(main): When using obsolescent +pos -pos syntax, allow

[coreutils.git] / doc / textutils.texi

\input texinfo
@c %**start of header
@setfilename textutils.info
@settitle GNU text utilities
@c %**end of header

@include version.texi

@c Define new indices.
@defcodeindex op

@c Put everything in one index (arbitrarily chosen to be the concept index).
@syncodeindex fn cp
@syncodeindex ky cp
@syncodeindex op cp
@syncodeindex pg cp
@syncodeindex vr cp

@ifinfo
@set Francois Franc,ois
@end ifinfo
@tex
@set Francois Fran\noexpand\ptexc cois
@end tex

@ifinfo
@format
START-INFO-DIR-ENTRY
* Text utilities: (textutils).          GNU text utilities.
* cat: (textutils)cat invocation.               Concatenate and write files.
* cksum: (textutils)cksum invocation.           Print POSIX CRC checksum.
* comm: (textutils)comm invocation.             Compare sorted files by line.
* csplit: (textutils)csplit invocation.         Split by context.
* cut: (textutils)cut invocation.               Print selected parts of lines.
* expand: (textutils)expand invocation.         Convert tabs to spaces.
* fmt: (textutils)fmt invocation.               Reformat paragraph text.
* fold: (textutils)fold invocation.             Wrap long input lines.
* head: (textutils)head invocation.             Output the first part of files.
* join: (textutils)join invocation.             Join lines on a common field.
* md5sum: (textutils)md5sum invocation.         Print or check message-digests.
* nl: (textutils)nl invocation.                 Number lines and write files.
* od: (textutils)od invocation.                 Dump files in octal, etc.
* paste: (textutils)paste invocation.           Merge lines of files.
* pr: (textutils)pr invocation.                 Paginate or columnate files.
* sort: (textutils)sort invocation.             Sort text files.
* split: (textutils)split invocation.           Split into fixed-size pieces.
* sum: (textutils)sum invocation.               Print traditional checksum.
* tac: (textutils)tac invocation.               Reverse files.
* tail: (textutils)tail invocation.             Output the last part of files.
* tr: (textutils)tr invocation.                 Translate characters.
* unexpand: (textutils)unexpand invocation.     Convert spaces to tabs.
* uniq: (textutils)uniq invocation.             Uniqify files.
* wc: (textutils)wc invocation.                 Byte, word, and line counts.
END-INFO-DIR-ENTRY
@end format
@end ifinfo

@ifinfo
This file documents the GNU text utilities.

Copyright (C) 1994, 95 Free Software Foundation, Inc.

Permission is granted to make and distribute verbatim copies of
this manual provided the copyright notice and this permission notice
are preserved on all copies.

@ignore
Permission is granted to process this file through TeX 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).

@end ignore
Permission is granted to copy and distribute modified versions of this
manual 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.

Permission is granted to copy and distribute translations of this manual
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.
@end ifinfo

@titlepage
@title GNU @code{textutils}
@subtitle A set of text utilities
@subtitle for version @value{VERSION}, @value{RELEASEDATE}
@author David MacKenzie et al.

@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1994, 95 Free Software Foundation, Inc.

Permission is granted to make and distribute verbatim copies of
this manual provided the copyright notice and this permission notice
are preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual 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.

Permission is granted to copy and distribute translations of this manual
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.
@end titlepage


@ifinfo
@node Top
@top GNU text utilities

@cindex text utilities
@cindex utilities for text handling

This manual minimally documents version @value{VERSION} of the GNU text
utilities.

@menu
* Introduction::                       Caveats, overview, and authors.
* Common options::                     Common options.
* Output of entire files::             cat tac nl od
* Formatting file contents::           fmt pr fold
* Output of parts of files::           head tail split csplit
* Summarizing files::                  wc sum cksum md5sum
* Operating on sorted files::          sort uniq comm
* Operating on fields within a line::  cut paste join
* Operating on characters::            tr expand unexpand
* Opening the software toolbox::       The software tools philosophy.
* Index::                              General index.
@end menu
@end ifinfo


@node Introduction
@chapter Introduction

@cindex introduction

This manual is incomplete: No attempt is made to explain basic concepts
in a way suitable for novices.  Thus, if you are interested, please get
involved in improving this manual.  The entire GNU community will
benefit.

@cindex POSIX.2
The GNU text utilities are mostly compatible with the POSIX.2 standard.

@c This paragraph appears in all of fileutils.texi, textutils.texi, and
@c sh-utils.texi too -- so be sure to keep them consistent.
@cindex bugs, reporting
Please report bugs to @samp{bug-gnu-utils@@prep.ai.mit.edu}.  Remember
to include the version number, machine architecture, input files, and
any other information needed to reproduce the bug: your input, what you
expected, what you got, and why it is wrong.  Diffs are welcome, but
please include a description of the problem as well, since this is
sometimes difficult to infer. @xref{Bugs, , , gcc, GNU CC}.

This manual is based on the Unix man pages in the distribution, which
were originally written by David MacKenzie and updated by Jim Meyering.
The original @code{fmt} man page was written by Ross Paterson.
@value{Francois} Pinard did the initial conversion to Texinfo format.
Karl Berry did the indexing, some reorganization, and editing of the results.
Richard Stallman contributed his usual invaluable insights to the
overall process.


@node Common options
@chapter Common options

@cindex common options

Certain options are available in all these programs.  Rather than
writing identical descriptions for each of the programs, they are
described here.  (In fact, every GNU program accepts (or should accept)
these options.)

A few of these programs take arbitrary strings as arguments.  In those
cases, @samp{--help} and @samp{--version} are taken as these options
only if there is one and exactly one command line argument.

@table @samp

@item --help
@opindex --help
@cindex help, online
Print a usage message listing all available options, then exit successfully.

@item --version
@opindex --version
@cindex version number, finding
Print the version number, then exit successfully.

@end table


@node Output of entire files
@chapter Output of entire files

@cindex output of entire files
@cindex entire files, output of

These commands read and write entire files, possibly transforming them
in some way.

@menu
* cat invocation::              Concatenate and write files.
* tac invocation::              Concatenate and write files in reverse.
* nl invocation::               Number lines and write files.
* od invocation::               Write files in octal or other formats.
@end menu

@node cat invocation
@section @code{cat}: Concatenate and write files

@pindex cat
@cindex concatenate and write files
@cindex copying files

@code{cat} copies each @var{file} (@samp{-} means standard input), or
standard input if none are given, to standard output.  Synopsis:

@example
cat [@var{option}] [@var{file}]@dots{}
@end example

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -A
@itemx --show-all
@opindex -A
@opindex --show-all
Equivalent to @samp{-vET}.

@item -b
@itemx --number-nonblank
@opindex -b
@opindex --number-nonblank
Number all nonblank output lines, starting with 1.

@item -e
@opindex -e
Equivalent to @samp{-vE}.

@item -E
@itemx --show-ends
@opindex -E
@opindex --show-ends
Display a @samp{$} after the end of each line.

@item -n
@itemx --number
@opindex -n
@opindex --number
Number all output lines, starting with 1.

@item -s
@itemx --squeeze-blank
@opindex -s
@opindex --squeeze-blank
@cindex squeezing blank lines
Replace multiple adjacent blank lines with a single blank line.

@item -t
@opindex -t
Equivalent to @samp{-vT}.

@item -T
@itemx --show-tabs
@opindex -T
@opindex --show-tabs
Display @key{TAB} characters as @samp{^I}.

@item -u
@opindex -u
Ignored; for Unix compatibility.

@item -v
@itemx --show-nonprinting
@opindex -v
@opindex --show-nonprinting
Display control characters except for @key{LFD} and @key{TAB} using
@samp{^} notation and precede characters that have the high bit set
with @samp{M-}.

@end table


@node tac invocation
@section @code{tac}: Concatenate and write files in reverse

@pindex tac
@cindex reversing files

@code{tac} copies each @var{file} (@samp{-} means standard input), or
standard input if none are given, to standard output, reversing the
records (lines by default) in each separately.  Synopsis:

@example
tac [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@dfn{Records} are separated by instances of a string (newline by
default).  By default, this separator string is attached to the end of
the record that it follows in the file.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -b
@itemx --before
@opindex -b
@opindex --before
The separator is attached to the beginning of the record that it
precedes in the file.

@item -r
@itemx --regex
@opindex -r
@opindex --regex
Treat the separator string as a regular expression.

@item -s @var{separator}
@itemx --separator=@var{separator}
@opindex -s
@opindex --separator
Use @var{separator} as the record separator, instead of newline.

@end table


@node nl invocation
@section @code{nl}: Number lines and write files

@pindex nl
@cindex numbering lines
@cindex line numbering

@code{nl} writes each @var{file} (@samp{-} means standard input), or
standard input if none are given, to standard output, with line numbers
added to some or all of the lines.  Synopsis:

@example
nl [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@cindex logical pages, numbering on
@code{nl} decomposes its input into (logical) pages; by default, the
line number is reset to 1 at the top of each logical page.  @code{nl}
treats all of the input files as a single document; it does not reset
line numbers or logical pages between files.

@cindex headers, numbering
@cindex body, numbering
@cindex footers, numbering
A logical page consists of three sections: header, body, and footer.
Any of the sections can be empty.  Each can be numbered in a different
style from the others.

The beginnings of the sections of logical pages are indicated in the
input file by a line containing exactly one of these delimiter strings:

@table @samp
@item \:\:\:
start of header;
@item \:\:
start of body;
@item \:
start of footer.
@end table

The two characters from which these strings are made can be changed from
@samp{\} and @samp{:} via options (see below), but the pattern and
length of each string cannot be changed.

A section delimiter is replaced by an empty line on output.  Any text
that comes before the first section delimiter string in the input file
is considered to be part of a body section, so @code{nl} treats a
file that contains no section delimiters as a single body section.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -b @var{style}
@itemx --body-numbering=@var{style}
@opindex -b
@opindex --body-numbering
Select the numbering style for lines in the body section of each
logical page.  When a line is not numbered, the current line number
is not incremented, but the line number separator character is still
prepended to the line.  The styles are:

@table @samp
@item a
number all lines,
@item t
number only nonempty lines (default for body),
@item n
do not number lines (default for header and footer),
@item p@var{regexp}
number only lines that contain a match for @var{regexp}.
@end table

@item -d @var{cd}
@itemx --section-delimiter=@var{cd}
@opindex -d
@opindex --section-delimiter
@cindex section delimiters of pages
Set the section delimiter characters to @var{cd}; default is
@samp{\:}. If only @var{c} is given, the second remains @samp{:}.
(Remember to protect @samp{\} or other metacharacters from shell
expansion with quotes or extra backslashes.)

@item -f @var{style}
@itemx --footer-numbering=@var{style}
@opindex -f
@opindex --footer-numbering
Analogous to @samp{--body-numbering}.

@item -h @var{style}
@itemx --header-numbering=@var{style}
@opindex -h
@opindex --header-numbering
Analogous to @samp{--body-numbering}.

@item -i @var{number}
@itemx --page-increment=@var{number}
@opindex -i
@opindex --page-increment
Increment line numbers by @var{number} (default 1).

@item -l @var{number}
@itemx --join-blank-lines=@var{number}
@opindex -l
@opindex --join-blank-lines
@cindex empty lines, numbering
@cindex blank lines, numbering
Consider @var{number} (default 1) consecutive empty lines to be one
logical line for numbering, and only number the last one.  Where fewer
than @var{number} consecutive empty lines occur, do not number them.
An empty line is one that contains no characters, not even spaces
or tabs.

@item -n @var{format}
@itemx --number-format=@var{format}
@opindex -n
@opindex --number-format
Select the line numbering format (default is @code{rn}):

@table @samp
@item ln
@opindex ln @r{format for @code{nl}}
left justified, no leading zeros;
@item rn
@opindex rn @r{format for @code{nl}}
right justified, no leading zeros;
@item rz
@opindex rz @r{format for @code{nl}}
right justified, leading zeros.
@end table

@item -p
@itemx --no-renumber
@opindex -p
@opindex --no-renumber
Do not reset the line number at the start of a logical page.

@item -s @var{string}
@itemx --number-separator=@var{string}
@opindex -s
@opindex --number-separator
Separate the line number from the text line in the output with
@var{string} (default is @key{TAB}).

@item -v @var{number}
@itemx --first-page=@var{number}
@opindex -v
@opindex --first-page
Set the initial line number on each logical page to @var{number} (default 1).

@item -w @var{number}
@itemx --number-width=@var{number}
@opindex -w
@opindex --number-width
Use @var{number} characters for line numbers (default 6).

@end table


@node od invocation
@section @code{od}: Write files in octal or other formats

@pindex od
@cindex octal dump of files
@cindex hex dump of files
@cindex ASCII dump of files
@cindex file contents, dumping unambiguously

@code{od} writes an unambiguous representation of each @var{file}
(@samp{-} means standard input), or standard input if none are given.
Synopsis:

@example
od [@var{option}]@dots{} [@var{file}]@dots{}
od -C [@var{file}] [[+]@var{offset} [[+]@var{label}]]
@end example

Each line of output consists of the offset in the input, followed by
groups of data from the file. By default, @code{od} prints the offset in
octal, and each group of file data is two bytes of input printed as a
single octal number.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -A @var{radix}
@itemx --address-radix=@var{radix}
@opindex -A
@opindex --address-radix
@cindex radix for file offsets
@cindex file offset radix
Select the base in which file offsets are printed.  @var{radix} can
be one of the following:

@table @samp
@item d
decimal;
@item o
octal;
@item x
hexadecimal;
@item n
none (do not print offsets).
@end table

The default is octal.

@item -j @var{bytes}
@itemx --skip-bytes=@var{bytes}
@opindex -j
@opindex --skip-bytes
Skip @var{bytes} input bytes before formatting and writing.  If
@var{bytes} begins with @samp{0x} or @samp{0X}, it is interpreted in
hexadecimal; otherwise, if it begins with @samp{0}, in octal; otherwise,
in decimal.  Appending @samp{b} multiplies @var{bytes} by 512, @samp{k}
by 1024, and @samp{m} by 1048576.

@item -N @var{bytes}
@itemx --read-bytes=@var{bytes}
@opindex -N
@opindex --read-bytes
Output at most @var{bytes} bytes of the input.  Prefixes and suffixes on
@code{bytes} are interpreted as for the @samp{-j} option.

@item -s [@var{n}]
@itemx --strings[=@var{n}]
@opindex -s
@opindex --strings
@cindex string constants, outputting
Instead of the normal output, output only @dfn{string constants}: at
least @var{n} (3 by default) consecutive ASCII graphic characters,
followed by a null (zero) byte.

@item -t @var{type}
@itemx --format=@var{type}
@opindex -t
@opindex --format
Select the format in which to output the file data.  @var{type} is a
string of one or more of the below type indicator characters.  If you
include more than one type indicator character in a single @var{type}
string, or use this option more than once, @code{od} writes one copy
of each output line using each of the data types that you specified,
in the order that you specified.

@table @samp
@item a
named character,
@item c
ASCII character or backslash escape,
@item d
signed decimal,
@item f
floating point,
@item o
octal,
@item u
unsigned decimal,
@item x
hexadecimal.
@end table

The type @code{a} outputs things like @samp{sp} for space, @samp{nl} for
newline, and @samp{nul} for a null (zero) byte.  Type @code{c} outputs
@samp{ }, @samp{\n}, and @code{\0}, respectively.

@cindex type size
Except for types @samp{a} and @samp{c}, you can specify the number
of bytes to use in interpreting each number in the given data type
by following the type indicator character with a decimal integer.
Alternately, you can specify the size of one of the C compiler's
built-in data types by following the type indicator character with
one of the following characters.  For integers (@samp{d}, @samp{o},
@samp{u}, @samp{x}):

@table @samp
@item C
char,
@item S
short,
@item I
int,
@item L
long.
@end table

For floating point (@code{f}):

@table @asis
@item F
float,
@item D
double,
@item L
long double.
@end table

@item -v
@itemx --output-duplicates
@opindex -v
@opindex --output-duplicates
Output consecutive lines that are identical.  By default, when two or
more consecutive output lines would be identical, @code{od} outputs only
the first line, and puts just an asterisk on the following line to
indicate the elision.

@item -w[@var{n}]
@itemx --width[=@var{n}]
@opindex -w
@opindex --width
Dump @code{n} input bytes per output line.  This must be a multiple of
the least common multiple of the sizes associated with the specified
output types.  If @var{n} is omitted, the default is 32.  If this option
is not given at all, the default is 16.

@end table

The next several options map the old, pre-POSIX format specification
options to the corresponding POSIX format specs.  GNU @code{od} accepts
any combination of old- and new-style options.  Format specification
options accumulate.

@table @samp

@item -a
@opindex -a
Output as named characters.  Equivalent to @samp{-ta}.

@item -b
@opindex -b
Output as octal bytes.  Equivalent to @samp{-toC}.

@item -c
@opindex -c
Output as ASCII characters or backslash escapes.  Equivalent to
@samp{-tc}.

@item -d
@opindex -d
Output as unsigned decimal shorts.  Equivalent to @samp{-tu2}.

@item -f
@opindex -f
Output as floats.  Equivalent to @samp{-tfF}.

@item -h
@opindex -h
Output as hexadecimal shorts.  Equivalent to @samp{-tx2}.

@item -i
@opindex -i
Output as decimal shorts.  Equivalent to @samp{-td2}.

@item -l
@opindex -l
Output as decimal longs.  Equivalent to @samp{-td4}.

@item -o
@opindex -o
Output as octal shorts.  Equivalent to @samp{-to2}.

@item -x
@opindex -x
Output as hexadecimal shorts.  Equivalent to @samp{-tx2}.

@item -C
@itemx --traditional
@opindex --traditional
Recognize the pre-POSIX non-option arguments that traditional @code{od}
accepted.  The following syntax:

@example
od --traditional [@var{file}] [[+]@var{offset}[.][b] [[+]@var{label}[.][b]]]
@end example

@noindent
can be used to specify at most one file and optional arguments
specifying an offset and a pseudo-start address, @var{label}.  By
default, @var{offset} is interpreted as an octal number specifying how
many input bytes to skip before formatting and writing.  The optional
trailing decimal point forces the interpretation of @var{offset} as a
decimal number.  If no decimal is specified and the offset begins with
@samp{0x} or @samp{0X} it is interpreted as a hexadecimal number.  If
there is a trailing @samp{b}, the number of bytes skipped will be
@var{offset} multiplied by 512.  The @var{label} argument is interpreted
just like @var{offset}, but it specifies an initial pseudo-address.  The
pseudo-addresses are displayed in parentheses following any normal
address.

@end table


@node Formatting file contents
@chapter Formatting file contents

@cindex formatting file contents

These commands reformat the contents of files.

@menu
* fmt invocation::              Reformat paragraph text.
* pr invocation::               Paginate or columnate files for printing.
* fold invocation::             Wrap input lines to fit in specified width.
@end menu


@node fmt invocation
@section @code{fmt}: Reformat paragraph text

@pindex fmt
@cindex reformatting paragraph text
@cindex paragraphs, reformatting
@cindex text, reformatting

@code{fmt} fills and joins lines to produce output lines of (at most)
a given number of characters (75 by default).  Synopsis:

@example
fmt [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@code{fmt} reads from the specified @var{file} arguments (or standard
input if none are given), and writes to standard output.

By default, blank lines, spaces between words, and indentation are
preserved in the output; successive input lines with different
indentation are not joined; tabs are expanded on input and introduced on
output.

@cindex line-breaking
@cindex sentences and line-breaking
@cindex Knuth, Donald E.
@cindex Plass, Michael F.
@code{fmt} prefers breaking lines at the end of a sentence, and tries to
avoid line breaks after the first word of a sentence or before the last
word of a sentence.  A @dfn{sentence break} is defined as either the end
of a paragraph or a word ending in any of @samp{.?!}, followed by two
spaces or end of line, ignoring any intervening parentheses or quotes.
Like @TeX{}, @code{fmt} reads entire ``paragraphs'' before choosing line
breaks; the algorithm is a variant of that in ``Breaking Paragraphs Into
Lines'' (Donald E. Knuth and Michael F. Plass, @cite{Software---Practice
and Experience}, 11 (1981), 1119--1184).

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -c
@itemx --crown-margin
@opindex -c
@opindex --crown-margin
@cindex crown margin
@dfn{Crown margin} mode: preserve the indentation of the first two
lines within a paragraph, and align the left margin of each subsequent
line with that of the second line.

@item -t
@itemx --tagged-paragraph
@opindex -t
@opindex --tagged-paragraph
@cindex tagged paragraphs
@dfn{Tagged paragraph} mode: like crown margin mode, except that if
indentation of the first line of a paragraph is the same as the
indentation of the second, the first line is treated as a one-line
paragraph.

@item -s
@itemx --split-only
@opindex -s
@opindex --split-only
Split lines only.  Do not join short lines to form longer ones.  This
prevents sample lines of code, and other such ``formatted'' text from
being unduly combined.

@item -u
@itemx --uniform-spacing
@opindex -u
@opindex --uniform-spacing
Uniform spacing.  Reduce spacing between words to one space, and spacing
between sentences to two spaces.

@item -@var{width}
@itemx -w @var{width}
@itemx --width=@var{width}
@opindex -@var{width}
@opindex -w
@opindex --width
Fill output lines up to @var{width} characters (default 75).  @code{fmt}
initially tries to make lines about 7% shorter than this, to give it
room to balance line lengths.

@item -p @var{prefix}
@itemx --prefix=@var{prefix}
Only lines beginning with @var{prefix} (possibly preceded by whitespace)
are subject to formatting. The prefix and any preceding whitespace is
stripped for the formatting and then re-attached to each formatted output
line.  One use is to format certain kinds of program comments, while
leaving the code unchanged.

@end table


@node pr invocation
@section @code{pr}: Paginate or columnate files for printing

@pindex pr
@cindex printing, preparing files for
@cindex multicolumn output, generating

@code{pr} writes each @var{file} (@samp{-} means standard input), or
standard input if none are given, to standard output, paginating and
optionally outputting in multicolumn format.  Synopsis:

@example
pr [@var{option}]@dots{} [@var{file}]@dots{}
@end example

By default, a 5-line header is printed: two blank lines; a line with the
date, the file name, and the page count; and two more blank lines.  A
five line footer (entirely) is also printed.

Form feeds in the input cause page breaks in the output.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item +@var{page}
Begin printing with page @var{page}.

@item -@var{column}
@opindex -@var{column}
Produce @var{column}-column output and print columns down.  The column
width is automatically decreased as @var{column} increases; unless you
use the @samp{-w} option to increase the page width as well, this option
might well cause some input to be truncated.

@item -a
@opindex -a
@cindex across columns
Print columns across rather than down.

@item -b
@opindex -b
@cindex balancing columns
Balance columns on the last page.

@item -c
@opindex -c
Print control characters using hat notation (e.g., @samp{^G}); print
other unprintable characters in octal backslash notation.  By default,
unprintable characters are not changed.

@item -d
@opindex -d
@cindex double spacing
Double space the output.

@item -e[@var{in-tabchar}[@var{in-tabwidth}]]
@opindex -e
@cindex input tabs
Expand tabs to spaces on input.  Optional argument @var{in-tabchar} is
the input tab character (default is @key{TAB}).  Second optional
argument @var{in-tabwidth} is the input tab character's width (default
is 8).

@item -f
@itemx -F
@opindex -F
@opindex -f
Use a formfeed instead of newlines to separate output pages.

@item -h @var{header}
@opindex -h
Replace the file name in the header with the string @var{header}.

@item -i[@var{out-tabchar}[@var{out-tabwidth}]]
@opindex -i
@cindex output tabs
Replace spaces with tabs on output.  Optional argument @var{out-tabchar}
is the output tab character (default is @key{TAB}).  Second optional
argument @var{out-tabwidth} is the output tab character's width (default
is 8).

@item -l @var{n}
@opindex -l
Set the page length to @var{n} (default 66) lines.  If @var{n} is less
than 10, the headers and footers are omitted, as if the @samp{-t} option
had been given.

@item -m
@opindex -m
Print all files in parallel, one in each column.

@item -n[@var{number-separator}[@var{digits}]]
@opindex -n
Precede each column with a line number; with parallel files (@samp{-m}),
precede each line with a line number.  Optional argument
@var{number-separator} is the character to print after each number
(default is @key{TAB}).  Optional argument @var{digits} is the number of
digits per line number (default is 5).

@item -o @var{n}
@opindex -o
@cindex indenting lines
@cindex left margin
Indent each line with @var{n} (default is zero) spaces wide, i.e., set
the left margin.  The total page width is @samp{n} plus the width set
with the @samp{-w} option.

@item -r
@opindex -r
Do not print a warning message when an argument @var{file} cannot be
opened.  (The exit status will still be nonzero, however.)

@item -s[@var{c}]
@opindex -s
Separate columns by the single character @var{c}.  If @var{c} is
omitted, the default is space; if this option is omitted altogether, the
default is @key{TAB}.

@item -t
@opindex -t
Do not print the usual 5-line header and the 5-line footer on each page,
and do not fill out the bottoms of pages (with blank lines or
formfeeds).

@item -v
@opindex -v
Print unprintable characters in octal backslash notation.

@item -w @var{n}
@opindex -w
Set the page width to @var{n} (default is 72) columns.

@end table


@node fold invocation
@section @code{fold}: Wrap input lines to fit in specified width

@pindex fold
@cindex wrapping long input lines
@cindex folding long input lines

@code{fold} writes each @var{file} (@samp{-} means standard input), or
standard input if none are given, to standard output, breaking long
lines.  Synopsis:

@example
fold [@var{option}]@dots{} [@var{file}]@dots{}
@end example

By default, @code{fold} breaks lines wider than 80 columns. The output
is split into as many lines as necessary.

@cindex screen columns
@code{fold} counts screen columns by default; thus, a tab may count more
than one column, backspace decreases the column count, and carriage
return sets the column to zero.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -b
@itemx --bytes
@opindex -b
@opindex --bytes
Count bytes rather than columns, so that tabs, backspaces, and carriage
returns are each counted as taking up one column, just like other
characters.

@item -s
@itemx --spaces
@opindex -s
@opindex --spaces
Break at word boundaries: the line is broken after the last blank before
the maximum line length.  If the line contains no such blanks, the line
is broken at the maximum line length as usual.

@item -w @var{width}
@itemx --width=@var{width}
@opindex -w
@opindex --width
Use a maximum line length of @var{width} columns instead of 80.

@end table


@node Output of parts of files
@chapter Output of parts of files

@cindex output of parts of files
@cindex parts of files, output of

These commands output pieces of the input.

@menu
* head invocation::             Output the first part of files.
* tail invocation::             Output the last part of files.
* split invocation::            Split a file into fixed-size pieces.
* csplit invocation::           Split a file into context-determined pieces.
@end menu

@node head invocation
@section @code{head}: Output the first part of files

@pindex head
@cindex initial part of files, outputting
@cindex first part of files, outputting

@code{head} prints the first part (10 lines by default) of each
@var{file}; it reads from standard input if no files are given or
when given a @var{file} of @samp{-}.  Synopses:

@example
head [@var{option}]@dots{} [@var{file}]@dots{}
head -@var{number} [@var{option}]@dots{} [@var{file}]@dots{}
@end example

If more than one @var{file} is specicified, @code{head} prints a
one-line header consisting of
@example
==> @var{file name} <==
@end example
@noindent
before the output for each @var{file}.

@code{head} accepts two option formats: the new one, in which numbers
are arguments to the options (@samp{-q -n 1}), and the old one, in which
the number precedes any option letters (@samp{-1q}).

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{count}@var{options}
@opindex -@var{count}
This option is only recognized if it is specified first.  @var{count} is
a decimal number optionally followed by a size letter (@samp{b},
@samp{k}, @samp{m}) as in @code{-c}, or @samp{l} to mean count by lines,
or other option letters (@samp{cqv}).

@item -c @var{bytes}
@itemx --bytes=@var{bytes}
@opindex -c
@opindex --bytes
Print the first @var{bytes} bytes, instead of initial lines.  Appending
@samp{b} multiplies @var{bytes} by 512, @samp{k} by 1024, and @samp{m}
by 1048576.

@itemx -n @var{n}
@itemx --lines=@var{n}
@opindex -n
@opindex --lines
Output the first @var{n} lines.

@item -q
@itemx --quiet
@itemx --silent
@opindex -q
@opindex --quiet
@opindex --silent
Never print file name headers.

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Always print file name headers.

@end table


@node tail invocation
@section @code{tail}: Output the last part of files

@pindex tail
@cindex last part of files, outputting

@code{tail} prints the last part (10 lines by default) of each
@var{file}; it reads from standard input if no files are given or
when given a @var{file} of @samp{-}.  Synopses:

@example
tail [@var{option}]@dots{} [@var{file}]@dots{}
tail -@var{number} [@var{option}]@dots{} [@var{file}]@dots{}
tail +@var{number} [@var{option}]@dots{} [@var{file}]@dots{}
@end example

If more than one @var{file} is specified, @code{tail} prints a
one-line header consisting of
@example
==> @var{file name} <==
@end example
@noindent
before the output for each @var{file}.

@cindex BSD @code{tail}
GNU @code{tail} can output any amount of data (some other versions of
@code{tail} cannot).  It also has no @samp{-r} option (print in
reverse), since reversing a file is really a different job from printing
the end of a file; BSD @code{tail} (which is the one with @code{-r}) can
only reverse files that are at most as large as its buffer, which is
typically 32k.  A more reliable and versatile way to reverse files is
the GNU @code{tac} command.

@code{tail} accepts two option formats: the new one, in which numbers
are arguments to the options (@samp{-n 1}), and the old one, in which
the number precedes any option letters (@samp{-1} or @samp{+1}).

If any option-argument is a number @var{n} starting with a @samp{+},
@code{tail} begins printing with the @var{n}th item from the start of
each file, instead of from the end.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{count}
@itemx +@var{count}
@opindex -@var{count}
@opindex +@var{count}
This option is only recognized if it is specified first.  @var{count} is
a decimal number optionally followed by a size letter (@samp{b},
@samp{k}, @samp{m}) as in @code{-c}, or @samp{l} to mean count by lines,
or other option letters (@samp{cfqv}).

@item -c @var{bytes}
@itemx --bytes=@var{bytes}
@opindex -c
@opindex --bytes
Output the last @var{bytes} bytes, instead of final lines.  Appending
@samp{b} multiplies @var{bytes} by 512, @samp{k} by 1024, and @samp{m}
by 1048576.

@item -f
@itemx --follow
@opindex -f
@opindex --follow
@cindex growing files
Loop forever trying to read more characters at the end of the file,
presumably because the file is growing.  Ignored if reading from a pipe.
If more than one file is given, @code{tail} prints a header whenever it
gets output from a different file, to indicate which file that output is
from.

@itemx -n @var{n}
@itemx --lines=@var{n}
@opindex -n
@opindex --lines
Output the last @var{n} lines.

@item -q
@itemx -quiet
@itemx --silent
@opindex -q
@opindex --quiet
@opindex --silent
Never print file name headers.

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Always print file name headers.

@end table


@node split invocation
@section @code{split}: Split a file into fixed-size pieces

@pindex split
@cindex splitting a file into pieces
@cindex pieces, splitting a file into

@code{split} creates output files containing consecutive sections of
@var{input} (standard input if none is given or @var{input} is
@samp{-}).  Synopsis:

@example
split [@var{option}] [@var{input} [@var{prefix}]]
@end example

By default, @code{split} puts 1000 lines of @var{input} (or whatever is
left over for the last section), into each output file.

@cindex output file name prefix
The output files' names consist of @var{prefix} (@samp{x} by default)
followed by a group of letters @samp{aa}, @samp{ab}, and so on, such
that concatenating the output files in sorted order by file name produces
the original input file.  (If more than 676 output files are required,
@code{split} uses @samp{zaa}, @samp{zab}, etc.)

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{lines}
@itemx -l @var{lines}
@itemx --lines=@var{lines}
@opindex -l
@opindex --lines
Put @var{lines} lines of @var{input} into each output file.

@item -b @var{bytes}
@itemx --bytes=@var{bytes}
@opindex -b
@opindex --bytes
Put the first @var{bytes} bytes of @var{input} into each output file.
Appending @samp{b} multiplies @var{bytes} by 512, @samp{k} by 1024, and
@samp{m} by 1048576.

@item -C @var{bytes}
@itemx --line-bytes=@var{bytes}
@opindex -C
@opindex --line-bytes
Put into each output file as many complete lines of @var{input} as
possible without exceeding @var{bytes} bytes.  For lines longer than
@var{bytes} bytes, put @var{bytes} bytes into each output file until
less than @var{bytes} bytes of the line are left, then continue
normally.  @var{bytes} has the same format as for the @samp{--bytes}
option.

@end table


@node csplit invocation
@section @code{csplit}: Split a file into context-determined pieces

@pindex csplit
@cindex context splitting
@cindex splitting a file into pieces by context

@code{csplit} creates zero or more output files containing sections of
@var{input} (standard input if @var{input} is @samp{-}).  Synopsis:

@example
csplit [@var{option}]@dots{} @var{input} @var{pattern}@dots{}
@end example

The contents of the output files are determined by the @var{pattern}
arguments, as detailed below.  An error occurs if a @var{pattern}
argument refers to a nonexistent line of the input file (e.g., if no
remaining line matches a given regular expression).  After every
@var{pattern} has been matched, any remaining input is copied into one
last output file.

By default, @code{csplit} prints the number of bytes written to each
output file after it has been created.

The types of pattern arguments are:

@table @samp

@item @var{n}
Create an output file containing the input up to but not including line
@var{n} (a positive integer).  If followed by a repeat count, also
create an output file containing the next @var{line} lines of the input
file once for each repeat.

@item /@var{regexp}/[@var{offset}]
Create an output file containing the current line up to (but not
including) the next line of the input file that contains a match for
@var{regexp}.  The optional @var{offset} is a @samp{+} or @samp{-}
followed by a positive integer.  If it is given, the input up to the
matching line plus or minus @var{offset} is put into the output file,
and the line after that begins the next section of input.

@item %@var{regexp}%[@var{offset}]
Like the previous type, except that it does not create an output
file, so that section of the input file is effectively ignored.

@item @{@var{repeat-count}@}
Repeat the previous pattern @var{repeat-count} additional
times. @var{repeat-count} can either be a positive integer or an
asterisk, meaning repeat as many times as necessary until the input is
exausted.

@end table

The output files' names consist of a prefix (@samp{xx} by default)
followed by a suffix.  By default, the suffix is an ascending sequence
of two-digit decimal numbers from @samp{00} and up to @samp{99}.  In any
case, concatenating the output files in sorted order by filename
produces the original input file.

By default, if @code{csplit} encounters an error or receives a hangup,
interrupt, quit, or terminate signal, it removes any output files
that it has created so far before it exits.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -f @var{prefix}
@itemx --prefix=@var{prefix}
@opindex -f
@opindex --prefix
@cindex output file name prefix
Use @var{prefix} as the output file name prefix.

@item -b @var{suffix}
@itemx --suffix=@var{suffix}
@opindex -b
@opindex --suffix
@cindex output file name suffix
Use @var{suffix} as the output file name suffix.  When this option is
specified, the suffix string must include exactly one
@code{printf(3)}-style conversion specification, possibly including
format specification flags, a field width, a precision specifications,
or all of these kinds of modifiers.  The format letter must convert a
binary integer argument to readable form; thus, only @samp{d}, @samp{i},
@samp{u}, @samp{o}, @samp{x}, and @samp{X} conversions are allowed.  The
entire @var{suffix} is given (with the current output file number) to
@code{sprintf(3)} to form the file name suffixes for each of the
individual output files in turn.  If this option is used, the
@samp{--digits} option is ignored.

@item -n @var{digits}
@itemx --digits=@var{digits}
@opindex -n
@opindex --digits
Use output file names containing numbers that are @var{digits} digits
long instead of the default 2.

@item -k
@itemx --keep-files
@opindex -k
@opindex --keep-files
Do not remove output files when errors are encountered.

@item -z
@itemx --elide-empty-files
@opindex -z
@opindex --elide-empty-files
Suppress the generation of zero-length output files.  (In cases where
the section delimiters of the input file are supposed to mark the first
lines of each of the sections, the first output file will generally be a
zero-length file unless you use this option.)  The output file sequence
numbers always run consecutively starting from 0, even when this option
is specified.

@item -s
@itemx -q
@itemx --silent
@itemx --quiet
@opindex -s
@opindex -q
@opindex --silent
@opindex --quiet
Do not print counts of output file sizes.

@end table


@node Summarizing files
@chapter Summarizing files

@cindex summarizing files

These commands generate just a few numbers representing entire
contents of files.

@menu
* wc invocation::               Print byte, word, and line counts.
* sum invocation::              Print checksum and block counts.
* cksum invocation::            Print CRC checksum and byte counts.
* md5sum invocation::           Print or check message-digests.
@end menu


@node wc invocation
@section @code{wc}: Print byte, word, and line counts

@pindex wc
@cindex byte count
@cindex word count
@cindex line count

@code{wc} counts the number of bytes, whitespace-separated words, and
newlines in each given @var{file}, or standard input if none are given
or for a @var{file} of @samp{-}.  Synopsis:

@example
wc [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@cindex total counts
@code{wc} prints one line of counts for each file, and if the file was
given as an argument, it prints the file name following the counts.  If
more than one @var{file} is given, @code{wc} prints a final line
containing the cumulative counts, with the file name @file{total}.  The
counts are printed in this order: newlines, words, bytes.

By default, @code{wc} prints all three counts.  Options can specify
that only certain counts be printed.  Options do not undo others
previously given, so

@example
wc --bytes --words
@end example

@noindent
prints both the byte counts and the word counts.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -c
@itemx --bytes
@itemx --chars
@opindex -c
@opindex --bytes
@opindex --chars
Print only the byte counts.

@item -w
@itemx --words
@opindex -w
@opindex --words
Print only the word counts.

@item -l
@itemx --lines
@opindex -l
@opindex --lines
Print only the newline counts.

@end table


@node sum invocation
@section @code{sum}: Print checksum and block counts

@pindex sum
@cindex 16-bit checksum
@cindex checksum, 16-bit

@code{sum} computes a 16-bit checksum for each given @var{file}, or
standard input if none are given or for a @var{file} of @samp{-}.  Synopsis:

@example
sum [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@code{sum} prints the checksum for each @var{file} followed by the
number of blocks in the file (rounded up).  If more than one @var{file}
is given, file names are also printed (by default).  (With the
@samp{--sysv} option, corresponding file name are printed when there is
at least one file argument.)

By default, GNU @code{sum} computes checksums using an algorithm
compatible with BSD @code{sum} and prints file sizes in units of
1024-byte blocks.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -r
@opindex -r
@cindex BSD @code{sum}
Use the default (BSD compatible) algorithm.  This option is included for
compatibility with the System V @code{sum}.  Unless @samp{-s} was also
given, it has no effect.

@item -s
@itemx --sysv
@opindex -s
@opindex --sysv
@cindex System V @code{sum}
Compute checksums using an algorithm compatible with System V
@code{sum}'s default, and print file sizes in units of 512-byte blocks.

@end table

@code{sum} is provided for compatibility; the @code{cksum} program (see
next section) is preferable in new applications.


@node cksum invocation
@section @code{cksum}: Print CRC checksum and byte counts

@pindex cksum
@cindex cyclic redundancy check
@cindex CRC checksum

@code{cksum} computes a cyclic redundancy check (CRC) checksum for each
given @var{file}, or standard input if none are given or for a
@var{file} of @samp{-}.  Synopsis:

@example
cksum [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@code{cksum} prints the CRC checksum for each file along with the number
of bytes in the file, and the filename unless no arguments were given.

@code{cksum} is typically used to ensure that files have been
transferred by unreliable means (e.g., netnews) have not been corrupted,
by comparing the @code{cksum} output for the received files with the
@code{cksum} output for the original files (typically given in the
distribution).

The CRC algorithm is specified by the POSIX.2 standard.  It is not
compatible with the BSD or System V @code{sum} algorithms (see the
previous section); it is more robust.

The only options are @samp{--help} and @samp{--version}.  @xref{Common
options}.


@node md5sum invocation
@section @code{md5sum}: Print or check message-digests

@pindex md5sum
@cindex 128-bit checksum
@cindex checksum, 128-bit
@cindex fingerprint, 128-bit
@cindex message-digest, 128-bit

@code{md5sum} computes a 128-bit checksum (or @dfn{fingerprint} or
@dfn{message-digest} for each given @var{file}, or standard input if
none are given or for a @var{file} of @samp{-}.  It can also check if the
checksum has changed. Synopsis:

@example
md5sum [@var{option}]@dots{} [@var{file}]@dots{}
@end example

For each @var{file}, @samp{md5sum} outputs the MD5 checksum, a flag
indicating a binary or text input file, and the filename.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -b
@itemx --binary
@opindex -b
@opindex --binary
@cindex binary input files
Treat input files as binary.  This makes no difference on Unix systems,
but other systems have different internal and external character
representations, notably to mark end-of-line.

@item -c
@itemx --check=@var{file}
@var{file} is taken as the output of a former run of @samp{md5sum}: each
line consists of an MD5 checksum, a binary/text flag, and a filename.
The file will be opened (with each possible relative path) and the its
message-digest computed.  If this computed message digest is not the
same as that given in the line, the file will be marked as failed.

@item -s
@itemx --string=@var{string}
@opindex -s
@opindex --string
Compute the message digest for @var{string}, instead of for a file.  The
result is the same as for a file with contains exactly @var{string}.

@item -t
@itemx --text
@opindex -t
@opindex --text
@cindex text input files
Treat all input files as text files.  This is the reverse of
@samp{--binary}.
@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Print progress information.

@end table


@node Operating on sorted files
@chapter Operating on sorted files

@cindex operating on sorted files
@cindex sorted files, operations on

These commands work with (or produce) sorted files.

@menu
* sort invocation::             Sort text files.
* uniq invocation::             Uniqify files.
* comm invocation::             Compare two sorted files line by line.
@end menu


@node sort invocation
@section @code{sort}: Sort text files

@pindex sort
@cindex sorting files

@code{sort} sorts, merges, or compares all the lines from the given
files, or standard input if none are given or for a @var{file} of
@samp{-}.  By default, @code{sort} writes the results to standard
output.  Synopsis:

@example
sort [@var{option}]@dots{} [@var{file}]@dots{}
@end example

@code{sort} has three modes of operation: sort (the default), merge,
and check for sortedness.  The following options change the operation
mode:

@table @samp

@item -c
@opindex -c
@cindex checking for sortedness
Check whether the given files are already sorted: if they are not all
sorted, print an error message and exit with a status of 1.

@item -m
@opindex -m
@cindex merging sorted files
Merge the given files by sorting them as a group.  Each input file must
always be individually sorted.  It always works to sort instead of
merge; merging is provided because it is faster, in the case where it
works.

@end table

A pair of lines is compared as follows: if any key fields have been
specified, @code{sort} compares each pair of fields, in the order
specified on the command line, according to the associated ordering
options, until a difference is found or no fields are left.

If any of the global options @samp{Mbdfinr} are given but no key fields
are specified, @code{sort} compares the entire lines according to the
global options.

Finally, as a last resort when all keys compare equal (or if no
ordering options were specified at all), @code{sort} compares the lines
byte by byte in machine collating sequence.  The last resort comparison
honors the @samp{-r} global option.  The @samp{-s} (stable) option
disables this last-resort comparison so that lines in which all fields
compare equal are left in their original relative order.  If no fields
or global options are specified, @samp{-s} has no effect.

GNU @code{sort} (as specified for all GNU utilities) has no limits on
input line length or restrictions on bytes allowed within lines.  In
addition, if the final byte of an input file is not a newline, GNU
@code{sort} silently supplies one.

@vindex TMPDIR
If the environment variable @code{TMPDIR} is set, @code{sort} uses its
value as the directory for temporary files instead of @file{/tmp}.  The
@samp{-T @var{tempdir}} option in turn overrides the environment
variable.

The following options affect the ordering of output lines.  They may be
specified globally or as part of a specific key field.  If no key
fields are specified, global options apply to comparison of entire
lines; otherwise the global options are inherited by key fields that do
not specify any special options of their own.

@table @samp

@item -b
@opindex -b
@cindex blanks, ignoring leading
Ignore leading blanks when finding sort keys in each line.

@item -d
@opindex -d
@cindex phone directory order
@cindex telephone directory order
Sort in @dfn{phone directory} order: ignore all characters except
letters, digits and blanks when sorting.

@item -f
@opindex -f
@cindex case folding
Fold lowercase characters into the equivalent uppercase characters when
sorting so that, for example, @samp{b} and @samp{B} sort as equal.

@item -i
@opindex -i
@cindex unprintable characters, ignoring
Ignore characters outside the printable ASCII range 040-0176 octal
(inclusive) when sorting.

@item -M
@opindex -M
@cindex months, sorting by
An initial string, consisting of any amount of whitespace, followed
by three letters abbreviating a month name, is folded to UPPER case and
compared in the order @samp{JAN} < @samp{FEB} < @dots{} < @samp{DEC}.
Invalid names compare low to valid names.

@item -n
@opindex -n
@cindex numeric sort
Sort numerically: the number begins each line; specifically, it consists
of optional whitespace, an optional @samp{-} sign, and zero or more
digits, optionally followed by a decimal point and zero or more digits.

@item -r
@opindex -r
@cindex reverse sorting
Reverse the result of comparison, so that lines with greater key values
appear earlier in the output instead of later.

@end table

Other options are:

@table @samp

@item -o @var{output-file}
@opindex -o
@cindex overwriting of input, allowed
Write output to @var{output-file} instead of standard output.
If @var{output-file} is one of the input files, @code{sort} copies
it to a temporary file before sorting and writing the output to
@var{output-file}.

@item -t @var{separator}
@opindex -t
@cindex field separator character
Use character @var{separator} as the field separator when finding the
sort keys in each line.  By default, fields are separated by the empty
string between a non-whitespace character and a whitespace character.
That is, given the input line @w{@samp{ foo bar}}, @code{sort} breaks it
into fields @w{@samp{ foo}} and @w{@samp{ bar}}.  The field separator is
not considered to be part of either the field preceding or the field
following.

@item -u
@opindex -u
@cindex uniqifying output
For the default case or the @samp{-m} option, only output the first
of a sequence of lines that compare equal.  For the @samp{-c} option,
check that no pair of consecutive lines compares equal.

@item -k @var{pos1}[,@var{pos2}]
@opindex -k
@cindex sort field
The recommended, POSIX, option for specifying a sort field.  The field
consists of the line between @var{pos1} and @var{pos2} (or the end of
the line, if @var{pos2} is omitted), inclusive.  Fields and character
positions are numbered starting with 1.  See below.

@item +@var{pos1}[-@var{pos2}]
The obsolete, traditional option for specifying a sort field.  The field
consists of the line between @var{pos1} and up to but not including
@var{pos2} (or the end of the line if @var{pos2} is omitted).  Fields
and character positions are numbered starting with 0.  See below.

@end table

In addition, when GNU @code{sort} is invoked with exactly one argument,
options @samp{--help} and @samp{--version} are recognized.  @xref{Common
options}.

Historical (BSD and System V) implementations of @code{sort} have
differed in their interpretation of some options, particularly
@samp{-b}, @samp{-f}, and @samp{-n}.  GNU sort follows the POSIX
behavior, which is usually (but not always!) like the System V behavior.
According to POSIX, @samp{-n} no longer implies @samp{-b}.  For
consistency, @samp{-M} has been changed in the same way.  This may
affect the meaning of character positions in field specifications in
obscure cases.  The only fix is to add an explicit @samp{-b}.

A position in a sort field specified with the @samp{-k} or @samp{+}
option has the form @samp{@var{f}.@var{c}}, where @var{f} is the number
of the field to use and @var{c} is the number of the first character
from the beginning of the field (for @samp{+@var{pos}}) or from the end
of the previous field (for @samp{-@var{pos}}).  If the @samp{.@var{c}}
is omitted, it's taken to be the first character in the field.  If the
@samp{-b} option was specified, the @samp{.@var{c}} part of a field
specification is counted from the first nonblank character of the field
(for @samp{+@var{pos}}) or from the first nonblank character following
the previous field (for @samp{-@var{pos}}).

A sort key option may also have any of the option letters @samp{Mbdfinr}
appended to it, in which case the global ordering options are not used
for that particular field.  The @samp{-b} option may be independently
attached to either or both of the @samp{+@var{pos}} and
@samp{-@var{pos}} parts of a field specification, and if it is inherited
from the global options it will be attached to both.  If a @samp{-n} or
@samp{-M} option is used, thus implying a @samp{-b} option, the
@samp{-b} option is taken to apply to both the @samp{+@var{pos}} and the
@samp{-@var{pos}} parts of a key specification.  Keys may span multiple
fields.


@node uniq invocation
@section @code{uniq}: Uniqify files

@pindex uniq
@cindex uniqify files

@code{uniq} writes the unique lines in the given @file{input}, or
standard input if nothing is given or for an @var{input} name of
@samp{-}.  Synopsis:

@example
uniq [@var{option}]@dots{} [@var{input} [@var{output}]]
@end example

By default, @code{uniq} prints the unique lines in a sorted file, i.e.,
discards all but one of identical successive lines.  Optionally, it can
instead show only lines that appear exactly once, or lines that appear
more than once.

The input must be sorted.  If your input is not sorted, perhaps you want
to use @code{sort -u}.

If no @var{output} file is specified, @code{uniq} writes to standard
output.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{n}
@itemx -f @var{n}
@itemx --skip-fields=@var{n}
@opindex -@var{n}
@opindex -f
@opindex --skip-fields
Skip @var{n} fields on each line before checking for uniqueness.  Fields
are sequences of non-space non-tab characters that are separated from
each other by at least one spaces or tabs.

@item +@var{n}
@itemx -s @var{n}
@itemx --skip-chars=@var{n}
@opindex +@var{n}
@opindex -s
@opindex --skip-chars
Skip @var{n} characters before checking for uniqueness.  If you use both
the field and character skipping options, fields are skipped over first.

@item -c
@itemx --count
@opindex -c
@opindex --count
Print the number of times each line occurred along with the line.

@item -d
@itemx --repeated
@opindex -d
@opindex --repeated
@cindex duplicate lines, outputting
Print only duplicate lines.

@item -u
@itemx --unique
@opindex -u
@opindex --unique
@cindex unique lines, outputting
Print only unique lines.

@item -w @var{n}
@itemx --check-chars=@var{n}
@opindex -w
@opindex --check-chars
Compare @var{n} characters on each line (after skipping any specified
fields and characters).  By default the entire rest of the lines are
compared.

@end table


@node comm invocation
@section @code{comm}: Compare two sorted files line by line

@pindex comm
@cindex line-by-line comparison
@cindex comparing sorted files

@code{comm} writes to standard output lines that are common, and lines
that are unique, to two input files; a file name of @samp{-} means
standard input.  Synopsis:

@example
comm [@var{option}]@dots{} @var{file1} @var{file2}
@end example

The input files must be sorted before @code{comm} can be used.

@cindex differing lines
@cindex common lines
With no options, @code{comm} produces three column output.  Column one
contains lines unique to @var{file1}, column two contains lines unique
to @var{file2}, and column three contains lines common to both files.

@opindex -1
@opindex -2
@opindex -3
The options @samp{-1}, @samp{-2}, and @samp{-3} suppress printing of
the corresponding columns.  Also see @ref{Common options}.


@node Operating on fields within a line
@chapter Operating on fields within a line

@menu
* cut invocation::              Print selected parts of lines.
* paste invocation::            Merge lines of files.
* join invocation::             Join lines on a common field.
@end menu


@node cut invocation
@section @code{cut}: Print selected parts of lines

@pindex cut
@code{cut} writes to standard output selected parts of each line of each
input file, or standard input if no files are given or for a file name of
@samp{-}.  Synopsis:

@example
cut [@var{option}]@dots{} [@var{file}]@dots{}
@end example

In the table which follows, the @var{byte-list}, @var{character-list},
and @var{field-list} are one or more numbers or ranges (two numbers
separated by a dash) separated by commas.  Bytes, characters, and
fields are numbered from starting at 1.  Incomplete ranges may be
given: @samp{-@var{m}} means @samp{1-@var{m}}; @samp{@var{n}-} means
@samp{@var{n}} through end of line or last field.

The program accepts the following options.  Also see @ref{Common
options}.

@table @samp

@item -b @var{byte-list}
@itemx --bytes=@var{byte-list}
@opindex -b
@opindex --bytes
Print only the bytes in positions listed in @var{byte-list}.  Tabs and
backspaces are treated like any other character; they take up 1 byte.

@item -c @var{character-list}
@itemx --characters=@var{character-list}
@opindex -c
@opindex --characters
Print only characters in positions listed in @var{character-list}.
The same as @samp{-b} for now, but internationalization will change
that.  Tabs and backspaces are treated like any other character; they
take up 1 character.

@item -f @var{field-list}
@itemx --fields=@var{field-list}
@opindex -f
@opindex --fields
Print only the fields listed in @var{field-list}.  Fields are
separated by a @key{TAB} by default.

@item -d @var{delim}
@itemx --delimiter=@var{delim}
@opindex -d
@opindex --delimiter
For @samp{-f}, fields are separated by the first character in @var{delim}
(default is @key{TAB}).

@item -n
@opindex -n
Do not split multibyte characters (no-op for now).

@item -s
@itemx --only-delimited
@opindex -s
@opindex --only-delimited
For @samp{-f}, do not print lines that do not contain the field separator
character.

@end table


@node paste invocation
@section @code{paste}: Merge lines of files

@pindex paste
@cindex merging files

@code{paste} writes to standard output lines consisting of sequentially
corresponding lines of each given file, separated by @key{TAB}.
Standard input is used for a file name of @samp{-} or if no input files
are given.

Synopsis:

@example
paste [@var{option}]@dots{} [@var{file}]@dots{}
@end example

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -s
@itemx --serial
@opindex -s
@opindex --serial
Paste the lines of one file at a time rather than one line from each
file.

@item -d @var{delim-list}
@itemx --delimiters @var{delim-list}
@opindex -d
@opindex --delimiters
Consecutively use the characters in @var{delim-list} instead of
@key{TAB} to separate merged lines.  When @var{delim-list} is
exhausted, start again at its beginning.

@end table


@node join invocation
@section @code{join}: Join lines on a common field

@pindex join
@cindex common field, joining on

@code{join} writes to standard output a line for each pair of input
lines that have identical join fields.  Synopsis:

@example
join [@var{option}]@dots{} @var{file1} @var{file2}
@end example

Either @var{file1} or @var{file2} (but not both) can be @samp{-},
meaning standard input.  @var{file1} and @var{file2} should be already
sorted in increasing order (not numerically) on the join fields; unless
the @samp{-t} option is given, they should be sorted ignoring blanks at
the start of the line, as in @code{sort -b}.

The defaults are: the join field is the first field in each line;
fields in the input are separated by one or more blanks, with leading
blanks on the line ignored; fields in the output are separated by a
space; each output line consists of the join field, the remaining
fields from @var{file1}, then the remaining fields from @var{file2}.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -a @var{file-number}
@opindex -a
Print a line for each unpairable line in file @var{file-number} (either
@samp{1} or @samp{2}), in addition to the normal output.

@item -e @var{string}
@opindex -e
Replace those output fields that are missing in the input with
@var{string}.

@item -1 @var{field}
@itemx -j1 @var{field}
@opindex -1
@opindex -j1
Join on field @var{field} (a positive integer) of file 1.

@item -2 @var{field}
@itemx -j2 @var{field}
@opindex -2
@opindex -j2
Join on field @var{field} (a positive integer) of file 2.

@item -j @var{field}
Equivalent to @samp{-1 @var{field} -2 @var{field}}.

@item -o @var{field-list}@dots{}
Construct each output line according to the format in @var{field-list}.
Each element in @var{field-list} consists of a file number (either 1 or
2), a period, and a field number (a positive integer).  The elements in
the list are separated by commas or blanks.  Multiple @var{field-list}
arguments can be given after a single @samp{-o} option; the values
of all lists given with @samp{-o} are concatenated together.

@item -t @var{char}
Use character @var{char} as the input and output field separator.

@item -v @var{file-number}
Print a line for each unpairable line in file @var{file-number}
(either 1 or 2), instead of the normal output.

@end table

In addition, when GNU @code{join} is invoked with exactly one argument,
options @samp{--help} and @samp{--version} are recognized.  @xref{Common
options}.


@node Operating on characters
@chapter Operating on characters

@cindex operating on characters

This commands operate on individual characters.

@menu
* tr invocation::               Translate, squeeze, and/or delete characters.
* expand invocation::           Convert tabs to spaces.
* unexpand invocation::         Convert spaces to tabs.
@end menu


@node tr invocation
@section @code{tr}: Translate, squeeze, and/or delete characters

@pindex tr

Synopsis:

@example
tr [@var{option}]@dots{} @var{set1} [@var{set2}]
@end example

@code{tr} copies standard input to standard output, performing
one of the following operations:

@itemize @bullet
@item
translate, and optionally squeeze repeated characters in the result,
@item
squeeze repeated characters,
@item
delete characters,
@item
delete characters, then squeeze repeated characters from the result.
@end itemize

The @var{set1} and (if given) @var{set2} arguments define ordered
sets of characters, referred to below as @var{set1} and @var{set2}.  These
sets are the characters of the input that @code{tr} operates on.
The @samp{--complement} (@samp{-c}) option replaces @var{set1} with its
complement (all of the characters that are not in @var{set1}).

@menu
* Character sets::              Specifying sets of characters.
* Translating::                 Changing one characters to another.
* Squeezing::                   Squeezing repeats and deleting.
* Warnings in tr::              Warning messages.
@end menu


@node Character sets
@subsection Specifying sets of characters

@cindex specifying sets of characters

The format of the @var{set1} and @var{set2} arguments resembles
the format of regular expressions; however, they are not regular
expressions, only lists of characters.  Most characters simply
represent themselves in these strings, but the strings can contain
the shorthands listed below, for convenience.  Some of them can be
used only in @var{set1} or @var{set2}, as noted below.

@table @asis

@item Backslash escapes.
@cindex backslash escapes

A backslash followed by a character not listed below causes an error
message.

@table @samp
@item \a
Control-G,
@item \b
Control-H,
@item \f
Control-L,
@item \n
Control-J,
@item \r
Control-M,
@item \t
Control-I,
@item \v
Control-K,
@item \@var{ooo}
The character with the value given by @var{ooo}, which is 1 to 3
octal digits,
@item \\
A backslash.
@end table

@item Ranges.
@cindex ranges

The notation @samp{@var{m}-@var{n}} expands to all of the characters
from @var{m} through @var{n}, in ascending order.  @var{m} should
collate before @var{n}; if it doesn't, an error results.  As an example,
@samp{0-9} is the same as @samp{0123456789}.  Although GNU @code{tr}
does not support the System V syntax that uses square brackets to
enclose ranges, translations specified in that format will still work as
long as the brackets in @var{string1} correspond to identical brackets
in @var{string2}.

@item Repeated characters.
@cindex repeated characters

The notation @samp{[@var{c}*@var{n}]} in @var{set2} expands to @var{n}
copies of character @var{c}.  Thus, @samp{[y*6]} is the same as
@samp{yyyyyy}.  The notation @samp{[@var{c}*]} in @var{string2} expands
to as many copies of @var{c} as are needed to make @var{set2} as long as
@var{set1}.  If @var{n} begins with @samp{0}, it is interpreted in
octal, otherwise in decimal.

@item Character classes.
@cindex characters classes

The notation @samp{[:@var{class}:]} expands to all of the characters in
the (predefined) class @var{class}.  The characters expand in no
particular order, except for the @code{upper} and @code{lower} classes,
which expand in ascending order.  When the @samp{--delete} (@samp{-d})
and @samp{--squeeze-repeats} (@samp{-s}) options are both given, any
character class can be used in @var{set2}.  Otherwise, only the
character classes @code{lower} and @code{upper} are accepted in
@var{set2}, and then only if the corresponding character class
(@code{upper} and @code{lower}, respectively) is specified in the same
relative position in @var{set1}.  Doing this specifies case conversion.
The class names are given below; an error results when an invalid class
name is given.

@table @code
@item alnum
@opindex alnum
Letters and digits.
@item alpha
@opindex alpha
Letters.
@item blank
@opindex blank
Horizontal whitespace.
@item cntrl
@opindex cntrl
Control characters.
@item digit
@opindex digit
Digits.
@item graph
@opindex graph
Printable characters, not including space.
@item lower
@opindex lower
Lowercase letters.
@item print
@opindex print
Printable characters, including space.
@item punct
@opindex punct
Punctuation characters.
@item space
@opindex space
Horizontal or vertical whitespace.
@item upper
@opindex upper
Uppercase letters.
@item xdigit
@opindex xdigit
Hexadecimal digits.
@end table

@item Equivalence classes.
@cindex equivalence classes

The syntax @samp{[=@var{c}=]} expands to all of the characters that are
equivalent to @var{c}, in no particular order.  Equivalence classes are
a relatively recent invention intended to support non-English alphabets.
But there seems to be no standard way to define them or determine their
contents.  Therefore, they are not fully implemented in GNU @code{tr};
each character's equivalence class consists only of that character,
which is of no particular use.

@end table


@node Translating
@subsection Translating

@cindex translating characters

@code{tr} performs translation when @var{set1} and @var{set2} are
both given and the @samp{--delete} (@samp{-d}) option is not given.
@code{tr} translates each character of its input that is in @var{set1}
to the corresponding character in @var{set2}.  Characters not in
@var{set1} are passed through unchanged.  When a character appears more
than once in @var{set1} and the corresponding characters in @var{set2}
are not all the same, only the final one is used.  For example, these
two commands are equivalent:

@example
tr aaa xyz
tr a z
@end example

A common use of @code{tr} is to convert lowercase characters to
uppercase.  This can be done in many ways.  Here are three of them:

@example
tr abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ
tr a-z A-Z
tr '[:lower:]' '[:upper:]'
@end example

When @code{tr} is performing translation, @var{set1} and @var{set2}
typically have the same length.  If @var{set1} is shorter than
@var{set2}, the extra characters at the end of @var{set2} are ignored.

On the other hand, making @var{set1} longer than @var{set2} is not
portable; POSIX.2 says that the result is undefined.  In this situation,
BSD @code{tr} pads @var{set2} to the length of @var{set1} by repeating
the last character of @var{set2} as many times as necessary.  System V
@code{tr} truncates @var{set1} to the length of @var{set2}.

By default, GNU @code{tr} handles this case like BSD @code{tr}.  When
the @samp{--truncate-set1} (@samp{-t}) option is given, GNU @code{tr}
handles this case like the System V @code{tr} instead.  This option is
ignored for operations other than translation.

Acting like System V @code{tr} in this case breaks the relatively common
BSD idiom:

@example
tr -cs A-Za-z0-9 '\012'
@end example

@noindent
because it converts only zero bytes (the first element in the
complement of @var{set1}), rather than all non-alphanumerics, to
newlines.


@node Squeezing
@subsection Squeezing repeats and deleting

@cindex squeezing repeat characters
@cindex deleting characters

When given just the @samp{--delete} (@samp{-d}) option, @code{tr}
removes any input characters that are in @var{set1}.

When given just the @samp{--squeeze-repeats} (@samp{-s}) option,
@code{tr} replaces each input sequence of a repeated character that
is in @var{set1} with a single occurrence of that character.

When given both @samp{--delete} and @samp{--squeeze-repeats}, @code{tr}
first performs any deletions using @var{set1}, then squeezes repeats
from any remaining characters using @var{set2}.

The @samp{--squeeze-repeats} option may also be used when translating,
in which case @code{tr} first performs translation, then squeezes
repeats from any remaining characters using @var{set2}.

Here are some examples to illustrate various combinations of options:

@itemize @bullet

@item
Remove all zero bytes:

@example
tr -d '\000'
@end example

@item
Put all words on lines by themselves.  This converts all
non-alphanumeric characters to newlines, then squeezes each string
of repeated newlines into a single newline:

@example
tr -cs '[a-zA-Z0-9]' '[\n*]'
@end example

@item
Convert each sequence of repeated newlines to a single newline:

@example
tr -s '\n'
@end example

@end itemize


@node Warnings in tr
@subsection Warning messages

@vindex POSIXLY_CORRECT
Setting the environment variable @code{POSIXLY_CORRECT} turns off the
following warning and error messages, for strict compliance with
POSIX.2.  Otherwise, the following diagnostics are issued:

@enumerate

@item
When the @samp{--delete} option is given but @samp{--squeeze-repeats}
is not, and @var{set2} is given, GNU @code{tr} by default prints
a usage message and exits, because @var{set2} would not be used.
The POSIX specification says that @var{set2} must be ignored in
this case. Silently ignoring arguments is a bad idea.

@item
When an ambiguous octal escape is given.  For example, @samp{\400}
is actually @samp{\40} followed by the digit @samp{0}, because the
value 400 octal does not fit into a single byte.

@end enumerate

GNU @code{tr} does not provide complete BSD or System V compatibility.
For example, it is impossible to disable interpretation of the POSIX
constructs @samp{[:alpha:]}, @samp{[=c=]}, and @samp{[c*10]}.  Also, GNU
@code{tr} does not delete zero bytes automatically, unlike traditional
Unix versions, which provide no way to preserve zero bytes.


@node expand invocation
@section @code{expand}: Convert tabs to spaces

@pindex expand
@cindex tabs to spaces, converting
@cindex converting tabs to spaces

@code{expand} writes the contents of each given @var{file}, or standard
input if none are given or for a @var{file} of @samp{-}, to standard
output, with tab characters converted to the appropriate number of
spaces.  Synopsis:

@example
expand [@var{option}]@dots{} [@var{file}]@dots{}
@end example

By default, @code{expand} converts all tabs to spaces.  It preserves
backspace characters in the output; they decrement the column count for
tab calculations.  The default action is equivalent to @samp{-8} (set
tabs every 8 columns).

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{tab1}[,@var{tab2}]@dots{}
@itemx -t @var{tab1}[,@var{tab2}]@dots{}
@itemx --tabs=@var{tab1}[,@var{tab2}]@dots{}
@opindex -@var{tab}
@opindex -t
@opindex --tabs
@cindex tabstops, setting
If only one tab stop is given, set the tabs @var{tab1} spaces apart
(default is 8).  Otherwise, set the tabs at columns @var{tab1},
@var{tab2}, @dots{} (numbered from 0), and replace any tabs beyond the
last tabstop given with single spaces.  If the tabstops are specified
with the @samp{-t} or @samp{--tabs} option, they can be separated by
blanks as well as by commas.

@item -i
@itemx --initial
@opindex -i
@opindex --initial
@cindex initial tabs, converting
Only convert initial tabs (those that precede all non-space or non-tab
characters) on each line to spaces.

@end table


@node unexpand invocation
@section @code{unexpand}: Convert spaces to tabs

@pindex unexpand

@code{unexpand} writes the contents of each given @var{file}, or
standard input if none are given or for a @var{file} of @samp{-}, to
standard output, with strings of two or more space or tab characters
converted to as many tabs as possible followed by as many spaces as are
needed.  Synopsis:

@example
unexpand [@var{option}]@dots{} [@var{file}]@dots{}
@end example

By default, @code{unexpand} converts only initial spaces and tabs (those
that precede all non space or tab characters) on each line.  It
preserves backspace characters in the output; they decrement the column
count for tab calculations.  By default, tabs are set at every 8th
column.

The program accepts the following options.  Also see @ref{Common options}.

@table @samp

@item -@var{tab1}[,@var{tab2}]@dots{}
@itemx -t @var{tab1}[,@var{tab2}]@dots{}
@itemx --tabs=@var{tab1}[,@var{tab2}]@dots{}
@opindex -@var{tab}
@opindex -t
@opindex --tabs
If only one tab stop is given, set the tabs @var{tab1} spaces apart
instead of the default 8.  Otherwise, set the tabs at columns
@var{tab1}, @var{tab2}, @dots{} (numbered from 0), and leave spaces and
tabs beyond the tabstops given unchanged.  If the tabstops are specified
with the @samp{-t} or @samp{--tabs} option, they can be separated by
blanks as well as by commas.  This option implies the @samp{-a} option.

@item -a
@itemx --all
@opindex -a
@opindex --all
Convert all strings of two or more spaces or tabs, not just initial
ones, to tabs.

@end table


@c              What's GNU?
@c              Arnold Robbins
@node Opening the software toolbox
@chapter Opening the software toolbox

This chapter originally appeared in @cite{Linux Journal}, volume 1,
number 2, in the @cite{What's GNU?} column. It was written by Arnold
Robbins.

@menu
* Toolbox introduction::
* I/O redirection::
* The @code{who} command::
* The @code{cut} command::
* The @code{sort} command::
* The @code{uniq} command::
* Putting the tools together::
@end menu


@node Toolbox introduction
@unnumberedsec Toolbox introduction

This month's column is only peripherally related to the GNU Project, in
that it describes a number of the GNU tools on your Linux system and how they
might be used.  What it's really about is the ``Software Tools'' philosophy
of program development and usage.

The software tools philosophy was an important and integral concept
in the initial design and development of Unix (of which Linux and GNU are
essentially clones).  Unfortunately, in the modern day press of
Internetworking and flashy GUIs, it seems to have fallen by the
wayside.  This is a shame, since it provides a powerful mental model
for solving many kinds of problems.

Many people carry a Swiss Army knife around in their pants pockets (or
purse).  A Swiss Army knife is a handy tool to have: it has several knife
blades, a screwdriver, tweezers, toothpick, nail file, corkscrew, and perhaps
a number of other things on it.  For the everyday, small miscellaneous jobs
where you need a simple, general purpose tool, it's just the thing.

On the other hand, an experienced carpenter doesn't build a house using
a Swiss Army knife.  Instead, he has a toolbox chock full of specialized
tools---a saw, a hammer, a screwdriver, a plane, and so on.  And he knows
exactly when and where to use each tool; you won't catch him hammering nails
with the handle of his screwdriver.

The Unix developers at Bell Labs were all professional programmers and trained
computer scientists.  They had found that while a one-size-fits-all program
might appeal to a user because there's only one program to use, in practice
such programs are

@enumerate a
@item
difficult to write,

@item
difficult to maintain and
debug, and

@item
difficult to extend to meet new situations.
@end enumerate

Instead, they felt that programs should be specialized tools.  In short, each
program ``should do one thing well.''  No more and no less.  Such programs are
simpler to design, write, and get right---they only do one thing.

Furthermore, they found that with the right machinery for hooking programs
together, that the whole was greater than the sum of the parts.  By combining
several special purpose programs, you could accomplish a specific task
that none of the programs was designed for, and accomplish it much more
quickly and easily than if you had to write a special purpose program.
We will see some (classic) examples of this further on in the column.
(An important additional point was that, if necessary, take a detour
and build any software tools you may need first, if you don't already
have something appropriate in the toolbox.)

@node I/O redirection
@unnumberedsec I/O redirection

Hopefully, you are familiar with the basics of I/O redirection in the
shell, in particular the concepts of ``standard input,'' ``standard output,''
and ``standard error''.  Briefly, ``standard input'' is a data source, where
data comes from.  A program should not need to either know or care if the
data source is a disk file, a keyboard, a magnetic tape, or even a punched
card reader.  Similarly, ``standard output'' is a data sink, where data goes
to.  The program should neither know nor care where this might be.
Programs that only read their standard input, do something to the data,
and then send it on, are called ``filters'', by analogy to filters in a
water pipeline.

With the Unix shell, it's very easy to set up data pipelines:

@example
program_to_create_data | filter1 | .... | filterN > final.pretty.data
@end example

We start out by creating the raw data; each filter applies some successive
transformation to the data, until by the time it comes out of the pipeline,
it is in the desired form.

This is fine and good for standard input and standard output.  Where does the
standard error come in to play?  Well, think about @code{filter1} in
the pipeline above.  What happens if it encounters an error in the data it
sees?  If it writes an error message to standard output, it will just
disappear down the pipeline into @code{filter2}'s input, and the
user will probably never see it.  So programs need a place where they can send
error messages so that the user will notice them.  This is standard error,
and it is usually connected to your console or window, even if you have
redirected standard output of your program away from your screen.

For filter programs to work together, the format of the data has to be
agreed upon.  The most straightforward and easiest format to use is simply
lines of text.  Unix data files are generally just streams of bytes, with
lines delimited by the @sc{ASCII} @sc{LF} (Line Feed) character,
conventionally called a ``newline'' in the Unix literature. (This is
@code{'\n'} if you're a C programmer.)  This is the format used by all
the traditional filtering programs.  (Many earlier operating systems
had elaborate facilities and special purpose programs for managing
binary data.  Unix has always shied away from such things, under the
philosophy that it's easiest to simply be able to view and edit your
data with a text editor.)

OK, enough introduction. Let's take a look at some of the tools, and then
we'll see how to hook them together in interesting ways.   In the following
discussion, we will only present those command line options that interest
us.  As you should always do, double check your system documentation
for the full story.

@node The @code{who} command
@unnumberedsec The @code{who} command

The first program is the @code{who} command.  By itself, it generates a
list of the users who are currently logged in.  Although I'm writing
this on a single-user system, we'll pretend that several people are
logged in:

@example
$ who
arnold   console Jan 22 19:57
miriam   ttyp0   Jan 23 14:19(:0.0)
bill     ttyp1   Jan 21 09:32(:0.0)
arnold   ttyp2   Jan 23 20:48(:0.0)
@end example

Here, the @samp{$} is the usual shell prompt, at which I typed @code{who}.
There are three people logged in, and I am logged in twice.  On traditional
Unix systems, user names are never more than eight characters long.  This
little bit of trivia will be useful later.  The output of @code{who} is nice,
but the data is not all that exciting.

@node The @code{cut} command
@unnumberedsec The @code{cut} command

The next program we'll look at is the @code{cut} command.  This program
cuts out columns or fields of input data.  For example, we can tell it
to print just the login name and full name from the @file{/etc/passwd
file}.  The @file{/etc/passwd} file has seven fields, separated by
colons:

@example
arnold:xyzzy:2076:10:Arnold D. Robbins:/home/arnold:/bin/ksh
@end example

To get the first and fifth fields, we would use cut like this:

@example
$ cut -d: -f1,5 /etc/passwd
root:Operator
@dots{}
arnold:Arnold D. Robbins
miriam:Miriam A. Robbins
@dots{}
@end example

With the @samp{-c} option, @code{cut} will cut out specific characters
(i.e., columns) in the input lines.  This command looks like it might be
useful for data filtering.


@node The @code{sort} command
@unnumberedsec The @code{sort} command

Next we'll look at the @code{sort} command.  This is one of the most
powerful commands on a Unix-style system; one that you will often find
yourself using when setting up fancy data plumbing. The @code{sort}
command reads and sorts each file named on the command line.  It then
merges the sorted data and writes it to standard output.  It will read
standard input if no files are given on the command line (thus
making it into a filter).  The sort is based on the machine collating
sequence (@sc{ASCII}) or based on  user-supplied ordering criteria.


@node The @code{uniq} command
@unnumberedsec The @code{uniq} command

Finally (at least for now), we'll look at the @code{uniq} program.  When
sorting data, you will often end up with duplicate lines, lines that
are identical.  Usually, all you need is one instance of each line.
This is where @code{uniq} comes in. The @code{uniq} program reads its
standard input, which it expects to be sorted.  It only prints out one
copy of each duplicated line.  It does have several options.  Later on,
we'll use the @samp{-c} option, which prints each unique line, preceded
by a count of the number of times that line occurred in the input.


@node Putting the tools together
@unnumberedsec Putting the tools together

Now, let's suppose this is a large BBS system with dozens of users
logged in.  The management wants the SysOp to write a program that will
generate a sorted list of logged in users.  Furthermore, even if a user
is logged in multiple times, his or her name should only show up in the
output once.

The SysOp could sit down with the system documentation and write a C
program that did this. It would take perhaps a couple of hundred lines
of code and about two hours to write it, test it, and debug it.
However, knowing the software toolbox, the SysOp can instead start out
by generating just a list of logged on users:

@example
$ who | cut -c1-8
arnold
miriam
bill
arnold
@end example

Next, sort the list:

@example
$ who | cut -c1-8 | sort
arnold
arnold
bill
miriam
@end example

Finally, run the sorted list through @code{uniq}, to weed out duplicates:

@example
$ who | cut -c1-8 | sort | uniq
arnold
bill
miriam
@end example

The @code{sort} command actually has a @samp{-u} option that does what
@code{uniq} does. However, @code{uniq} has other uses for which one
cannot substitute @samp{sort -u}.

The SysOp puts this pipeline into a shell script, and makes it available for
all the users on the system:

@example
# cat > /usr/local/bin/listusers
who | cut -c1-8 | sort | uniq
^D
# chmod +x /usr/local/bin/listusers
@end example

There are four major points to note here.  First, with just four
programs, on one command line, the SysOp was able to save about two
hours worth of work.  Furthermore, the shell pipeline is just about as
efficient as the C program would be, and it is much more efficient in
terms of programmer time.  People time is much more expensive than
computer time, and in our modern ``there's never enough time to do
everything'' society, saving two hours of programmer time is no mean
feat.

Second, it is also important to emphasize that with the
@emph{combination} of the tools, it is possible to do a special
purpose job never imagined by the authors of the individual programs.

Third, it is also valuable to build up your pipeline in stages, as we did here.
This allows you to view the data at each stage in the pipeline, which helps
you acquire the confidence that you are indeed using these tools correctly.

Finally, by bundling the pipeline in a shell script, other users can use
your command, without having to remember the fancy plumbing you set up for
them. In terms of how you run them, shell scripts and compiled programs are
indistinguishable.

After the previous warm-up exercise, we'll look at two additional, more
complicated pipelines.  For them, we need to introduce two more tools.

The first is the @code{tr} command, which stands for ``transliterate.''
The @code{tr} command works on a character-by-character basis, changing
characters. Normally it is used for things like mapping upper case to
lower case:

@example
$ echo ThIs ExAmPlE HaS MIXED case! | tr '[A-Z]' '[a-z]'
this example has mixed case!
@end example

There are several options of interest:

@table @samp
@item -c
work on the complement of the listed characters, i.e.,
operations apply to characters not in the given set

@item -d
delete characters in the first set from the output

@item -s
squeeze repeated characters in the output into just one character.
@end table

We will be using all three options in a moment.

The other command we'll look at is @code{comm}.  The @code{comm}
command takes two sorted input files as input data, and prints out the
files' lines in three columns.  The output columns are the data lines
unique to the first file, the data lines unique to the second file, and
the data lines that are common to both.  The @samp{-1}, @samp{-2}, and
@samp{-3} command line options omit the respective columns. (This is
non-intuitive and takes a little getting used to.)  For example:

@example
$ cat f1
11111
22222
33333
44444
$ cat f2
00000
22222
33333
55555
$ comm f1 f2
        00000
11111
                22222
                33333
44444
        55555
@end example

The single dash as a filename tells @code{comm} to read standard input
instead of a regular file.

Now we're ready to build a fancy pipeline.  The first application is a word
frequency counter.  This helps an author determine if he or she is over-using
certain words.

The first step is to change the case of all the letters in our input file
to one case.  ``The'' and ``the'' are the same word when doing counting.

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | ...
@end example

The next step is to get rid of punctuation.  Quoted words and unquoted words
should be treated identically; it's easiest to just get the punctuation out of
the way.

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' | ...
@end example

The second @code{tr} command operates on the complement of the listed
characters, which are all the letters, the digits, the underscore, and
the blank.  The @samp{\012} represents the newline character; it has to
be left alone.  (The ASCII TAB character should also be included for
good measure in a production script.)

At this point, we have data consisting of words separated by blank space.
The words only contain alphanumeric characters (and the underscore).  The
next step is break the data apart so that we have one word per line. This
makes the counting operation much easier, as we will see shortly.

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' |
> tr -s '[ ]' '\012' | ...
@end example

This command turns blanks into newlines.  The @samp{-s} option squeezes
multiple newline characters in the output into just one.  This helps us
avoid blank lines. (The @samp{>} is the shell's ``secondary prompt.''
This is what the shell prints when it notices you haven't finished
typing in all of a command.)

We now have data consisting of one word per line, no punctuation, all one
case.  We're ready to count each word:

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' |
> tr -s '[ ]' '\012' | sort | uniq -c | ...
@end example

At this point, the data might look something like this:

@example
  60 a
   2 able
   6 about
   1 above
   2 accomplish
   1 acquire
   1 actually
   2 additional
@end example

The output is sorted by word, not by count!  What we want is the most
frequently used words first.  Fortunately, this is easy to accomplish,
with the help of two more @code{sort} options:

@table @samp
@item -n
do a numeric sort, not an ASCII one

@item -r
reverse the order of the sort
@end table

The final pipeline looks like this:

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' |
> tr -s '[ ]' '\012' | sort | uniq -c | sort -nr
 156 the
  60 a
  58 to
  51 of
  51 and
 ...
@end example

Whew!  That's a lot to digest.  Yet, the same principles apply. With six
commands, on two lines (really one long one split for convenience), we've
created a program that does something interesting and useful, in much
less time than we could have written a C program to do the same thing.

A minor modification to the above pipeline can give us a simple spelling
checker!  To determine if you've spelled a word correctly, all you have to
do is look it up in a dictionary.  If it is not there, then chances are
that your spelling is incorrect.  So, we need a dictionary.  If you
have the Slackware Linux distribution, you have the file
@file{/usr/lib/ispell/ispell.words}, which is a sorted, 38,400 word
dictionary.

Now, how to compare our file with the dictionary?  As before, we generate
a sorted list of words, one per line:

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' |
> tr -s '[ ]' '\012' | sort -u | ...
@end example

Now, all we need is a list of words that are @emph{not} in the
dictionary.  Here is where the @code{comm} command comes in.

@example
$ tr '[A-Z]' '[a-z]' < whats.gnu | tr -cd '[A-Za-z0-9_ \012]' |
> tr -s '[ ]' '\012' | sort -u |
> comm -23 - /usr/lib/ispell/ispell.words
@end example

The @samp{-2} and @samp{-3} options eliminate lines that are only in the
dictionary (the second file), and lines that are in both files.  Lines
only in the first file (standard input, our stream of words), are
words that are not in the dictionary.  These are likely candidates for
spelling errors.  This pipeline was the first cut at a production
spelling checker on Unix.

There are some other tools that deserve brief mention.

@table @code
@item grep
search files for text that matches a regular expression

@item egrep
like @code{grep}, but with more powerful regular expressions

@item wc
count lines, words, characters

@item tee
a T-fitting for data pipes, copies data to files and to standard output

@item sed
the stream editor, an advanced tool

@item awk
a data manipulation language, another advanced tool
@end table

The software tools philosophy also espoused the following bit of
advice: ``Let someone else do the hard part.'' This means, take
something that gives you most of what you need, and then massage it the
rest of the way until it's in the form that you want.

To summarize:

@enumerate 1
@item
Each program should do one thing well. No more, no less.

@item
Combining programs with appropriate plumbing leads to results where
the whole is greater than the sum of the parts.  It also leads to novel
uses of programs that the authors might never have imagined.

@item
Programs should never print extraneous header or trailer data, since these
could get sent on down a pipeline. (A point we didn't mention earlier.)

@item
Let someone else do the hard part.

@item
Know your toolbox! Use each program appropriately. If you don't have an
appropriate tool, build one.
@end enumerate

As of this writing, all the programs we've discussed are available via
anonymous @code{ftp} from @code{prep.ai.mit.edu} as
@file{/pub/gnu/textutils-1.9.tar.gz} directory.@footnote{Version 1.9 was
current when this column was written. Check the nearest GNU archive for
the current version.}

None of what I have presented in this column is new. The Software Tools
philosophy was first introduced in the book @cite{Software Tools},
by Brian Kernighan and P.J. Plauger (Addison-Wesley, ISBN
0-201-03669-X).   This book showed how to write and use software
tools.   It was written in 1976, using a preprocessor for FORTRAN named
@code{ratfor} (RATional FORtran).  At the time, C was not as ubiquitous
as it is now; FORTRAN was.  The last chapter presented a @code{ratfor}
to FORTRAN processor, written in @code{ratfor}. @code{ratfor} looks an
awful lot like C; if you know C, you won't have any problem following
the code.

In 1981, the book was updated and made available as @cite{Software
Tools in Pascal} (Addison-Wesley, ISBN 0-201-10342-7).  Both books
remain in print, and are well worth reading if you're a programmer.
They certainly made a major change in how I view programming.

Initially, the programs in both books were available (on 9-track tape)
from Addison-Wesley.  Unfortunately, this is no longer the case,
although you might be able to find copies floating around the Internet.
For a number of years, there was an active Software Tools Users Group,
whose members had ported the original @code{ratfor} programs to essentially
every computer system with a FORTRAN compiler.  The popularity of the
group waned in the middle '80s as Unix began to spread beyond universities.

With the current proliferation of GNU code and other clones of Unix programs,
these programs now receive little attention; modern C versions are
much more efficient and do more than these programs do.  Nevertheless, as
exposition of good programming style, and evangelism for a still-valuable
philosophy, these books are unparalleled, and I recommend them highly.

Acknowledgement: I would like to express my gratitude to Brian Kernighan
of Bell Labs, the original Software Toolsmith, for reviewing this column.


@node Index
@unnumbered Index

@printindex cp

@contents
@bye

@c Local variables:
@c texinfo-column-for-description: 32
@c End: