B BEGIN
and
.B END
patterns cannot be combined with other patterns in pattern expressions.
.B BEGIN
and
.B END
patterns cannot have missing action parts.
.PP
.B BEGINFILE
and
.B ENDFILE
are additional special patterns whose bodies are executed
before reading the first record of each command line input file
and after reading the last record of each file.
Inside the
.B BEGINFILE
rule, the value of
.B ERRNO
will be the empty string if the file was opened successfully.
Otherwise, there is some problem with the file and the code should
use
.B nextfile
to skip it. If that is not done,
.I gawk
produces its usual fatal error for files that cannot be opened.
.PP
For
.BI / "regular expression" /
patterns, the associated statement is executed for each input record that matches
the regular expression.
Regular expressions are the same as those in
.IR egrep (1),
and are summarized below.
.PP
A
.I "relational expression"
may use any of the operators defined below in the section on actions.
These generally test whether certain fields match certain regular expressions.
.PP
The
.BR && ,
.BR || ,
and
.B !
operators are logical AND, logical OR, and logical NOT, respectively, as in C.
They do short-circuit evaluation, also as in C, and are used for combining
more primitive pattern expressions.  As in most languages, parentheses
may be used to change the order of evaluation.
.PP
The
.B ?\^:
operator is like the same operator in C.  If the first pattern is true
then the pattern used for testing is the second pattern, otherwise it is
the third.  Only one of the second and third patterns is evaluated.
.PP
The
.IB pattern1 ", " pattern2
form of an expression is called a
.IR "range pattern" .
It matches all input records starting with a record that matches
.IR pattern1 ,
and continuing until a record that matches
.IR pattern2 ,
inclusive.  It does not combine with any other sort of pattern expression.
.SS Regular Expressions
Regular expressions are the extended kind found in
.IR egrep .
They are composed of characters as follows:
.TP "\w'\fB[^\fIabc.\|.\|.\fB]\fR'u+2n"
.I c
Matches the non-metacharacter
.IR c .
.TP
.I \ec
Matches the literal character
.IR c .
.TP
.B .
Matches any character
.I including
newline.
.TP
.B ^
Matches the beginning of a string.
.TP
.B $
Matches the end of a string.
.TP
.BI [ abc.\|.\|. ]
A character list: matches any of the characters
.IR abc.\|.\|. .
You may include a range of characters by separating them with a dash.
.TP
\fB[^\fIabc.\|.\|.\fB]\fR
A negated character list: matches any character except
.IR abc.\|.\|. .
.TP
.IB r1 | r2
Alternation: matches either
.I r1
or
.IR r2 .
.TP
.I r1r2
Concatenation: matches
.IR r1 ,
and then
.IR r2 .
.TP
.IB r\^ +
Matches one or more
.IR r\^ "'s."
.TP
.IB r *
Matches zero or more
.IR r\^ "'s."
.TP
.IB r\^ ?
Matches zero or one
.IR r\^ "'s."
.TP
.BI ( r )
Grouping: matches
.IR r .
.TP
.PD 0
.IB r { n }
.TP
.PD 0
.IB r { n ,}
.TP
.PD
.IB r { n , m }
One or two numbers inside braces denote an
.IR "interval expression" .
If there is one number in the braces, the preceding regular expression
.I r
is repeated
.I n
times.  If there are two numbers separated by a comma,
.I r
is repeated
.I n
to
.I m
times.
If there is one number followed by a comma, then
.I r
is repeated at least
.I n
times.
.TP
.B \ey
Matches the empty string at either the beginning or the
end of a word.
.TP
.B \eB
Matches the empty string within a word.
.TP
.B \e<
Matches the empty string at the beginning of a word.
.TP
.B \e>
Matches the empty string at the end of a word.
.TP
.B \es
Matches any whitespace character.
.TP
.B \eS
Matches any nonwhitespace character.
.TP
.B \ew
Matches any word-constituent character (letter, digit, or underscore).
.TP
.B \eW
Matches any character that is not word-constituent.
.TP
.B \e`
Matches the empty string at the beginning of a buffer (string).
.TP
.B \e'
Matches the empty string at the end of a buffer.
.PP
The escape sequences that are valid in string constants (see
.BR "String Constants" )
are also valid in regular expressions.
.PP
.I "Character classes"
are a feature introduced in the \*(PX standard.
A character class is a special notation for describing
lists of characters that have a specific attribute, but where the
actual characters themselves 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 in the USA and in France.
.PP
A character class is only valid in a regular expression
.I inside
the brackets of a character list.  Character classes consist of
.BR [: ,
a keyword denoting the class, and
.BR :] .
The character
classes defined by the \*(PX standard are:
.TP "\w'\fB[:alnum:]\fR'u+2n"
.B [:alnum:]
Alphanumeric characters.
.TP
.B [:alpha:]
Alphabetic characters.
.TP
.B [:blank:]
Space or tab characters.
.TP
.B [:cntrl:]
Control characters.
.TP
.B [:digit:]
Numeric characters.
.TP
.B [:graph:]
Characters that are both printable and visible.
(A space is printable, but not visible, while an
.B a
is both.)
.TP
.B [:lower:]
Lowercase alphabetic characters.
.TP
.B [:print:]
Printable characters (characters that are not control characters.)
.TP
.B [:punct:]
Punctuation characters (characters that are not letter, digits,
control characters, or space characters).
.TP
.B [:space:]
Space characters (such as space, tab, and formfeed, to name a few).
.TP
.B [:upper:]
Uppercase alphabetic characters.
.TP
.B [:xdigit:]
Characters that are hexadecimal digits.
.PP
For example, before the \*(PX standard, to match alphanumeric
characters, you would have had to write
.BR /[A\-Za\-z0\-9]/ .
If your character set had other alphabetic characters in it, this would not
match them, and if your character set collated differently from
\s-1ASCII\s+1, this might not even match the
\s-1ASCII\s+1 alphanumeric characters.
With the \*(PX character classes, you can write
.BR /[[:alnum:]]/ ,
and this matches
the alphabetic and numeric characters in your character set,
no matter what it is.
.PP
Two additional special sequences can appear in character lists.
These apply to non-\s-1ASCII\s+1 character sets, which can have single symbols
(called
.IR "collating elements" )
that are represented with more than one
character, as well as several characters that are equivalent for
.IR collating ,
or sorting, purposes.  (E.g., in French, a plain \*(lqe\*(rq
and a grave-accented \*(lqe\h'-\w:e:u'\`\*(rq are equivalent.)
.TP
Collating Symbols
A collating symbol is a multi-character collating element enclosed in
.B [.
and
.BR .] .
For example, if
.B ch
is a collating element, then
.B [[.ch.]]
is a regular expression that matches this collating element, while
.B [ch]
is a regular expression that matches either
.B c
or
.BR h .
.TP
Equivalence Classes
An equivalence class is a locale-specific name for a list of
characters that are equivalent.  The name is enclosed in
.B [=
and
.BR =] .
For example, the name
.B e
might be used to represent all of
\*(lqe,\*(rq \*(lqe\h'-\w:e:u'\',\*(rq and \*(lqe\h'-\w:e:u'\`.\*(rq
In this case,
.B [[=e=]]
is a regular expression
that matches any of
.BR e ,
.BR "e\h'-\w:e:u'\'" ,
or
.BR "e\h'-\w:e:u'\`" .
.PP
These features are very valuable in non-English speaking locales.
The library functions that
.I gawk
uses for regular expression matching
currently only recognize \*(PX character classes; they do not recognize
collating symbols or equivalence classes.
.PP
The
.BR \ey ,
.BR \eB ,
.BR \e< ,
.BR \e> ,
.BR \es ,
.BR \eS ,
.BR \ew ,
.BR \eW ,
.BR \e` ,
and
.B \e'
operators are specific to
.IR gawk ;
they are extensions based on facilities in the \*(GN regular expression libraries.
.PP
The various command line options
control how
.I gawk
interprets characters in regular expressions.
.TP
No options
In the default case,
.I gawk
provides all the facilities of
\*(PX regular expressions and the \*(GN regular expression operators described above.
.TP
.B \-\^\-posix
Only \*(PX regular expressions are supported, the \*(GN operators are not special.
(E.g.,
.B \ew
matches a literal
.BR w ).
.TP
.B \-\^\-traditional
Traditional \*(UX
.I awk
regular expressions are matched.  The \*(GN operators
are not special, and interval expressions are not available.
Characters described by octal and hexadecimal escape sequences are
treated literally, even if they represent regular expression metacharacters.
.TP
.B \-\^\-re\-interval
Allow interval expressions in regular expressions, even if
.B \-\^\-traditional
has been provided.
.SS Actions
Action statements are enclosed in braces,
.B {
and
.BR } .
Action statements consist of the usual assignment, conditional, and looping
statements found in most languages.  The operators, control statements,
and input/output statements
available are patterned after those in C.
.SS Operators
.PP
The operators in \*(AK, in order of decreasing precedence, are:
.PP
.TP "\w'\fB*= /= %= ^=\fR'u+1n"
.BR ( \&.\|.\|. )
Grouping
.TP
.B $
Field reference.
.TP
.B "++ \-\^\-"
Increment and decrement, both prefix and postfix.
.TP
.B ^
Exponentiation (\fB**\fR may also be used, and \fB**=\fR for
the assignment operator).
.TP
.B "+ \- !"
Unary plus, unary minus, and logical negation.
.TP
.B "* / %"
Multiplication, division, and modulus.
.TP
.B "+ \-"
Addition and subtraction.
.TP
.I space
String concatenation.
.TP
.B "|   |&"
Piped I/O for
.BR getline ,
.BR print ,
and
.BR printf .
.TP
.B "< > <= >= != =="
The regular relational operators.
.TP
.B "~ !~"
Regular expression match, negated match.
.BR NOTE :
Do not use a constant regular expression
.RB ( /foo/ )
on the left-hand side of a
.B ~
or
.BR !~ .
Only use one on the right-hand side.  The expression
.BI "/foo/ ~ " exp
has the same meaning as \fB(($0 ~ /foo/) ~ \fIexp\fB)\fR.
This is usually
.I not
what you want.
.TP
.B in
Array membership.
.TP
.B &&
Logical AND.
.TP
.B ||
Logical OR.
.TP
.B ?:
The C conditional expression.  This has the form
.IB expr1 " ? " expr2 " : " expr3\c
\&.
If
.I expr1
is true, the value of the expression is
.IR expr2 ,
otherwise it is
.IR expr3 .
Only one of
.I expr2
and
.I expr3
is evaluated.
.TP
.B "= += \-= *= /= %= ^="
Assignment.  Both absolute assignment
.BI ( var " = " value )
and operator-assignment (the other forms) are supported.
.SS Control Statements
.PP
The control statements are
as follows:
.PP
.RS
.nf
\fBif (\fIcondition\fB) \fIstatement\fR [ \fBelse\fI statement \fR]
\fBwhile (\fIcondition\fB) \fIstatement \fR
\fBdo \fIstatement \fBwhile (\fIcondition\fB)\fR
\fBfor (\fIexpr1\fB; \fIexpr2\fB; \fIexpr3\fB) \fIstatement\fR
\fBfor (\fIvar \fBin\fI array\fB) \fIstatement\fR
\fBbreak\fR
\fBcontinue\fR
\fBdelete \fIarray\^\fB[\^\fIindex\^\fB]\fR
\fBdelete \fIarray\^\fR
\fBexit\fR [ \fIexpression\fR ]
\fB{ \fIstatements \fB}\fR
\fBswitch (\fIexpression\fB) {
\fBcase \fIvalue\fB|\fIregex\fB : \fIstatement
\&.\^.\^.
\fR[ \fBdefault: \fIstatement \fR]
\fB}\fR
.fi
.RE
.SS "I/O Statements"
.PP
The input/output statements are as follows:
.PP
.TP "\w'\fBprintf \fIfmt, expr-list\fR'u+1n"
\fBclose(\fIfile \fR[\fB, \fIhow\fR]\fB)\fR
Close file, pipe or co-process.
The optional
.I how
should only be used when closing one end of a
two-way pipe to a co-process.
It must be a string value, either
\fB"to"\fR or \fB"from"\fR.
.TP
.B getline
Set
.B $0
from next input record; set
.BR NF ,
.BR NR ,
.BR FNR ,
.BR RT .
.TP
.BI "getline <" file
Set
.B $0
from next record of
.IR file ;
set
.BR NF ,
.BR RT .
.TP
.BI getline " var"
Set
.I var
from next input record; set
.BR NR ,
.BR FNR ,
.BR RT .
.TP
.BI getline " var" " <" file
Set
.I var
from next record of
.IR file ,
.BR RT .
.TP
\fIcommand\fB | getline \fR[\fIvar\fR]
Run
.I command
piping the output either into
.B $0
or
.IR var ,
as above, and
.BR RT .
.TP
\fIcommand\fB |& getline \fR[\fIvar\fR]
Run
.I command
as a co-process
piping the output either into
.B $0
or
.IR var ,
as above, and
.BR RT .
Co-processes are a
.I gawk
extension.
.RI ( command
can also be a socket.  See the subsection
.BR "Special File Names" ,
below.)
.TP
.B next
Stop processing the current input record.  The next input record
is read and processing starts over with the first pattern in the
\*(AK program.
Upon reaching the end of the input data,
.I gawk
executes any
.B END
rule(s).
.TP
.B "nextfile"
Stop processing the current input file.  The next input record read
comes from the next input file.
.B FILENAME
and
.B ARGIND
are updated,
.B FNR
is reset to 1, and processing starts over with the first pattern in the
\*(AK program.
Upon reaching the end of the input data,
.I gawk
executes any
.B END
rule(s).
.TP
.B print
Print the current record.
The output record is terminated with the value of
.BR ORS .
.TP
.BI print " expr-list"
Print expressions.
Each expression is separated by the value of
.BR OFS .
The output record is terminated with the value of
.BR ORS .
.TP
.BI print " expr-list" " >" file
Print expressions on
.IR file .
Each expression is separated by the value of
.BR OFS .
The output record is terminated with the value of
.BR ORS .
.TP
.BI printf " fmt, expr-list"
Format and print.
See \fBThe \fIprintf \fBStatement\fR, below.
.TP
.BI printf " fmt, expr-list" " >" file
Format and print on
.IR file .
.TP
.BI system( cmd-line )
Execute the command
.IR cmd-line ,
and return the exit status.
(This may not be available on non-\*(PX systems.)
.TP
\&\fBfflush(\fR[\fIfile\^\fR]\fB)\fR
Flush any buffers associated with the open output file or pipe
.IR file .
If
.I file
is missing or if it
is the null string,
then flush all open output files and pipes.
.PP
Additional output redirections are allowed for
.B print
and
.BR printf .
.TP
.BI "print .\|.\|. >>" " file"
Appends output to the
.IR file .
.TP
.BI "print .\|.\|. |" " command"
Writes on a pipe.
.TP
.BI "print .\|.\|. |&" " command"
Sends data to a co-process or socket.
(See also the subsection
.BR "Special File Names" ,
below.)
.PP
The
.B getline
command returns 1 on success, 0 on end of file, and \-1 on an error.
Upon an error,
.B ERRNO
is set to a string describing the problem.
.PP
.BR NOTE :
Failure in opening a two-way socket results in a non-fatal error being
returned to the calling function. If using a pipe, co-process, or socket to
.BR getline ,
or from
.B print
or
.B printf
within a loop, you
.I must
use
.B close()
to create new instances of the command or socket.
\*(AK does not automatically close pipes, sockets, or co-processes when
they return EOF.
.SS The \fIprintf\fP\^ Statement
.PP
The \*(AK versions of the
.B printf
statement and
.B sprintf()
function
(see below)
accept the following conversion specification formats:
.TP "\w'\fB%g\fR, \fB%G\fR'u+2n"
.B %c
A single character.
If the argument used for
.B %c
is numeric, it is treated as a character and printed.
Otherwise, the argument is assumed to be a string, and the only first
character of that string is printed.
.TP
.BR "%d" "," " %i"
A decimal number (the integer part).
.TP
.BR %e , " %E"
A floating point number of the form
[\fB\-\fP]\fId\fB.\fIdddddd\^\fBe\fR[\fB+\-\fR]\fIdd\fR.
The
.B %E
format uses
.B E
instead of
.BR e .
.TP
.BR %f , " %F"
A floating point number of the form
[\fB\-\fP]\fIddd\fB.\fIdddddd\fR.
If the system library supports it,
.B %F
is available as well. This is like
.BR %f ,
but uses capital letters for special \*(lqnot a number\*(rq
and \*(lqinfinity\*(rq values. If
.B %F
is not available,
.I gawk
uses
.BR %f .
.TP
.BR %g , " %G"
Use
.B %e
or
.B %f
conversion, whichever is shorter, with nonsignificant zeros suppressed.
The
.B %G
format uses
.B %E
instead of
.BR %e .
.TP
.B %o
An unsigned octal number (also an integer).
.TP
.PD
.B %u
An unsigned decimal number (again, an integer).
.TP
.B %s
A character string.
.TP
.BR %x , " %X"
An unsigned hexadecimal number (an integer).
The
.B %X
format uses
.B ABCDEF
instead of
.BR abcdef .
.TP
.B %%
A single
.B %
character; no argument is converted.
.PP
Optional, additional parameters may lie between the
.B %
and the control letter:
.TP
.IB count $
Use the
.IR count "'th"
argument at this point in the formatting.
This is called a
.I "positional specifier"
and
is intended primarily for use in translated versions of
format strings, not in the original text of an AWK program.
It is a
.I gawk
extension.
.TP
.B \-
The expression should be left-justified within its field.
.TP
.I space
For numeric conversions, prefix positive values with a space, and
negative values with a minus sign.
.TP
.B +
The plus sign, used before the width modifier (see below),
says to always supply a sign for numeric conversions, even if the data
to be formatted is positive.  The
.B +
overrides the space modifier.
.TP
.B #
Use an \*(lqalternate form\*(rq for certain control letters.
For
.BR %o ,
supply a leading zero.
For
.BR %x ,
and
.BR %X ,
supply a leading
.B 0x
or
.B 0X
for
a nonzero result.
For
.BR %e ,
.BR %E ,
.B %f
and
.BR %F ,
the result always contains a
decimal point.
For
.BR %g ,
and
.BR %G ,
trailing zeros are not removed from the result.
.TP
.B 0
A leading
.B 0
(zero) acts as a flag, that indicates output should be
padded with zeroes 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 be printed.
.TP
.I width
The field should be padded to this width.  The field is normally padded
with spaces.  With the
.B 0
flag, it is padded with zeroes.
.TP
.BI \&. prec
A number that specifies the precision to use when printing.
For the
.BR %e ,
.BR %E ,
.B %f
and
.BR %F ,
formats, this specifies the
number of digits you want printed to the right of the decimal point.
For the
.BR %g ,
and
.B %G
formats, it specifies the maximum number
of significant digits.  For the
.BR %d ,
.BR %i ,
.BR %o ,
.BR %u ,
.BR %x ,
and
.B %X
formats, it specifies the minimum number of
digits to print.  For
.BR %s ,
it specifies the maximum number of
characters from the string that should be printed.
.PP
The dynamic
.I width
and
.I prec
capabilities of the ISO C
.B printf()
routines are supported.
A
.B *
in place of either the
.I width
or
.I prec
specifications causes their values to be taken from
the argument list to
.B printf
or
.BR sprintf() .
To use a positional specifier with a dynamic width or precision,
supply the
.IB count $
after the
.B *
in the format string.
For example, \fB"%3$*2$.*1$s"\fP.
.SS Special File Names
.PP
When doing I/O redirection from either
.B print
or
.B printf
into a file,
or via
.B getline
from a file,
.I gawk
recognizes certain special filenames internally.  These filenames
allow access to open file descriptors inherited from
.IR gawk\^ "'s"
parent process (usually the shell).
These file names may also be used on the command line to name data files.
The filenames are:
.TP "\w'\fB/dev/stdout\fR'u+1n"
.B \-
The standard input.
.TP
.B /dev/stdin
The standard input.
.TP
.B /dev/stdout
The standard output.
.TP
.B /dev/stderr
The standard error output.
.TP
.BI /dev/fd/\^ n
The file associated with the open file descriptor
.IR n .
.PP
These are particularly useful for error messages.  For example:
.PP
.RS
.ft B
print "You blew it!" > "/dev/stderr"
.ft R
.RE
.PP
whereas you would otherwise have to use
.PP
.RS
.ft B
print "You blew it!" | "cat 1>&2"
.ft R
.RE
.PP
The following special filenames may be used with the
.B |&
co-process operator for creating TCP/IP network connections:
.TP
.PD 0
.BI /inet/tcp/ lport / rhost / rport
.TP
.PD 0
.BI /inet4/tcp/ lport / rhost / rport
.TP
.PD
.BI /inet6/tcp/ lport / rhost / rport
Files for a TCP/IP connection on local port
.I lport
to
remote host
.I rhost
on remote port
.IR rport .
Use a port of
.B 0
to have the system pick a port.
Use
.B /inet4
to force an IPv4 connection,
and
.B /inet6
to force an IPv6 connection.
Plain
.B /inet
uses the system default (most likely IPv4).
.TP
.PD 0
.BI /inet/udp/ lport / rhost / rport
.TP
.PD 0
.BI /inet4/udp/ lport / rhost / rport
.TP
.PD
.BI /inet6/udp/ lport / rhost / rport
Similar, but use UDP/IP instead of TCP/IP.
.SS Numeric Functions
.PP
\*(AK has the following built-in arithmetic functions:
.PP
.TP "\w'\fBsrand(\fR[\fIexpr\^\fR]\fB)\fR'u+1n"
.BI atan2( y , " x" )
Return the arctangent of
.I y/x
in radians.
.TP
.BI cos( expr )
Return the cosine of
.IR expr ,
which is in radians.
.TP
.BI exp( expr )
The exponential function.
.TP
.BI int( expr )
Truncate to integer.
.TP
.BI log( expr )
The natural logarithm function.
.TP
.B rand()
Return a random number
.IR N ,
between 0 and 1,
such that 0 \(<= \fIN\fP < 1.
.TP
.BI sin( expr )
Return the sine of
.IR expr ,
which is in radians.
.TP
.BI sqrt( expr )
Return the square root of
.IR expr .
.TP
\&\fBsrand(\fR[\fIexpr\^\fR]\fB)\fR
Use
.I expr
as the new seed for the random number generator.  If no
.I expr
is provided, use the time of day.
Return the previous seed for the random
number generator.
.SS String Functions
.PP
.I Gawk
has the following built-in string functions:
.PP
.TP "\w'\fBsprintf(\^\fIfmt\fB\^, \fIexpr-list\^\fB)\fR'u+1n"
\fBasort(\fIs \fR[\fB, \fId\fR [\fB, \fIhow\fR] ]\fB)\fR
Return the number of elements in the source
array
.IR s .
Sort
the contents of
.I s
using
.IR gawk\^ "'s"
normal rules for
comparing values, and replace the indices of the
sorted values
.I s
with sequential
integers starting with 1. If the optional
destination array
.I d
is specified,
first duplicate
.I s
into
.IR d ,
and then sort
.IR d ,
leaving the indices of the
source array
.I s
unchanged. The optional string
.I how
controls the direction and the comparison mode.
Valid values for
.I how
are
any of the strings valid for
\fBPROCINFO["sorted_in"]\fR.
It can also be the name of a user-defined
comparison function as described in
\fBPROCINFO["sorted_in"]\fR.
.TP "\w'\fBsprintf(\^\fIfmt\fB\^, \fIexpr-list\^\fB)\fR'u+1n"
\fBasorti(\fIs \fR[\fB, \fId\fR [\fB, \fIhow\fR] ]\fB)\fR
Return the number of elements in the source
array
.IR s .
The behavior is the same as that of
.BR asort() ,
except that the array
.I indices
are used for sorting, not the array values.
When done, the array is indexed numerically, and
the values are those of the original indices.
The original values are lost; thus provide
a second array if you wish to preserve the original.
The purpose of the optional string
.I how
is the same as described in
.B asort()
above.
.TP
\fBgensub(\fIr\fB, \fIs\fB, \fIh \fR[\fB, \fIt\fR]\fB)\fR
Search the target string
.I t
for matches of the regular expression
.IR r .
If
.I h
is a string beginning with
.B g
or
.BR G ,
then replace all matches of
.I r
with
.IR s .
Otherwise,
.I h
is a number indicating which match of
.I r
to replace.
If
.I t
is not supplied, use
.B $0
instead.
Within the replacement text
.IR s ,
the sequence
.BI \e n\fR,
where
.I n
is a digit from 1 to 9, may be used to indicate just the text that
matched the
.IR n 'th
parenthesized subexpression.  The sequence
.B \e0
represents the entire matched text, as does the character
.BR & .
Unlike
.B sub()
and
.BR gsub() ,
the modified string is returned as the result of the function,
and the original target string is
.I not
changed.
.TP "\w'\fBsprintf(\^\fIfmt\fB\^, \fIexpr-list\^\fB)\fR'u+1n"
\fBgsub(\fIr\fB, \fIs \fR[\fB, \fIt\fR]\fB)\fR
For each substring matching the regular expression
.I r
in the string
.IR t ,
substitute the string
.IR s ,
and return the number of substitutions.
If
.I t
is not supplied, use
.BR $0 .
An
.B &
in the replacement text is replaced with the text that was actually matched.
Use
.B \e&
to get a literal
.BR & .
(This must be typed as \fB"\e\e&"\fP;
see \*(EP
for a fuller discussion of the rules for
.BR & 's
and backslashes in the replacement text of
.BR sub() ,
.BR gsub() ,
and
.BR gensub() .)
.TP
.BI index( s , " t" )
Return the index of the string
.I t
in the string
.IR s ,
or 0 if
.I t
is not present.
(This implies that character indices start at one.)
It is a fatal error to use a regexp constant for
.IR t .
.TP
\fBlength(\fR[\fIs\fR]\fB)
Return the length of the string
.IR s ,
or the length of
.B $0
if
.I s
is not supplied.
As a non-standard extension, with an array argument,
.B length()
returns the number of elements in the array.
.TP
\fBmatch(\fIs\fB, \fIr \fR[\fB, \fIa\fR]\fB)\fR
Return the position in
.I s
where the regular expression
.I r
occurs, or 0 if
.I r
is not present, and set the values of
.B RSTART
and
.BR RLENGTH .
Note that the argument order is the same as for the
.B ~
operator:
.IB str " ~"
.IR re .
.ft R
If array
.I a
is provided,
.I a
is cleared and then elements 1 through
.I n
are filled with the portions of
.I s
that match the corresponding parenthesized
subexpression in
.IR r .
The 0'th element of
.I a
contains the portion
of
.I s
matched by the entire regular expression
.IR r .
Subscripts
\fBa[\fIn\^\fB, "start"]\fR,
and
\fBa[\fIn\^\fB, "length"]\fR
provide the starting index in the string and length
respectively, of each matching substring.
.TP
\fBpatsplit(\fIs\fB, \fIa \fR[\fB, \fIr\fR [\fB, \fIseps\fR] ]\fB)\fR
Split the string
.I s
into the array
.I a
and the separators array
.I seps
on the regular expression
.IR r ,
and return the number of fields.
Element values are the portions of
.I s
that matched
.IR r .
The value of
.BI seps[ i ]
is the separator that appeared in
front of
.BI a[ i +1]\fR.
\&\fRIf
.I r
is omitted,
.B FPAT
is used instead.
The arrays
.I a
and
.I seps
are cleared first.
Splitting behaves identically to field splitting with
.BR FPAT ,
described above.
.TP
\fBsplit(\fIs\fB, \fIa \fR[\fB, \fIr\fR [\fB, \fIseps\fR] ]\fB)\fR
Split the string
.I s
into the array
.I a
and the separators array
.I seps
on the regular expression
.IR r ,
and return the number of fields.  If
.I r
is omitted,
.B FS
is used instead.
The arrays
.I a
and
.I seps
are cleared first.
.BI seps[ i ]
is the field separator matched by
.I r
between
.BI a[ i ]
and
.BI a[ i +1]\fR.
\&\fRIf
.I r
is a single space, then leading whitespace in
.I s
goes into the extra array element
.B seps[0]
and trailing whitespace goes into the extra array element
.BI seps[ n ]\fR,
where
.I n
is the return value of
.BI split( s ", " a ", " r ", " seps )\fR.
Splitting behaves identically to field splitting, described above.
.TP
.BI sprintf( fmt , " expr-list" )
Print
.I expr-list
according to
.IR fmt ,
and return the resulting string.
.TP
.BI strtonum( str )
Examine
.IR str ,
and return its numeric value.
If
.I str
begins
with a leading
.BR 0 ,
treat it
as an octal number.
If
.I str
begins
with a leading
.B 0x
or
.BR 0X ,
treat it
as a hexadecimal number.
Otherwise, assume it is a decimal number.
.TP
\fBsub(\fIr\fB, \fIs \fR[\fB, \fIt\fR]\fB)\fR
Just like
.BR gsub() ,
but replace only the first matching substring.
.TP
\fBsubstr(\fIs\fB, \fIi \fR[\fB, \fIn\fR]\fB)\fR
Return the at most
.IR n -character
substring of
.I s
starting at
.IR i .
If
.I n
is omitted, use the rest of
.IR s .
.TP
.BI tolower( str )
Return a copy of the string
.IR str ,
with all the uppercase characters in
.I str
translated to their corresponding lowercase counterparts.
Non-alphabetic characters are left unchanged.
.TP
.BI toupper( str )
Return a copy of the string
.IR str ,
with all the lowercase characters in
.I str
translated to their corresponding uppercase counterparts.
Non-alphabetic characters are left unchanged.
.PP
.I Gawk
is multibyte aware.  This means that
.BR index() ,
.BR length() ,
.B substr()
and
.B match()
all work in terms of characters, not bytes.
.SS Time Functions
Since one of the primary uses of \*(AK programs is processing log files
that contain time stamp information,
.I gawk
provides the following functions for obtaining time stamps and
formatting them.
.PP
.TP "\w'\fBsystime()\fR'u+1n"
\fBmktime(\fIdatespec\fB)\fR
Turn
.I datespec
into a time stamp of the same form as returned by
.BR systime() ,
and return the result.
The
.I datespec
is a string of the form
.IR "YYYY MM DD HH MM SS[ DST]" .
The contents of the string are 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,
and an optional daylight saving flag.
The values of these numbers need not be within the ranges specified;
for example, an hour of \-1 means 1 hour before midnight.
The origin-zero Gregorian calendar is assumed,
with year 0 preceding year 1 and year \-1 preceding year 0.
The time is assumed to be in the local timezone.
If the daylight saving flag is positive,
the time is assumed to be daylight saving time;
if zero, the time is assumed to be standard time;
and if negative (the default),
.B mktime()
attempts to determine whether daylight saving time is in effect
for the specified time.
If
.I datespec
does not contain enough elements or if the resulting time
is out of range,
.B mktime()
returns \-1.
.TP
\fBstrftime(\fR[\fIformat \fR[\fB, \fItimestamp\fR[\fB, \fIutc-flag\fR]]]\fB)\fR
Format
.I timestamp
according to the specification in
.IR format .
If
.I utc-flag
is present and is non-zero or non-null, the result
is in UTC, otherwise the result is in local time.
The
.I timestamp
should be of the same form as returned by
.BR systime() .
If
.I timestamp
is missing, the current time of day is used.
If
.I format
is missing, a default format equivalent to the output of
.IR date (1)
is used.
The default format is available in
.BR PROCINFO["strftime"] .
See the specification for the
.B strftime()
function in ISO C for the format conversions that are
guaranteed to be available.
.TP
.B systime()
Return the current time of day as the number of seconds since the Epoch
(1970-01-01 00:00:00 UTC on \*(PX systems).
.SS Bit Manipulations Functions
.I Gawk
supplies the following bit manipulation functions.
They work by converting double-precision floating point
values to
.B uintmax_t
integers, doing the operation, and then converting the
result back to floating point.
The functions are:
.TP "\w'\fBrshift(\fIval\fB, \fIcount\fB)\fR'u+2n"
\fBand(\fIv1\fB, \fIv2 \fR[, ...]\fB)\fR
Return the bitwise AND of the values provided in the argument list.
There must be at least two.
.TP
\fBcompl(\fIval\fB)\fR
Return the bitwise complement of
.IR val .
.TP
\fBlshift(\fIval\fB, \fIcount\fB)\fR
Return the value of
.IR val ,
shifted left by
.I count
bits.
.TP
\fBor(\fIv1\fB, \fIv2 \fR[, ...]\fB)\fR
Return the bitwise OR of the values provided in the argument list.
There must be at least two.
.TP
\fBrshift(\fIval\fB, \fIcount\fB)\fR
Return the value of
.IR val ,
shifted right by
.I count
bits.
.TP
\fBxor(\fIv1\fB, \fIv2 \fR[, ...]\fB)\fR
Return the bitwise XOR of the values provided in the argument list.
There must be at least two.
.PP
.SS Type Function
The following function is for use with multidimensional arrays.
.TP
\fBisarray(\fIx\fB)\fR
Return true if
.I x
is an array, false otherwise.
.SS Internationalization Functions
The following functions may be used from within your AWK program for
translating strings at run-time.
For full details, see \*(EP.
.TP
\fBbindtextdomain(\fIdirectory \fR[\fB, \fIdomain\fR]\fB)\fR
Specify the directory where
.I gawk
looks for the
.B \&.gmo
files, in case they
will not or cannot be placed in the ``standard'' locations
(e.g., during testing).
It returns the directory where
.I domain
is ``bound.''
.sp .5
The default
.I domain
is the value of
.BR TEXTDOMAIN .
If
.I directory
is the null string (\fB""\fR), then
.B bindtextdomain()
returns the current binding for the
given
.IR domain .
.TP
\fBdcgettext(\fIstring \fR[\fB, \fIdomain \fR[\fB, \fIcategory\fR]]\fB)\fR
Return the translation of
.I string
in text domain
.I domain
for locale category
.IR category .
The default value for
.I domain
is the current value of
.BR TEXTDOMAIN .
The default value for
.I category
is \fB"LC_MESSAGES"\fR.
.sp .5
If you supply a value for
.IR category ,
it must be a string equal to
one of the known locale categories described
in \*(EP.
You must also supply a text domain.  Use
.B TEXTDOMAIN
if you want to use the current domain.
.TP
\fBdcngettext(\fIstring1\fB, \fIstring2\fB, \fInumber \fR[\fB, \fIdomain \fR[\fB, \fIcategory\fR]]\fB)\fR
Return the plural form used for
.I number
of the translation of
.I string1
and
.I string2
in
text domain
.I domain
for locale category
.IR category .
The default value for
.I domain
is the current value of
.BR TEXTDOMAIN .
The default value for
.I category
is \fB"LC_MESSAGES"\fR.
.sp .5
If you supply a value for
.IR category ,
it must be a string equal to
one of the known locale categories described
in \*(EP.
You must also supply a text domain.  Use
.B TEXTDOMAIN
if you want to use the current domain.
.SH USER-DEFINED FUNCTIONS
Functions in \*(AK are defined as follows:
.PP
.RS
\fBfunction \fIname\fB(\fIparameter list\fB) { \fIstatements \fB}\fR
.RE
.PP
Functions are executed when they are called from within expressions
in either patterns or actions.  Actual parameters supplied in the function
call are used to instantiate the formal parameters declared in the function.
Arrays are passed by reference, other variables are passed by value.
.PP
Since functions were not originally part of the \*(AK language, the provision
for local variables is rather clumsy: They are declared as extra parameters
in the parameter list.  The convention is to separate local variables from
real parameters by extra spaces in the parameter list.  For example:
.PP
.RS
.ft B
.nf
function  f(p, q,     a, b)	# a and b are local
{
	\&.\|.\|.
}

/abc/	{ .\|.\|. ; f(1, 2) ; .\|.\|. }
.fi
.ft R
.RE
.PP
The left parenthesis in a function call is required
to immediately follow the function name,
without any intervening whitespace.
This avoids a syntactic ambiguity with the concatenation operator.
This restriction does not apply to the built-in functions listed above.
.PP
Functions may call each other and may be recursive.
Function parameters used as local variables are initialized
to the null string and the number zero upon function invocation.
.PP
Use
.BI return " expr"
to return a value from a function.  The return value is undefined if no
value is provided, or if the function returns by \*(lqfalling off\*(rq the
end.
.PP
As a
.I gawk
extension, functions may be called indirectly. To do this, assign
the name of the function to be called, as a string, to a variable.
Then use the variable as if it were the name of a function, prefixed with an
.B @
sign, like so:
.RS
.ft B
.nf
function  myfunc()
{
	print "myfunc called"
	\&.\|.\|.
}

{	.\|.\|.
	the_func = "myfunc"
	@the_func()	# call through the_func to myfunc
	.\|.\|.
}
.fi
.ft R
.RE
.PP
If
.B \-\^\-lint
has been provided,
.I gawk
warns about calls to undefined functions at parse time,
instead of at run time.
Calling an undefined function at run time is a fatal error.
.PP
The word
.B func
may be used in place of
.BR function ,
although this is deprecated.
.SH DYNAMICALLY LOADING NEW FUNCTIONS
You can dynamically add new built-in functions to the running
.I gawk
interpreter with the
.B @load
statement.
The full details are beyond the scope of this manual page;
see \*(EP.
.SH SIGNALS
The
.I gawk
profiler accepts two signals.
.B SIGUSR1
causes it to dump a profile and function call stack to the
profile file, which is either
.BR awkprof.out ,
or whatever file was named with the
.B \-\^\-profile
option.  It then continues to run.
.B SIGHUP
causes
.I gawk
to dump the profile and function call stack and then exit.
.SH INTERNATIONALIZATION
.PP
String constants are sequences of characters enclosed in double
quotes.  In non-English speaking environments, it is possible to mark
strings in the \*(AK program as requiring translation to the local
natural language. Such strings are marked in the \*(AK program with
a leading underscore (\*(lq_\*(rq).  For example,
.sp
.RS
.ft B
gawk 'BEGIN { print "hello, world" }'
.RE
.sp
.ft R
always prints
.BR "hello, world" .
But,
.sp
.RS
.ft B
gawk 'BEGIN { print _"hello, world" }'
.RE
.sp
.ft R
might print
.B "bonjour, monde"
in France.
.PP
There are several steps involved in producing and running a localizable
\*(AK program.
.TP "\w'4.'u+2n"
1.
Add a
.B BEGIN
action to assign a value to the
.B TEXTDOMAIN
variable to set the text domain to a name associated with your program:
.sp
.in +5m
.ft B
BEGIN { TEXTDOMAIN = "myprog" }
.ft R
.in -5m
.sp
This allows
.I gawk
to find the
.B \&.gmo
file associated with your program.
Without this step,
.I gawk
uses the
.B messages
text domain,
which likely does not contain translations for your program.
.TP
2.
Mark all strings that should be translated with leading underscores.
.TP
3.
If necessary, use the
.B dcgettext()
and/or
.B bindtextdomain()
functions in your program, as appropriate.
.TP
4.
Run
.B "gawk \-\^\-gen\-pot \-f myprog.awk > myprog.pot"
to generate a
.B \&.pot
file for your program.
.TP
5.
Provide appropriate translations, and build and install the corresponding
.B \&.gmo
files.
.PP
The internationalization features are described in full detail in \*(EP.
.SH POSIX COMPATIBILITY
A primary goal for
.I gawk
is compatibility with the \*(PX standard, as well as with the
latest version of Brian Kernighan's
.IR awk .
To this end,
.I gawk
incorporates the following user visible
features which are not described in the \*(AK book,
but are part of the Brian Kernighan's version of
.IR awk ,
and are in the \*(PX standard.
.PP
The book indicates that command line variable assignment happens when
.I awk
would otherwise open the argument as a file, which is after the
.B BEGIN
rule is executed.  However, in earlier implementations, when such an
assignment appeared before any file names, the assignment would happen
.I before
the
.B BEGIN
rule was run.  Applications came to depend on this \*(lqfeature.\*(rq
When
.I awk
was changed to match its documentation, the
.B \-v
option for assigning variables before program execution was added to
accommodate applications that depended upon the old behavior.
(This feature was agreed upon by both the Bell Laboratories
and the \*(GN developers.)
.PP
When processing arguments,
.I gawk
uses the special option \*(lq\-\^\-\*(rq to signal the end of
arguments.
In compatibility mode, it warns about but otherwise ignores
undefined options.
In normal operation, such arguments are passed on to the \*(AK program for
it to process.
.PP
The \*(AK book does not define the return value of
.BR srand() .
The \*(PX standard
has it return the seed it was using, to allow keeping track
of random number sequences.  Therefore
.B srand()
in
.I gawk
also returns its current seed.
.PP
Other new features are:
The use of multiple
.B \-f
options (from MKS
.IR awk );
the
.B ENVIRON
array; the
.BR \ea ,
and
.B \ev
escape sequences (done originally in
.I gawk
and fed back into the Bell Laboratories version); the
.B tolower()
and
.B toupper()
built-in functions (from the Bell Laboratories version); and the ISO C conversion specifications in
.B printf
(done first in the Bell Laboratories version).
.SH HISTORICAL FEATURES
There is one feature of historical \*(AK implementations that
.I gawk
supports:
It is possible to call the
.B length()
built-in function not only with no argument, but even without parentheses!
Thus,
.RS
.PP
.ft B
a = length	# Holy Algol 60, Batman!
.ft R
.RE
.PP
is the same as either of
.RS
.PP
.ft B
a = length()
.br
a = length($0)
.ft R
.RE
.PP
Using this feature is poor practice, and
.I gawk
issues a warning about its use if
.B \-\^\-lint
is specified on the command line.
.SH GNU EXTENSIONS
.I Gawk
has a too-large number of extensions to \*(PX
.IR awk .
They are described in this section.  All the extensions described here
can be disabled by
invoking
.I gawk
with the
.B \-\^\-traditional
or
.B \-\^\-posix
options.
.PP
The following features of
.I gawk
are not available in
\*(PX
.IR awk .
.\" Environment vars and startup stuff
.TP "\w'\(bu'u+1n"
\(bu
No path search is performed for files named via the
.B \-f
option.  Therefore the
.B AWKPATH
environment variable is not special.
.\" POSIX and language recognition issues
.TP
\(bu
There is no facility for doing file inclusion
.RI ( gawk 's
.B @include
mechanism).
.TP
\(bu
There is no facility for dynamically adding new functions
written in C
.RI ( gawk 's
.B @load
mechanism).
.TP
\(bu
The
.B \ex
escape sequence.
(Disabled with
.BR \-\^\-posix .)
.TP
\(bu
The ability to continue lines after
.B ?
and
.BR : .
(Disabled with
.BR \-\^\-posix .)
.TP
\(bu
Octal and hexadecimal constants in AWK programs.
.\" Special variables
.TP
\(bu
The
.BR ARGIND ,
.BR BINMODE ,
.BR ERRNO ,
.BR LINT ,
.B  RT
and
.B  TEXTDOMAIN
variables are not special.
.TP
\(bu
The
.B IGNORECASE
variable and its side-effects are not available.
.TP
\(bu
The
.B FIELDWIDTHS
variable and fixed-width field splitting.
.TP
\(bu
The
.B FPAT
variable and field splitting based on field values.
.TP
\(bu
The
.B PROCINFO
array is not available.
.\" I/O stuff
.TP
\(bu
The use of
.B RS
as a regular expression.
.TP
\(bu
The special file names available for I/O redirection are not recognized.
.TP
\(bu
The
.B |&
operator for creating co-processes.
.TP
\(bu
The
.B BEGINFILE
and
.B ENDFILE
special patterns are not available.
.\" Changes to standard awk functions
.TP
\(bu
The ability to split out individual characters using the null string
as the value of
.BR FS ,
and as the third argument to
.BR split() .
.TP
\(bu
An optional fourth argument to
.B split()
to receive the separator texts.
.TP
\(bu
The optional second argument to the
.B close()
function.
.TP
\(bu
The optional third argument to the
.B match()
function.
.TP
\(bu
The ability to use positional specifiers with
.B printf
and
.BR sprintf() .
.TP
\(bu
The ability to pass an array to
.BR length() .
.\" New keywords or changes to keywords
.\" (As of 2012, these are in POSIX)
.\" .TP
.\" \(bu
.\" The use of
.\" .BI delete " array"
.\" to delete the entire contents of an array.
.\" .TP
.\" \(bu
.\" The use of
.\" .B "nextfile"
.\" to abandon processing of the current input file.
.\" New functions
.TP
\(bu
The
.BR and() ,
.BR asort() ,
.BR asorti() ,
.BR bindtextdomain() ,
.BR compl() ,
.BR dcgettext() ,
.BR dcngettext() ,
.BR gensub() ,
.BR lshift() ,
.BR mktime() ,
.BR or() ,
.BR patsplit() ,
.BR rshift() ,
.BR strftime() ,
.BR strtonum() ,
.B systime()
and
.B xor()
functions.
.\" I18N stuff
.TP
\(bu
Localizable strings.
.PP
The \*(AK book does not define the return value of the
.B close()
function.
.IR Gawk\^ "'s"
.B close()
returns the value from
.IR fclose (3),
or
.IR pclose (3),
when closing an output file or pipe, respectively.
It returns the process's exit status when closing an input pipe.
The return value is \-1 if the named file, pipe
or co-process was not opened with a redirection.
.PP
When
.I gawk
is invoked with the
.B \-\^\-traditional
option,
if the
.I fs
argument to the
.B \-F
option is \*(lqt\*(rq, then
.B FS
is set to the tab character.
Note that typing
.B "gawk \-F\et \&.\|.\|."
simply causes the shell to quote the \*(lqt,\*(rq and does not pass
\*(lq\et\*(rq to the
.B \-F
option.
Since this is a rather ugly special case, it is not the default behavior.
This behavior also does not occur if
.B \-\^\-posix
has been specified.
To really get a tab character as the field separator, it is best to use
single quotes:
.BR "gawk \-F'\et' \&.\|.\|." .
.ig
.PP
If
.I gawk
was compiled for debugging, it
accepts the following additional options:
.TP
.PD 0
.B \-Y
.TP
.PD
.B \-\^\-parsedebug
Turn on
.IR yacc (1)
or
.IR bison (1)
debugging output during program parsing.
This option should only be of interest to the
.I gawk
maintainers, and may not even be compiled into
.IR gawk .
..
.SH ENVIRONMENT VARIABLES
The
.B AWKPATH
environment variable can be used to provide a list of directories that
.I gawk
searches when looking for files named via the
.BR \-f ,
.RB \-\^\-file ,
.B \-i
and
.B \-\^\-include
options.  If the initial search fails, the path is searched again after
appending
.B \&.awk
to the filename.
.PP
The
.B AWKLIBPATH
environment variable can be used to provide a list of directories that
.I gawk
searches when looking for files named via the
.B \-l
and
.B \-\^\-load
options.
.PP
The
.B GAWK_READ_TIMEOUT
environment variable can be used to specify a timeout
in milliseconds for reading input from a terminal, pipe
or two-way communication including sockets.
.PP
For connection to a remote host via socket,
.B GAWK_SOCK_RETRIES
controls the number of retries, and
.B GAWK_MSEC_SLEEP
and the interval between retries.
The interval is in milliseconds. On systems that do not support
.IR usleep (3),
the value is rounded up to an integral number of seconds.
.PP
If
.B POSIXLY_CORRECT
exists in the environment, then
.I gawk
behaves exactly as if
.B \-\^\-posix
had been specified on the command line.
If
.B \-\^\-lint
has been specified,
.I gawk
issues a warning message to this effect.
.SH EXIT STATUS
If the
.B exit
statement is used with a value,
then
.I gawk
exits with
the numeric value given to it.
.PP
Otherwise, if there were no problems during execution,
.I gawk
exits with the value of the C constant
.BR EXIT_SUCCESS .
This is usually zero.
.PP
If an error occurs,
.I gawk
exits with the value of
the C constant
.BR EXIT_FAILURE .
This is usually one.
.PP
If
.I gawk
exits because of a fatal error, the exit
status is 2.  On non-POSIX systems, this value may be mapped to
.BR EXIT_FAILURE .
.SH VERSION INFORMATION
This man page documents
.IR gawk ,
version 4.1.
.SH AUTHORS
The original version of \*(UX
.I awk
was designed and implemented by Alfred Aho,
Peter Weinberger, and Brian Kernighan of Bell Laboratories.  Brian Kernighan
continues to maintain and enhance it.
.PP
Paul Rubin and Jay Fenlason,
of the Free Software Foundation, wrote
.IR gawk ,
to be compatible with the original version of
.I awk
distributed in Seventh Edition \*(UX.
John Woods contributed a number of bug fixes.
David Trueman, with contributions
from Arnold Robbins, made
.I gawk
compatible with the new version of \*(UX
.IR awk .
Arnold Robbins is the current maintainer.
.PP
See \*(EP for a full list of the contributors to
.I gawk
and its documentation.
.PP
See the
.B README
file in the
.I gawk
distribution for up-to-date information about maintainers
and which ports are currently supported.
.SH BUG REPORTS
If you find a bug in
.IR gawk ,
please send electronic mail to
.BR bug-gawk@gnu.org .
Please include your operating system and its revision, the version of
.I gawk
(from
.BR "gawk \-\^\-version" ),
which C compiler you used to compile it, and a test program
and data that are as small as possible for reproducing the problem.
.PP
Before sending a bug report, please do the following things.  First, verify that
you have the latest version of
.IR 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
.B LC_ALL
to
.B 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.
Finally, please read this man page and the reference manual carefully to
be sure that what you think is a bug really is, instead of just a quirk
in the language.
.PP
Whatever you do, do
.B NOT
post a bug report in
.BR comp.lang.awk .
While the
.I gawk
developers occasionally read this newsgroup, posting bug reports there
is an unreliable way to report bugs.  Instead, please use the electronic mail
addresses given above.
Really.
.PP
If you're using a GNU/Linux or BSD-based system,
you may wish to submit a bug report to the vendor of your distribution.
That's fine, but please send a copy to the official email address as well,
since there's no guarantee that the bug report will be forwarded to the
.I gawk
maintainer.
.SH BUGS
The
.B \-F
option is not necessary given the command line variable assignment feature;
it remains only for backwards compatibility.
.PP
Syntactically invalid single character programs tend to overflow
the parse stack, generating a rather unhelpful message.  Such programs
are surprisingly difficult to diagnose in the completely general case,
and the effort to do so really is not worth it.
.SH SEE ALSO
.IR egrep (1),
.IR sed (1),
.IR getpid (2),
.IR getppid (2),
.IR getpgrp (2),
.IR getuid (2),
.IR geteuid (2),
.IR getgid (2),
.IR getegid (2),
.IR getgroups (2),
.IR usleep (3)
.PP
.IR "The AWK Programming Language" ,
Alfred V. Aho, Brian W. Kernighan, Peter J. Weinberger,
Addison-Wesley, 1988.  ISBN 0-201-07981-X.
.PP
\*(EP,
Edition 4.1, shipped with the
.I gawk
source.
The current version of this document is available online at
.BR http://www.gnu.org/software/gawk/manual .
.SH EXAMPLES
.nf
Print and sort the login names of all users:

.ft B
	BEGIN	{ FS = ":" }
		{ print $1 | "sort" }

.ft R
Count lines in a file:

.ft B
		{ nlines++ }
	END	{ print nlines }

.ft R
Precede each line by its number in the file:

.ft B
	{ print FNR, $0 }

.ft R
Concatenate and line number (a variation on a theme):

.ft B
	{ print NR, $0 }

.ft R
Run an external command for particular lines of data:

.ft B
	tail \-f access_log |
	awk '/myhome.html/ { system("nmap " $1 ">> logdir/myhome.html") }'
.ft R
.fi
.SH ACKNOWLEDGEMENTS
Brian Kernighan
provided valuable assistance during testing and debugging.
We thank him.
.SH COPYING PERMISSIONS
Copyright \(co 1989, 1991, 1992, 1993, 1994, 1995, 1996,
1997, 1998, 1999, 2001, 2002, 2003, 2004, 2005, 2007, 2009,
2010, 2011, 2012, 2013, 2014
Free Software Foundation, Inc.
.PP
Permission is granted to make and distribute verbatim copies of
this manual page provided the copyright notice and this permission
notice are preserved on all copies.
.ig
Permission is granted to process this file through troff and print the
results, provided the printed document carries copying permission
notice identical to this one except for the removal of this paragraph
(this paragraph not being relevant to the printed manual page).
..
.PP
Permission is granted to copy and distribute modified versions of this
manual page under the conditions for verbatim copying, provided that
the entire resulting derived work is distributed under the terms of a
permission notice identical to this one.
.PP
Permission is granted to copy and distribute translations of this
manual page into another language, under the above conditions for
modified versions, except that this permission notice may be stated in
a translation approved by the Foundation.