\input texinfo @c -*-texinfo-*-
@c
@c -- Stuff that needs adding: ----------------------------------------------
@c (nothing!)
@c --------------------------------------------------------------------------
@c Check for consistency: regexps in @code, text that they match in @samp.
@c
@c Tips:
@c @command for command
@c @samp for command fragments: @samp{cat -s}
@c @code for sed commands and flags
@c Use ``quote'' not `quote' or "quote".
@c
@c %**start of header
@setfilename sed.info
@settitle sed, a stream editor
@c %**end of header
@c @smallbook
@include version.texi
@c Combine indices.
@syncodeindex ky cp
@syncodeindex pg cp
@syncodeindex tp cp
@defcodeindex op
@syncodeindex op fn
@include config.texi
@copying
This file documents version @value{VERSION} of
@value{SSED}, a stream editor.
Copyright @copyright{} 1998--2022 Free Software Foundation, Inc.
@quotation
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3
or any later version published by the Free Software Foundation;
with no Invariant Sections, no Front-Cover Texts, and no
Back-Cover Texts. A copy of the license is included in the
section entitled ``GNU Free Documentation License''.
@end quotation
@end copying
@setchapternewpage off
@titlepage
@title @value{SSED}, a stream editor
@subtitle version @value{VERSION}, @value{UPDATED}
@author by Ken Pizzini, Paolo Bonzini, Jim Meyering, Assaf Gordon
@page
@vskip 0pt plus 1filll
@insertcopying
@end titlepage
@contents
@ifnottex
@node Top
@top @value{SSED}
@insertcopying
@end ifnottex
@menu
* Introduction:: Introduction
* Invoking sed:: Invocation
* sed scripts:: @command{sed} scripts
* sed addresses:: Addresses: selecting lines
* sed regular expressions:: Regular expressions: selecting text
* advanced sed:: Advanced @command{sed}: cycles and buffers
* Examples:: Some sample scripts
* Limitations:: Limitations and (non-)limitations of @value{SSED}
* Other Resources:: Other resources for learning about @command{sed}
* Reporting Bugs:: Reporting bugs
* GNU Free Documentation License:: Copying and sharing this manual
* Concept Index:: A menu with all the topics in this manual.
* Command and Option Index:: A menu with all @command{sed} commands and
command-line options.
@end menu
@node Introduction
@chapter Introduction
@cindex Stream editor
@command{sed} is a stream editor.
A stream editor is used to perform basic text
transformations on an input stream
(a file or input from a pipeline).
While in some ways similar to an editor which
permits scripted edits (such as @command{ed}),
@command{sed} works by making only one pass over the
input(s), and is consequently more efficient.
But it is @command{sed}'s ability to filter text in a pipeline
which particularly distinguishes it from other types of
editors.
@node Invoking sed
@chapter Running sed
This chapter covers how to run @command{sed}. Details of @command{sed}
scripts and individual @command{sed} commands are discussed in the
next chapter.
@menu
* Overview::
* Command-Line Options::
* Exit status::
@end menu
@node Overview
@section Overview
Normally @command{sed} is invoked like this:
@example
sed SCRIPT INPUTFILE...
@end example
For example, to change every @samp{hello} to @samp{world}
in the file @file{input.txt}:
@example
sed 's/hello/world/g' input.txt > output.txt
@end example
Without the @samp{g} (global) modifier, @command{sed} affects
only the first instance per line.
@cindex stdin
@cindex standard input
If you do not specify @var{INPUTFILE}, or if @var{INPUTFILE} is @file{-},
@command{sed} filters the contents of the standard input. The following
commands are equivalent:
@example
sed 's/hello/world/g' input.txt > output.txt
sed 's/hello/world/g' < input.txt > output.txt
cat input.txt | sed 's/hello/world/g' - > output.txt
@end example
@cindex stdout
@cindex output
@cindex standard output
@cindex -i, example
@command{sed} writes output to standard output. Use @option{-i} to edit
files in-place instead of printing to standard output.
See also the @code{W} and @code{s///w} commands for writing output to
other files. The following command modifies @file{file.txt} and
does not produce any output:
@example
sed -i 's/hello/world/' file.txt
@end example
@cindex -n, example
@cindex p, example
@cindex suppressing output
@cindex output, suppressing
By default @command{sed} prints all processed input (except input
that has been modified/deleted by commands such as @command{d}).
Use @option{-n} to suppress output, and the @code{p} command
to print specific lines. The following command prints only line 45
of the input file:
@example
sed -n '45p' file.txt
@end example
@cindex multiple files
@cindex -s, example
@command{sed} treats multiple input files as one long stream.
The following example prints the first line of the first file
(@file{one.txt}) and the last line of the last file (@file{three.txt}).
Use @option{-s} to reverse this behavior.
@example
sed -n '1p ; $p' one.txt two.txt three.txt
@end example
@cindex -e, example
@cindex --expression, example
@cindex -f, example
@cindex --file, example
@cindex script parameter
@cindex parameters, script
Without @option{-e} or @option{-f} options, @command{sed} uses
the first non-option parameter as the @var{script}, and the following
non-option parameters as input files.
If @option{-e} or @option{-f} options are used to specify a @var{script},
all non-option parameters are taken as input files.
Options @option{-e} and @option{-f} can be combined, and can appear
multiple times (in which case the final effective @var{script} will be
concatenation of all the individual @var{script}s).
The following examples are equivalent:
@example
sed 's/hello/world/' input.txt > output.txt
sed -e 's/hello/world/' input.txt > output.txt
sed --expression='s/hello/world/' input.txt > output.txt
echo 's/hello/world/' > myscript.sed
sed -f myscript.sed input.txt > output.txt
sed --file=myscript.sed input.txt > output.txt
@end example
@node Command-Line Options
@section Command-Line Options
The full format for invoking @command{sed} is:
@example
sed OPTIONS... [SCRIPT] [INPUTFILE...]
@end example
@command{sed} may be invoked with the following command-line options:
@table @code
@item --version
@opindex --version
@cindex Version, printing
Print out the version of @command{sed} that is being run and a copyright notice,
then exit.
@item --help
@opindex --help
@cindex Usage summary, printing
Print a usage message briefly summarizing these command-line options
and the bug-reporting address,
then exit.
@item -n
@itemx --quiet
@itemx --silent
@opindex -n
@opindex --quiet
@opindex --silent
@cindex Disabling autoprint, from command line
By default, @command{sed} prints out the pattern space
at the end of each cycle through the script (@pxref{Execution Cycle, ,
How @code{sed} works}).
These options disable this automatic printing,
and @command{sed} only produces output when explicitly told to
via the @code{p} command.
@item --debug
@opindex --debug
@cindex @value{SSEDEXT}, debug
Print the input sed program in canonical form,
and annotate program execution.
@codequotebacktick on
@codequoteundirected on
@example
$ echo 1 | sed '\%1%s21232'
3
$ echo 1 | sed --debug '\%1%s21232'
SED PROGRAM:
/1/ s/1/3/
INPUT: 'STDIN' line 1
PATTERN: 1
COMMAND: /1/ s/1/3/
PATTERN: 3
END-OF-CYCLE:
3
@end example
@codequotebacktick off
@codequoteundirected off
@item -e @var{script}
@itemx --expression=@var{script}
@opindex -e
@opindex --expression
@cindex Script, from command line
Add the commands in @var{script} to the set of commands to be
run while processing the input.
@item -f @var{script-file}
@itemx --file=@var{script-file}
@opindex -f
@opindex --file
@cindex Script, from a file
Add the commands contained in the file @var{script-file}
to the set of commands to be run while processing the input.
@item -i[@var{SUFFIX}]
@itemx --in-place[=@var{SUFFIX}]
@opindex -i
@opindex --in-place
@cindex In-place editing, activating
@cindex @value{SSEDEXT}, in-place editing
This option specifies that files are to be edited in-place.
@value{SSED} does this by creating a temporary file and
sending output to this file rather than to the standard
output.@footnote{This applies to commands such as @code{=},
@code{a}, @code{c}, @code{i}, @code{l}, @code{p}. You can
still write to the standard output by using the @code{w}
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
or @code{W} commands together with the @file{/dev/stdout}
special file}.
This option implies @option{-s}.
When the end of the file is reached, the temporary file is
renamed to the output file's original name. The extension,
if supplied, is used to modify the name of the old file
before renaming the temporary file, thereby making a backup
copy@footnote{Note that @value{SSED} creates the backup
file whether or not any output is actually changed.}).
@cindex In-place editing, Perl-style backup file names
This rule is followed: if the extension doesn't contain a @code{*},
then it is appended to the end of the current filename as a
suffix; if the extension does contain one or more @code{*}
characters, then @emph{each} asterisk is replaced with the
current filename. This allows you to add a prefix to the
backup file, instead of (or in addition to) a suffix, or
even to place backup copies of the original files into another
directory (provided the directory already exists).
If no extension is supplied, the original file is
overwritten without making a backup.
Because @option{-i} takes an optional argument, it should
not be followed by other short options:
@table @code
@item sed -Ei '...' FILE
Same as @option{-E -i} with no backup suffix - @file{FILE} will be
edited in-place without creating a backup.
@item sed -iE '...' FILE
This is equivalent to @option{--in-place=E}, creating @file{FILEE} as backup
of @file{FILE}
@end table
Be cautious of using @option{-n} with @option{-i}: the former disables
automatic printing of lines and the latter changes the file in-place
without a backup. Used carelessly (and without an explicit @code{p} command),
the output file will be empty:
@codequotebacktick on
@codequoteundirected on
@example
# WRONG USAGE: 'FILE' will be truncated.
sed -ni 's/foo/bar/' FILE
@end example
@codequotebacktick off
@codequoteundirected off
@item -l @var{N}
@itemx --line-length=@var{N}
@opindex -l
@opindex --line-length
@cindex Line length, setting
Specify the default line-wrap length for the @code{l} command.
A length of 0 (zero) means to never wrap long lines. If
not specified, it is taken to be 70.
@item --posix
@opindex --posix
@cindex @value{SSEDEXT}, disabling
@value{SSED} includes several extensions to POSIX
sed. In order to simplify writing portable scripts, this
option disables all the extensions that this manual documents,
including additional commands.
@cindex @code{POSIXLY_CORRECT} behavior, enabling
Most of the extensions accept @command{sed} programs that
are outside the syntax mandated by POSIX, but some
of them (such as the behavior of the @command{N} command
described in @ref{Reporting Bugs}) actually violate the
standard. If you want to disable only the latter kind of
extension, you can set the @code{POSIXLY_CORRECT} variable
to a non-empty value.
@item -b
@itemx --binary
@opindex -b
@opindex --binary
This option is available on every platform, but is only effective where the
operating system makes a distinction between text files and binary files.
When such a distinction is made---as is the case for MS-DOS, Windows,
Cygwin---text files are composed of lines separated by a carriage return
@emph{and} a line feed character, and @command{sed} does not see the
ending CR. When this option is specified, @command{sed} will open
input files in binary mode, thus not requesting this special processing
and considering lines to end at a line feed.
@item --follow-symlinks
@opindex --follow-symlinks
This option is available only on platforms that support
symbolic links and has an effect only if option @option{-i}
is specified. In this case, if the file that is specified
on the command line is a symbolic link, @command{sed} will
follow the link and edit the ultimate destination of the
link. The default behavior is to break the symbolic link,
so that the link destination will not be modified.
@item -E
@itemx -r
@itemx --regexp-extended
@opindex -E
@opindex -r
@opindex --regexp-extended
@cindex Extended regular expressions, choosing
@cindex GNU extensions, extended regular expressions
Use extended regular expressions rather than basic
regular expressions. Extended regexps are those that
@command{egrep} accepts; they can be clearer because they
usually have fewer backslashes.
Historically this was a GNU extension,
but the @option{-E}
extension has since been added to the POSIX standard
(http://austingroupbugs.net/view.php?id=528),
so use @option{-E} for portability.
GNU sed has accepted @option{-E} as an undocumented option for years,
and *BSD seds have accepted @option{-E} for years as well,
but scripts that use @option{-E} might not port to other older systems.
@xref{ERE syntax, , Extended regular expressions}.
@item -s
@itemx --separate
@opindex -s
@opindex --separate
@cindex Working on separate files
By default, @command{sed} will consider the files specified on the
command line as a single continuous long stream. This @value{SSED}
extension allows the user to consider them as separate files:
range addresses (such as @samp{/abc/,/def/}) are not allowed
to span several files, line numbers are relative to the start
of each file, @code{$} refers to the last line of each file,
and files invoked from the @code{R} commands are rewound at the
start of each file.
@item --sandbox
@opindex --sandbox
@cindex Sandbox mode
In sandbox mode, @code{e/w/r} commands are rejected - programs containing
them will be aborted without being run. Sandbox mode ensures @command{sed}
operates only on the input files designated on the command line, and
cannot run external programs.
@item -u
@itemx --unbuffered
@opindex -u
@opindex --unbuffered
@cindex Unbuffered I/O, choosing
Buffer both input and output as minimally as practical.
(This is particularly useful if the input is coming from
the likes of @samp{tail -f}, and you wish to see the transformed
output as soon as possible.)
@item -z
@itemx --null-data
@itemx --zero-terminated
@opindex -z
@opindex --null-data
@opindex --zero-terminated
Treat the input as a set of lines, each terminated by a zero byte
(the ASCII @samp{NUL} character) instead of a newline. This option can
be used with commands like @samp{sort -z} and @samp{find -print0}
to process arbitrary file names.
@end table
If no @option{-e}, @option{-f}, @option{--expression}, or @option{--file}
options are given on the command-line,
then the first non-option argument on the command line is
taken to be the @var{script} to be executed.
@cindex Files to be processed as input
If any command-line parameters remain after processing the above,
these parameters are interpreted as the names of input files to
be processed.
@cindex Standard input, processing as input
A file name of @samp{-} refers to the standard input stream.
The standard input will be processed if no file names are specified.
@node Exit status
@section Exit status
@cindex exit status
An exit status of zero indicates success, and a nonzero value
indicates failure. @value{SSED} returns the following exit status
error values:
@table @asis
@item 0
Successful completion.
@item 1
Invalid command, invalid syntax, invalid regular expression or a
@value{SSED} extension command used with @option{--posix}.
@item 2
One or more of the input file specified on the command line could not be
opened (e.g. if a file is not found, or read permission is denied).
Processing continued with other files.
@item 4
An I/O error, or a serious processing error during runtime,
@value{SSED} aborted immediately.
@end table
@cindex Q, example
@cindex exit status, example
Additionally, the commands @code{q} and @code{Q} can be used to terminate
@command{sed} with a custom exit code value (this is a @value{SSED} extension):
@example
$ echo | sed 'Q42' ; echo $?
42
@end example
@node sed scripts
@chapter @command{sed} scripts
@menu
* sed script overview:: @command{sed} script overview
* sed commands list:: @command{sed} commands summary
* The "s" Command:: @command{sed}'s Swiss Army Knife
* Common Commands:: Often used commands
* Other Commands:: Less frequently used commands
* Programming Commands:: Commands for @command{sed} gurus
* Extended Commands:: Commands specific of @value{SSED}
* Multiple commands syntax:: Extension for easier scripting
@end menu
@node sed script overview
@section @command{sed} script overview
@cindex @command{sed} script structure
@cindex Script structure
A @command{sed} program consists of one or more @command{sed} commands,
passed in by one or more of the
@option{-e}, @option{-f}, @option{--expression}, and @option{--file}
options, or the first non-option argument if zero of these
options are used.
This document will refer to ``the'' @command{sed} script;
this is understood to mean the in-order concatenation
of all of the @var{script}s and @var{script-file}s passed in.
@xref{Overview}.
@cindex @command{sed} commands syntax
@cindex syntax, @command{sed} commands
@cindex addresses, syntax
@cindex syntax, addresses
@command{sed} commands follow this syntax:
@example
[addr]@var{X}[options]
@end example
@var{X} is a single-letter @command{sed} command.
@c TODO: add @pxref{commands} when there is a command-list section.
@code{[addr]} is an optional line address. If @code{[addr]} is specified,
the command @var{X} will be executed only on the matched lines.
@code{[addr]} can be a single line number, a regular expression,
or a range of lines (@pxref{sed addresses}).
Additional @code{[options]} are used for some @command{sed} commands.
@cindex @command{d}, example
@cindex address range, example
@cindex example, address range
The following example deletes lines 30 to 35 in the input.
@code{30,35} is an address range. @command{d} is the delete command:
@example
sed '30,35d' input.txt > output.txt
@end example
@cindex @command{q}, example
@cindex regular expression, example
@cindex example, regular expression
The following example prints all input until a line
starting with the string @samp{foo} is found. If such line is found,
@command{sed} will terminate with exit status 42.
If such line was not found (and no other error occurred), @command{sed}
will exit with status 0.
@code{/^foo/} is a regular-expression address.
@command{q} is the quit command. @code{42} is the command option.
@example
sed '/^foo/q42' input.txt > output.txt
@end example
@cindex multiple @command{sed} commands
@cindex @command{sed} commands, multiple
@cindex newline, command separator
@cindex semicolons, command separator
@cindex ;, command separator
@cindex -e, example
@cindex -f, example
Commands within a @var{script} or @var{script-file} can be
separated by semicolons (@code{;}) or newlines (ASCII 10).
Multiple scripts can be specified with @option{-e} or @option{-f}
options.
The following examples are all equivalent. They perform two @command{sed}
operations: deleting any lines matching the regular expression @code{/^foo/},
and replacing all occurrences of the string @samp{hello} with @samp{world}:
@example
sed '/^foo/d ; s/hello/world/g' input.txt > output.txt
sed -e '/^foo/d' -e 's/hello/world/g' input.txt > output.txt
echo '/^foo/d' > script.sed
echo 's/hello/world/g' >> script.sed
sed -f script.sed input.txt > output.txt
echo 's/hello/world/g' > script2.sed
sed -e '/^foo/d' -f script2.sed input.txt > output.txt
@end example
@cindex @command{a}, and semicolons
@cindex @command{c}, and semicolons
@cindex @command{i}, and semicolons
Commands @command{a}, @command{c}, @command{i}, due to their syntax,
cannot be followed by semicolons working as command separators and
thus should be terminated
with newlines or be placed at the end of a @var{script} or @var{script-file}.
Commands can also be preceded with optional non-significant
whitespace characters.
@xref{Multiple commands syntax}.
@node sed commands list
@section @command{sed} commands summary
The following commands are supported in @value{SSED}.
Some are standard POSIX commands, while other are @value{SSEDEXT}.
Details and examples for each command are in the following sections.
(Mnemonics) are shown in parentheses.
@table @code
@item a\
@itemx @var{text}
Append @var{text} after a line.
@item a @var{text}
Append @var{text} after a line (alternative syntax).
@item b @var{label}
Branch unconditionally to @var{label}.
The @var{label} may be omitted, in which case the next cycle is started.
@item c\
@itemx @var{text}
Replace (change) lines with @var{text}.
@item c @var{text}
Replace (change) lines with @var{text} (alternative syntax).
@item d
Delete the pattern space;
immediately start next cycle.
@item D
If pattern space contains newlines, delete text in the pattern
space up to the first newline, and restart cycle with the resultant
pattern space, without reading a new line of input.
If pattern space contains no newline, start a normal new cycle as if
the @code{d} command was issued.
@c TODO: add a section about D+N and D+n commands
@item e
Executes the command that is found in pattern space and
replaces the pattern space with the output; a trailing newline
is suppressed.
@item e @var{command}
Executes @var{command} and sends its output to the output stream.
The command can run across multiple lines, all but the last ending with
a back-slash.
@item F
(filename) Print the file name of the current input file (with a trailing
newline).
@item g
Replace the contents of the pattern space with the contents of the hold space.
@item G
Append a newline to the contents of the pattern space,
and then append the contents of the hold space to that of the pattern space.
@item h
(hold) Replace the contents of the hold space with the contents of the
pattern space.
@item H
Append a newline to the contents of the hold space,
and then append the contents of the pattern space to that of the hold space.
@item i\
@itemx @var{text}
insert @var{text} before a line.
@item i @var{text}
insert @var{text} before a line (alternative syntax).
@item l
Print the pattern space in an unambiguous form.
@item n
(next) If auto-print is not disabled, print the pattern space,
then, regardless, replace the pattern space with the next line of input.
If there is no more input then @command{sed} exits without processing
any more commands.
@item N
Add a newline to the pattern space,
then append the next line of input to the pattern space.
If there is no more input then @command{sed} exits without processing
any more commands.
@item p
Print the pattern space.
@c useful with @option{-n}
@item P
Print the pattern space, up to the first <newline>.
@item q@var{[exit-code]}
(quit) Exit @command{sed} without processing any more commands or input.
@item Q@var{[exit-code]}
(quit) This command is the same as @code{q}, but will not print the
contents of pattern space. Like @code{q}, it provides the
ability to return an exit code to the caller.
@c useful to quit on a conditional without printing
@item r filename
Reads file @var{filename}.
@item R filename
Queue a line of @var{filename} to be read and
inserted into the output stream at the end of the current cycle,
or when the next input line is read.
@c useful to interleave files
@item s@var{/regexp/replacement/[flags]}
(substitute) Match the regular-expression against the content of the
pattern space. If found, replace matched string with
@var{replacement}.
@item t @var{label}
(test) Branch to @var{label} only if there has been a successful
@code{s}ubstitution since the last input line was read or conditional
branch was taken. The @var{label} may be omitted, in which case the
next cycle is started.
@item T @var{label}
(test) Branch to @var{label} only if there have been no successful
@code{s}ubstitutions since the last input line was read or
conditional branch was taken. The @var{label} may be omitted,
in which case the next cycle is started.
@item v @var{[version]}
(version) This command does nothing, but makes @command{sed} fail if
@value{SSED} extensions are not supported, or if the requested version
is not available.
@item w filename
Write the pattern space to @var{filename}.
@item W filename
Write to the given filename the portion of the pattern space up to
the first newline
@item x
Exchange the contents of the hold and pattern spaces.
@item y/src/dst/
Transliterate any characters in the pattern space which match
any of the @var{source-chars} with the corresponding character
in @var{dest-chars}.
@item z
(zap) This command empties the content of pattern space.
@item #
A comment, until the next newline.
@item @{ @var{cmd ; cmd ...} @}
Group several commands together.
@c useful for multiple commands on same address
@item =
Print the current input line number (with a trailing newline).
@item : @var{label}
Specify the location of @var{label} for branch commands (@code{b},
@code{t}, @code{T}).
@end table
@node The "s" Command
@section The @code{s} Command
The @code{s} command (as in substitute) is probably the most important
in @command{sed} and has a lot of different options. The syntax of
the @code{s} command is
@samp{s/@var{regexp}/@var{replacement}/@var{flags}}.
Its basic concept is simple: the @code{s} command attempts to match
the pattern space against the supplied regular expression @var{regexp};
if the match is successful, then that portion of the
pattern space which was matched is replaced with @var{replacement}.
For details about @var{regexp} syntax @pxref{Regexp Addresses,,Regular
Expression Addresses}.
@cindex Backreferences, in regular expressions
@cindex Parenthesized substrings
The @var{replacement} can contain @code{\@var{n}} (@var{n} being
a number from 1 to 9, inclusive) references, which refer to
the portion of the match which is contained between the @var{n}th
@code{\(} and its matching @code{\)}.
Also, the @var{replacement} can contain unescaped @code{&}
characters which reference the whole matched portion
of the pattern space.
@c TODO: xref to backreference section mention @var{\'}.
The @code{/}
characters may be uniformly replaced by any other single
character within any given @code{s} command. The @code{/}
character (or whatever other character is used in its stead)
can appear in the @var{regexp} or @var{replacement}
only if it is preceded by a @code{\} character.
@cindex @value{SSEDEXT}, case modifiers in @code{s} commands
Finally, as a @value{SSED} extension, you can include a
special sequence made of a backslash and one of the letters
@code{L}, @code{l}, @code{U}, @code{u}, or @code{E}.
The meaning is as follows:
@table @code
@item \L
Turn the replacement
to lowercase until a @code{\U} or @code{\E} is found,
@item \l
Turn the
next character to lowercase,
@item \U
Turn the replacement to uppercase
until a @code{\L} or @code{\E} is found,
@item \u
Turn the next character
to uppercase,
@item \E
Stop case conversion started by @code{\L} or @code{\U}.
@end table
When the @code{g} flag is being used, case conversion does not
propagate from one occurrence of the regular expression to
another. For example, when the following command is executed
with @samp{a-b-} in pattern space:
@example
s/\(b\?\)-/x\u\1/g
@end example
@noindent
the output is @samp{axxB}. When replacing the first @samp{-},
the @samp{\u} sequence only affects the empty replacement of
@samp{\1}. It does not affect the @code{x} character that is
added to pattern space when replacing @code{b-} with @code{xB}.
On the other hand, @code{\l} and @code{\u} do affect the remainder
of the replacement text if they are followed by an empty substitution.
With @samp{a-b-} in pattern space, the following command:
@example
s/\(b\?\)-/\u\1x/g
@end example
@noindent
will replace @samp{-} with @samp{X} (uppercase) and @samp{b-} with
@samp{Bx}. If this behavior is undesirable, you can prevent it by
adding a @samp{\E} sequence---after @samp{\1} in this case.
To include a literal @code{\}, @code{&}, or newline in the final
replacement, be sure to precede the desired @code{\}, @code{&},
or newline in the @var{replacement} with a @code{\}.
@findex s command, option flags
@cindex Substitution of text, options
The @code{s} command can be followed by zero or more of the
following @var{flags}:
@table @code
@item g
@cindex Global substitution
@cindex Replacing all text matching regexp in a line
Apply the replacement to @emph{all} matches to the @var{regexp},
not just the first.
@item @var{number}
@cindex Replacing only @var{n}th match of regexp in a line
Only replace the @var{number}th match of the @var{regexp}.
@cindex GNU extensions, @code{g} and @var{number} modifier
interaction in @code{s} command
@cindex Mixing @code{g} and @var{number} modifiers in the @code{s} command
Note: the @sc{posix} standard does not specify what should happen
when you mix the @code{g} and @var{number} modifiers,
and currently there is no widely agreed upon meaning
across @command{sed} implementations.
For @value{SSED}, the interaction is defined to be:
ignore matches before the @var{number}th,
and then match and replace all matches from
the @var{number}th on.
@item p
@cindex Text, printing after substitution
If the substitution was made, then print the new pattern space.
Note: when both the @code{p} and @code{e} options are specified,
the relative ordering of the two produces very different results.
In general, @code{ep} (evaluate then print) is what you want,
but operating the other way round can be useful for debugging.
For this reason, the current version of @value{SSED} interprets
specially the presence of @code{p} options both before and after
@code{e}, printing the pattern space before and after evaluation,
while in general flags for the @code{s} command show their
effect just once. This behavior, although documented, might
change in future versions.
@item w @var{filename}
@cindex Text, writing to a file after substitution
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
@cindex @value{SSEDEXT}, @file{/dev/stderr} file
If the substitution was made, then write out the result to the named file.
As a @value{SSED} extension, two special values of @var{filename} are
supported: @file{/dev/stderr}, which writes the result to the standard
error, and @file{/dev/stdout}, which writes to the standard
output.@footnote{This is equivalent to @code{p} unless the @option{-i}
option is being used.}
@item e
@cindex Evaluate Bourne-shell commands, after substitution
@cindex Subprocesses
@cindex @value{SSEDEXT}, evaluating Bourne-shell commands
@cindex @value{SSEDEXT}, subprocesses
This command allows one to pipe input from a shell command
into pattern space. If a substitution was made, the command
that is found in pattern space is executed and pattern space
is replaced with its output. A trailing newline is suppressed;
results are undefined if the command to be executed contains
a @sc{nul} character. This is a @value{SSED} extension.
@item I
@itemx i
@cindex GNU extensions, @code{I} modifier
@cindex Case-insensitive matching
The @code{I} modifier to regular-expression matching is a GNU
extension which makes @command{sed} match @var{regexp} in a
case-insensitive manner.
@item M
@itemx m
@cindex @value{SSEDEXT}, @code{M} modifier
The @code{M} modifier to regular-expression matching is a @value{SSED}
extension which directs @value{SSED} to match the regular expression
in @cite{multi-line} mode. The modifier causes @code{^} and @code{$} to
match respectively (in addition to the normal behavior) the empty string
after a newline, and the empty string before a newline. There are
special character sequences
@ifclear PERL
(@code{\`} and @code{\'})
@end ifclear
which always match the beginning or the end of the buffer.
In addition,
the period character does not match a new-line character in
multi-line mode.
@end table
@node Common Commands
@section Often-Used Commands
If you use @command{sed} at all, you will quite likely want to know
these commands.
@table @code
@item #
[No addresses allowed.]
@findex # (comments)
@cindex Comments, in scripts
The @code{#} character begins a comment;
the comment continues until the next newline.
@cindex Portability, comments
If you are concerned about portability, be aware that
some implementations of @command{sed} (which are not @sc{posix}
conforming) may only support a single one-line comment,
and then only when the very first character of the script is a @code{#}.
@findex -n, forcing from within a script
@cindex Caveat --- #n on first line
Warning: if the first two characters of the @command{sed} script
are @code{#n}, then the @option{-n} (no-autoprint) option is forced.
If you want to put a comment in the first line of your script
and that comment begins with the letter @samp{n}
and you do not want this behavior,
then be sure to either use a capital @samp{N},
or place at least one space before the @samp{n}.
@item q [@var{exit-code}]
@findex q (quit) command
@cindex @value{SSEDEXT}, returning an exit code
@cindex Quitting
Exit @command{sed} without processing any more commands or input.
Example: stop after printing the second line:
@example
$ seq 3 | sed 2q
1
2
@end example
This command accepts only one address.
Note that the current pattern space is printed if auto-print is
not disabled with the @option{-n} options. The ability to return
an exit code from the @command{sed} script is a @value{SSED} extension.
See also the @value{SSED} extension @code{Q} command which quits silently
without printing the current pattern space.
@item d
@findex d (delete) command
@cindex Text, deleting
Delete the pattern space;
immediately start next cycle.
Example: delete the second input line:
@example
$ seq 3 | sed 2d
1
3
@end example
@item p
@findex p (print) command
@cindex Text, printing
Print out the pattern space (to the standard output).
This command is usually only used in conjunction with the @option{-n}
command-line option.
Example: print only the second input line:
@example
$ seq 3 | sed -n 2p
2
@end example
@item n
@findex n (next-line) command
@cindex Next input line, replace pattern space with
@cindex Read next input line
If auto-print is not disabled, print the pattern space,
then, regardless, replace the pattern space with the next line of input.
If there is no more input then @command{sed} exits without processing
any more commands.
This command is useful to skip lines (e.g. process every Nth line).
Example: perform substitution on every 3rd line (i.e. two @code{n} commands
skip two lines):
@codequoteundirected on
@codequotebacktick on
@example
$ seq 6 | sed 'n;n;s/./x/'
1
2
x
4
5
x
@end example
@value{SSED} provides an extension address syntax of @var{first}~@var{step}
to achieve the same result:
@example
$ seq 6 | sed '0~3s/./x/'
1
2
x
4
5
x
@end example
@codequotebacktick off
@codequoteundirected off
@item @{ @var{commands} @}
@findex @{@} command grouping
@cindex Grouping commands
@cindex Command groups
A group of commands may be enclosed between
@code{@{} and @code{@}} characters.
This is particularly useful when you want a group of commands
to be triggered by a single address (or address-range) match.
Example: perform substitution then print the second input line:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed -n '2@{s/2/X/ ; p@}'
X
@end example
@codequoteundirected off
@codequotebacktick off
@end table
@node Other Commands
@section Less Frequently-Used Commands
Though perhaps less frequently used than those in the previous
section, some very small yet useful @command{sed} scripts can be built with
these commands.
@table @code
@item y/@var{source-chars}/@var{dest-chars}/
@findex y (transliterate) command
@cindex Transliteration
Transliterate any characters in the pattern space which match
any of the @var{source-chars} with the corresponding character
in @var{dest-chars}.
Example: transliterate @samp{a-j} into @samp{0-9}:
@codequoteundirected on
@codequotebacktick on
@example
$ echo hello world | sed 'y/abcdefghij/0123456789/'
74llo worl3
@end example
@codequoteundirected off
@codequotebacktick off
(The @code{/} characters may be uniformly replaced by
any other single character within any given @code{y} command.)
Instances of the @code{/} (or whatever other character is used in its stead),
@code{\}, or newlines can appear in the @var{source-chars} or @var{dest-chars}
lists, provide that each instance is escaped by a @code{\}.
The @var{source-chars} and @var{dest-chars} lists @emph{must}
contain the same number of characters (after de-escaping).
See the @command{tr} command from GNU coreutils for similar functionality.
@item a @var{text}
Appending @var{text} after a line. This is a GNU extension
to the standard @code{a} command - see below for details.
Example: Add @samp{hello} after the second line:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2a hello'
1
2
hello
3
@end example
@codequoteundirected off
@codequotebacktick off
Leading whitespace after the @code{a} command is ignored.
The text to add is read until the end of the line.
@item a\
@itemx @var{text}
@findex a (append text lines) command
@cindex Appending text after a line
@cindex Text, appending
Appending @var{text} after a line.
Example: Add @samp{hello} after the second line
(@print{} indicates printed output lines):
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2a\
hello'
@print{}1
@print{}2
@print{}hello
@print{}3
@end example
@codequoteundirected off
@codequotebacktick off
The @code{a} command queues the lines of text which follow this command
(each but the last ending with a @code{\},
which are removed from the output)
to be output at the end of the current cycle,
or when the next input line is read.
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a GNU extension, this command accepts two addresses.
Escape sequences in @var{text} are processed, so you should
use @code{\\} in @var{text} to print a single backslash.
The commands resume after the last line without a backslash (@code{\}) -
@samp{world} in the following example:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2a\
hello\
world
3s/./X/'
@print{}1
@print{}2
@print{}hello
@print{}world
@print{}X
@end example
@codequoteundirected off
@codequotebacktick off
As a GNU extension, the @code{a} command and @var{text} can be
separated into two @code{-e} parameters, enabling easier scripting:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed -e '2a\' -e hello
1
2
hello
3
$ sed -e '2a\' -e "$VAR"
@end example
@codequoteundirected off
@codequotebacktick off
@item i @var{text}
insert @var{text} before a line. This is a GNU extension
to the standard @code{i} command - see below for details.
Example: Insert @samp{hello} before the second line:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2i hello'
1
hello
2
3
@end example
@codequoteundirected off
@codequotebacktick off
Leading whitespace after the @code{i} command is ignored.
The text to add is read until the end of the line.
@anchor{insert command}
@item i\
@itemx @var{text}
@findex i (insert text lines) command
@cindex Inserting text before a line
@cindex Text, insertion
Immediately output the lines of text which follow this command.
Example: Insert @samp{hello} before the second line
(@print{} indicates printed output lines):
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2i\
hello'
@print{}1
@print{}hello
@print{}2
@print{}3
@end example
@codequoteundirected off
@codequotebacktick off
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a GNU extension, this command accepts two addresses.
Escape sequences in @var{text} are processed, so you should
use @code{\\} in @var{text} to print a single backslash.
The commands resume after the last line without a backslash (@code{\}) -
@samp{world} in the following example:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2i\
hello\
world
s/./X/'
@print{}X
@print{}hello
@print{}world
@print{}X
@print{}X
@end example
@codequoteundirected off
@codequotebacktick off
As a GNU extension, the @code{i} command and @var{text} can be
separated into two @code{-e} parameters, enabling easier scripting:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed -e '2i\' -e hello
1
hello
2
3
$ sed -e '2i\' -e "$VAR"
@end example
@codequoteundirected off
@codequotebacktick off
@item c @var{text}
Replaces the line(s) with @var{text}. This is a GNU extension
to the standard @code{c} command - see below for details.
Example: Replace the 2nd to 9th lines with the word @samp{hello}:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 10 | sed '2,9c hello'
1
hello
10
@end example
@codequoteundirected off
@codequotebacktick off
Leading whitespace after the @code{c} command is ignored.
The text to add is read until the end of the line.
@item c\
@itemx @var{text}
@findex c (change to text lines) command
@cindex Replacing selected lines with other text
Delete the lines matching the address or address-range,
and output the lines of text which follow this command.
Example: Replace 2nd to 4th lines with the words @samp{hello} and
@samp{world} (@print{} indicates printed output lines):
@codequoteundirected on
@codequotebacktick on
@example
$ seq 5 | sed '2,4c\
hello\
world'
@print{}1
@print{}hello
@print{}world
@print{}5
@end example
@codequoteundirected off
@codequotebacktick off
If no addresses are given, each line is replaced.
A new cycle is started after this command is done,
since the pattern space will have been deleted.
In the following example, the @code{c} starts a
new cycle and the substitution command is not performed
on the replaced text:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2c\
hello
s/./X/'
@print{}X
@print{}hello
@print{}X
@end example
@codequoteundirected off
@codequotebacktick off
As a GNU extension, the @code{c} command and @var{text} can be
separated into two @code{-e} parameters, enabling easier scripting:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed -e '2c\' -e hello
1
hello
3
$ sed -e '2c\' -e "$VAR"
@end example
@codequoteundirected off
@codequotebacktick off
@item =
@findex = (print line number) command
@cindex Printing line number
@cindex Line number, printing
Print out the current input line number (with a trailing newline).
@codequoteundirected on
@codequotebacktick on
@example
$ printf '%s\n' aaa bbb ccc | sed =
1
aaa
2
bbb
3
ccc
@end example
@codequoteundirected off
@codequotebacktick off
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a GNU extension, this command accepts two addresses.
@item l @var{n}
@findex l (list unambiguously) command
@cindex List pattern space
@cindex Printing text unambiguously
@cindex Line length, setting
@cindex @value{SSEDEXT}, setting line length
Print the pattern space in an unambiguous form:
non-printable characters (and the @code{\} character)
are printed in C-style escaped form; long lines are split,
with a trailing @code{\} character to indicate the split;
the end of each line is marked with a @code{$}.
@var{n} specifies the desired line-wrap length;
a length of 0 (zero) means to never wrap long lines. If omitted,
the default as specified on the command line is used. The @var{n}
parameter is a @value{SSED} extension.
@item r @var{filename}
@findex r (read file) command
@cindex Read text from a file
Reads file @var{filename}. Example:
@codequoteundirected on
@codequotebacktick on
@example
$ seq 3 | sed '2r/etc/hostname'
1
2
fencepost.gnu.org
3
@end example
@codequoteundirected off
@codequotebacktick off
@cindex @value{SSEDEXT}, @file{/dev/stdin} file
Queue the contents of @var{filename} to be read and
inserted into the output stream at the end of the current cycle,
or when the next input line is read.
Note that if @var{filename} cannot be read, it is treated as
if it were an empty file, without any error indication.
As a @value{SSED} extension, the special value @file{/dev/stdin}
is supported for the file name, which reads the contents of the
standard input.
@cindex @value{SSEDEXT}, two addresses supported by most commands
As a GNU extension, this command accepts two addresses. The
file will then be reread and inserted on each of the addressed lines.
As a @value{SSED} extension, the @code{r} command accepts a zero address,
inserting a file @emph{before} the first line of the input
@pxref{Adding a header to multiple files}.
@item w @var{filename}
@findex w (write file) command
@cindex Write to a file
@cindex @value{SSEDEXT}, @file{/dev/stdout} file
@cindex @value{SSEDEXT}, @file{/dev/stderr} file
Write the pattern space to @var{filename}.
As a @value{SSED} extension, two special values of @var{filename} are
supported: @file{/dev/stderr}, which writes the result to the standard
error, and @file{/dev/stdout}, which writes to the standard
output.@footnote{This is equivalent to @code{p} unless the @option{-i}
option is being used.}
The file will be created (or truncated) before the first input line is
read; all @code{w} commands (including instances of the @code{w} flag
on successful @code{s} commands) which refer to the same @var{filename}
are output without closing and reopening the file.
@item D
@findex D (delete first line) command
@cindex Delete first line from pattern space
If pattern space contains no newline, start a normal new cycle as if
the @code{d} command was issued. Otherwise, delete text in the pattern
space up to the first newline, and restart cycle with the resultant
pattern space, without reading a new line of input.
@item N
@findex N (append Next line) command
@cindex Next input line, append to pattern space
@cindex Append next input line to pattern space
Add a newline to the pattern space,
then append the next line of input to the pattern space.
If there is no more input then @command{sed} exits without processing
any more commands.
When @option{-z} is used, a zero byte (the ascii @samp{NUL} character) is
added between the lines (instead of a new line).
By default @command{sed} does not terminate if there is no 'next' input line.
This is a GNU extension which can be disabled with @option{--posix}.
@xref{N_command_last_line,,N command on the last line}.
@item P
@findex P (print first line) command
@cindex Print first line from pattern space
Print out the portion of the pattern space up to the first newline.
@item h
@findex h (hold) command
@cindex Copy pattern space into hold space
@cindex Replace hold space with copy of pattern space
@cindex Hold space, copying pattern space into
Replace the contents of the hold space with the contents of the pattern space.
@item H
@findex H (append Hold) command
@cindex Append pattern space to hold space
@cindex Hold space, appending from pattern space
Append a newline to the contents of the hold space,
and then append the contents of the pattern space to that of the hold space.
@item g
@findex g (get) command
@cindex Copy hold space into pattern space
@cindex Replace pattern space with copy of hold space
@cindex Hold space, copy into pattern space
Replace the contents of the pattern space with the contents of the hold space.
@item G
@findex G (appending Get) command
@cindex Append hold space to pattern space
@cindex Hold space, appending to pattern space
Append a newline to the contents of the pattern space,
and then append the contents of the hold space to that of the pattern space.
@item x
@findex x (eXchange) command
@cindex Exchange hold space with pattern space
@cindex Hold space, exchange with pattern space
Exchange the contents of the hold and pattern spaces.
@end table
@node Programming Commands
@section Commands for @command{sed} gurus
In most cases, use of these commands indicates that you are
probably better off programming in something like @command{awk}
or Perl. But occasionally one is committed to sticking
with @command{sed}, and these commands can enable one to write
quite convoluted scripts.
@cindex Flow of control in scripts
@table @code
@item : @var{label}
[No addresses allowed.]
@findex : (label) command
@cindex Labels, in scripts
Specify the location of @var{label} for branch commands.
In all other respects, a no-op.
@item b @var{label}
@findex b (branch) command
@cindex Branch to a label, unconditionally
@cindex Goto, in scripts
Unconditionally branch to @var{label}.
The @var{label} may be omitted, in which case the next cycle is started.
@item t @var{label}
@findex t (test and branch if successful) command
@cindex Branch to a label, if @code{s///} succeeded
@cindex Conditional branch
Branch to @var{label} only if there has been a successful @code{s}ubstitution
since the last input line was read or conditional branch was taken.
The @var{label} may be omitted, in which case the next cycle is started.
@end table
@node Extended Commands
@section Commands Specific to @value{SSED}
These commands are specific to @value{SSED}, so you
must use them with care and only when you are sure that
hindering portability is not evil. They allow you to check
for @value{SSED} extensions or to do tasks that are required
quite often, yet are unsupported by standard @command{sed}s.
@table @code
@item e [@var{command}]
@findex e (evaluate) command
@cindex Evaluate Bourne-shell commands
@cindex Subprocesses
@cindex @value{SSEDEXT}, evaluating Bourne-shell commands
@cindex @value{SSEDEXT}, subprocesses
This command allows one to pipe input from a shell command
into pattern space. Without parameters, the @code{e} command
executes the command that is found in pattern space and
replaces the pattern space with the output; a trailing newline
is suppressed.
If a parameter is specified, instead, the @code{e} command
interprets it as a command and sends its output to the output stream.
The command can run across multiple lines, all but the last ending with
a back-slash.
In both cases, the results are undefined if the command to be
executed contains a @sc{nul} character.
Note that, unlike the @code{r} command, the output of the command will
be printed immediately; the @code{r} command instead delays the output
to the end of the current cycle.
@item F
@findex F (File name) command
@cindex Printing file name
@cindex File name, printing
Print out the file name of the current input file (with a trailing
newline).
@item Q [@var{exit-code}]
This command accepts only one address.
@findex Q (silent Quit) command
@cindex @value{SSEDEXT}, quitting silently
@cindex @value{SSEDEXT}, returning an exit code
@cindex Quitting
This command is the same as @code{q}, but will not print the
contents of pattern space. Like @code{q}, it provides the
ability to return an exit code to the caller.
This command can be useful because the only alternative ways
to accomplish this apparently trivial function are to use
the @option{-n} option (which can unnecessarily complicate
your script) or resorting to the following snippet, which
wastes time by reading the whole file without any visible effect:
@example
:eat
$d @i{@r{Quit silently on the last line}}
N @i{@r{Read another line, silently}}
g @i{@r{Overwrite pattern space each time to save memory}}
b eat
@end example
@item R @var{filename}
@findex R (read line) command
@cindex Read text from a file
@cindex @value{SSEDEXT}, reading a file a line at a time
@cindex @value{SSEDEXT}, @code{R} command
@cindex @value{SSEDEXT}, @file{/dev/stdin} file
Queue a line of @var{filename} to be read and
inserted into the output stream at the end of the current cycle,
or when the next input line is read.
Note that if @var{filename} cannot be read, or if its end is
reached, no line is appended, without any error indication.
As with the @code{r} command, the special value @file{/dev/stdin}
is supported for the file name, which reads a line from the
standard input.
@item T @var{label}
@findex T (test and branch if failed) command
@cindex @value{SSEDEXT}, branch if @code{s///} failed
@cindex Branch to a label, if @code{s///} failed
@cindex Conditional branch
Branch to @var{label} only if there have been no successful
@code{s}ubstitutions since the last input line was read or
conditional branch was taken. The @var{label} may be omitted,
in which case the next cycle is started.
@item v @var{version}
@findex v (version) command
@cindex @value{SSEDEXT}, checking for their presence
@cindex Requiring @value{SSED}
This command does nothing, but makes @command{sed} fail if
@value{SSED} extensions are not supported, simply because other
versions of @command{sed} do not implement it. In addition, you
can specify the version of @command{sed} that your script
requires, such as @code{4.0.5}. The default is @code{4.0}
because that is the first version that implemented this command.
This command enables all @value{SSEDEXT} even if
@env{POSIXLY_CORRECT} is set in the environment.
@item W @var{filename}
@findex W (write first line) command
@cindex Write first line to a file
@cindex @value{SSEDEXT}, writing first line to a file
Write to the given filename the portion of the pattern space up to
the first newline. Everything said under the @code{w} command about
file handling holds here too.
@item z
@findex z (Zap) command
@cindex @value{SSEDEXT}, emptying pattern space
@cindex Emptying pattern space
This command empties the content of pattern space. It is
usually the same as @samp{s/.*//}, but is more efficient
and works in the presence of invalid multibyte sequences
in the input stream. @sc{posix} mandates that such sequences
are @emph{not} matched by @samp{.}, so that there is no portable
way to clear @command{sed}'s buffers in the middle of the
script in most multibyte locales (including UTF-8 locales).
@end table
@node Multiple commands syntax
@section Multiple commands syntax
@c POSIX says:
@c Editing commands other than {...}, a, b, c, i, r, t, w, :, and #
@c can be followed by a <semicolon>, optional <blank> characters, and
@c another editing command. However, when an s editing command is used
@c with the w flag, following it with another command in this manner
@c produces undefined results.
There are several methods to specify multiple commands in a @command{sed}
program.
Using newlines is most natural when running a sed script from a file
(using the @option{-f} option).
On the command line, all @command{sed} commands may be separated by newlines.
Alternatively, you may specify each command as an argument to an @option{-e}
option:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 6 | sed '1d
3d
5d'
2
4
6
$ seq 6 | sed -e 1d -e 3d -e 5d
2
4
6
@end group
@end example
@codequoteundirected off
@codequotebacktick off
A semicolon (@samp{;}) may be used to separate most simple commands:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 6 | sed '1d;3d;5d'
2
4
6
@end group
@end example
@codequoteundirected off
@codequotebacktick off
The @code{@{},@code{@}},@code{b},@code{t},@code{T},@code{:} commands can
be separated with a semicolon (this is a non-portable @value{SSED} extension).
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 4 | sed '@{1d;3d@}'
2
4
$ seq 6 | sed '@{1d;3d@};5d'
2
4
6
@end group
@end example
@codequoteundirected off
@codequotebacktick off
Labels used in @code{b},@code{t},@code{T},@code{:} commands are read
until a semicolon. Leading and trailing whitespace is ignored. In
the examples below the label is @samp{x}. The first example works
with @value{SSED}. The second is a portable equivalent. For more
information about branching and labels @pxref{Branching and flow
control}.
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 3 | sed '/1/b x ; s/^/=/ ; :x ; 3d'
1
=2
$ seq 3 | sed -e '/1/bx' -e 's/^/=/' -e ':x' -e '3d'
1
=2
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@subsection Commands Requiring a newline
The following commands cannot be separated by a semicolon and
require a newline:
@table @asis
@item @code{a},@code{c},@code{i} (append/change/insert)
All characters following @code{a},@code{c},@code{i} commands are taken
as the text to append/change/insert. Using a semicolon leads to
undesirable results:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 2 | sed '1aHello ; 2d'
1
Hello ; 2d
2
@end group
@end example
@codequoteundirected off
@codequotebacktick off
Separate the commands using @option{-e} or a newline:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 2 | sed -e 1aHello -e 2d
1
Hello
$ seq 2 | sed '1aHello
2d'
1
Hello
@end group
@end example
@codequoteundirected off
@codequotebacktick off
Note that specifying the text to add (@samp{Hello}) immediately
after @code{a},@code{c},@code{i} is itself a @value{SSED} extension.
A portable, POSIX-compliant alternative is:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 2 | sed '1a\
Hello
2d'
1
Hello
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@item @code{#} (comment)
All characters following @samp{#} until the next newline are ignored.
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 3 | sed '# this is a comment ; 2d'
1
2
3
$ seq 3 | sed '# this is a comment
2d'
1
3
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@item @code{r},@code{R},@code{w},@code{W} (reading and writing files)
The @code{r},@code{R},@code{w},@code{W} commands parse the filename
until end of the line. If whitespace, comments or semicolons are found,
they will be included in the filename, leading to unexpected results:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 2 | sed '1w hello.txt ; 2d'
1
2
$ ls -log
total 4
-rw-rw-r-- 1 2 Jan 23 23:03 hello.txt ; 2d
$ cat 'hello.txt ; 2d'
1
@end group
@end example
@codequoteundirected off
@codequotebacktick off
Note that @command{sed} silently ignores read/write errors in
@code{r},@code{R},@code{w},@code{W} commands (such as missing files).
In the following example, @command{sed} tries to read a file named
@samp{@file{hello.txt ; N}}. The file is missing, and the error is silently
ignored:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo x | sed '1rhello.txt ; N'
x
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@item @code{e} (command execution)
Any characters following the @code{e} command until the end of the line
will be sent to the shell. If whitespace, comments or semicolons are found,
they will be included in the shell command, leading to unexpected results:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo a | sed '1e touch foo#bar'
a
$ ls -1
foo#bar
$ echo a | sed '1e touch foo ; s/a/b/'
sh: 1: s/a/b/: not found
a
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@item @code{s///[we]} (substitute with @code{e} or @code{w} flags)
In a substitution command, the @code{w} flag writes the substitution
result to a file, and the @code{e} flag executes the substitution result
as a shell command. As with the @code{r/R/w/W/e} commands, these
must be terminated with a newline. If whitespace, comments or semicolons
are found, they will be included in the shell command or filename, leading to
unexpected results:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo a | sed 's/a/b/w1.txt#foo'
b
$ ls -1
1.txt#foo
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@end table
@node sed addresses
@chapter Addresses: selecting lines
@menu
* Addresses overview:: Addresses overview
* Numeric Addresses:: selecting lines by numbers
* Regexp Addresses:: selecting lines by text matching
* Range Addresses:: selecting a range of lines
* Zero Address:: Using address @code{0}
@end menu
@node Addresses overview
@section Addresses overview
@cindex addresses, numeric
@cindex numeric addresses
Addresses determine on which line(s) the @command{sed} command will be
executed. The following command replaces any first occurrence of @samp{hello}
with @samp{world} only on line 144:
@codequoteundirected on
@codequotebacktick on
@example
sed '144s/hello/world/' input.txt > output.txt
@end example
@codequoteundirected off
@codequotebacktick off
If no address is specified, the command is performed on all lines.
The following command replaces @samp{hello} with @samp{world},
targeting every line of the input file.
However, note that it modifies only the first instance of @samp{hello}
on each line.
Use the @samp{g} modifier to affect every instance on each affected line.
@codequoteundirected on
@codequotebacktick on
@example
sed 's/hello/world/' input.txt > output.txt
@end example
@codequoteundirected off
@codequotebacktick off
@cindex addresses, regular expression
@cindex regular expression addresses
Addresses can contain regular expressions to match lines based
on content instead of line numbers. The following command replaces
@samp{hello} with @samp{world} only on lines
containing the string @samp{apple}:
@codequoteundirected on
@codequotebacktick on
@example
sed '/apple/s/hello/world/' input.txt > output.txt
@end example
@codequoteundirected off
@codequotebacktick off
@cindex addresses, range
@cindex range addresses
An address range is specified with two addresses separated by a comma
(@code{,}). Addresses can be numeric, regular expressions, or a mix of
both.
The following command replaces @samp{hello} with @samp{world}
only on lines 4 to 17 (inclusive):
@codequoteundirected on
@codequotebacktick on
@example
sed '4,17s/hello/world/' input.txt > output.txt
@end example
@codequoteundirected off
@codequotebacktick off
@cindex Excluding lines
@cindex Selecting non-matching lines
@cindex addresses, negating
@cindex addresses, excluding
Appending the @code{!} character to the end of an address
specification (before the command letter) negates the sense of the
match. That is, if the @code{!} character follows an address or an
address range, then only lines which do @emph{not} match the addresses
will be selected. The following command replaces @samp{hello}
with @samp{world} only on lines @emph{not} containing the string
@samp{apple}:
@example
sed '/apple/!s/hello/world/' input.txt > output.txt
@end example
The following command replaces @samp{hello} with
@samp{world} only on lines 1 to 3 and from line 18 to the last line of the
input file (i.e. excluding lines 4 to 17):
@example
sed '4,17!s/hello/world/' input.txt > output.txt
@end example
@node Numeric Addresses
@section Selecting lines by numbers
@cindex Addresses, in @command{sed} scripts
@cindex Line selection
@cindex Selecting lines to process
Addresses in a @command{sed} script can be in any of the following forms:
@table @code
@item @var{number}
@cindex Address, numeric
@cindex Line, selecting by number
Specifying a line number will match only that line in the input.
(Note that @command{sed} counts lines continuously across all input files
unless @option{-i} or @option{-s} options are specified.)
@item $
@cindex Address, last line
@cindex Last line, selecting
@cindex Line, selecting last
This address matches the last line of the last file of input, or
the last line of each file when the @option{-i} or @option{-s} options
are specified.
@item @var{first}~@var{step}
@cindex GNU extensions, @samp{@var{n}~@var{m}} addresses
This GNU extension matches every @var{step}th line
starting with line @var{first}.
In particular, lines will be selected when there exists
a non-negative @var{n} such that the current line-number equals
@var{first} + (@var{n} * @var{step}).
Thus, one would use @code{1~2} to select the odd-numbered lines and
@code{0~2} for even-numbered lines;
to pick every third line starting with the second, @samp{2~3} would be used;
to pick every fifth line starting with the tenth, use @samp{10~5};
and @samp{50~0} is just an obscure way of saying @code{50}.
The following commands demonstrate the step address usage:
@example
$ seq 10 | sed -n '0~4p'
4
8
$ seq 10 | sed -n '1~3p'
1
4
7
10
@end example
@end table
@node Regexp Addresses
@section selecting lines by text matching
@value{SSED} supports the following regular expression addresses.
The default regular expression is
@ref{BRE syntax, , Basic Regular Expression (BRE)}.
If @option{-E} or @option{-r} options are used, The regular expression should be
in @ref{ERE syntax, , Extended Regular Expression (ERE)} syntax.
@xref{BRE vs ERE}.
@table @code
@item /@var{regexp}/
@cindex Address, as a regular expression
@cindex Line, selecting by regular expression match
This will select any line which matches the regular expression @var{regexp}.
If @var{regexp} itself includes any @code{/} characters,
each must be escaped by a backslash (@code{\}).
The following command prints lines in @file{/etc/passwd}
which end with @samp{bash}@footnote{
There are of course many other ways to do the same,
e.g.
@example
grep 'bash$' /etc/passwd
awk -F: '$7 == "/bin/bash"' /etc/passwd
@end example
}:
@example
sed -n '/bash$/p' /etc/passwd
@end example
@cindex empty regular expression
@cindex @value{SSEDEXT}, modifiers and the empty regular expression
The empty regular expression @samp{//} repeats the last regular
expression match (the same holds if the empty regular expression is
passed to the @code{s} command). Note that modifiers to regular expressions
are evaluated when the regular expression is compiled, thus it is invalid to
specify them together with the empty regular expression.
@item \%@var{regexp}%
(The @code{%} may be replaced by any other single character.)
@cindex Slash character, in regular expressions
This also matches the regular expression @var{regexp},
but allows one to use a different delimiter than @code{/}.
This is particularly useful if the @var{regexp} itself contains
a lot of slashes, since it avoids the tedious escaping of every @code{/}.
If @var{regexp} itself includes any delimiter characters,
each must be escaped by a backslash (@code{\}).
The following commands are equivalent. They print lines
which start with @samp{/home/alice/documents/}:
@example
sed -n '/^\/home\/alice\/documents\//p'
sed -n '\%^/home/alice/documents/%p'
sed -n '\;^/home/alice/documents/;p'
@end example
@item /@var{regexp}/I
@itemx \%@var{regexp}%I
@cindex GNU extensions, @code{I} modifier
@cindex case insensitive, regular expression
The @code{I} modifier to regular-expression matching is a GNU
extension which causes the @var{regexp} to be matched in
a case-insensitive manner.
In many other programming languages, a lower case @code{i} is used
for case-insensitive regular expression matching. However, in @command{sed}
the @code{i} is used for the insert command (@pxref{insert command}).
Observe the difference between the following examples.
In this example, @code{/b/I} is the address: regular expression with @code{I}
modifier. @code{d} is the delete command:
@example
$ printf "%s\n" a b c | sed '/b/Id'
a
c
@end example
Here, @code{/b/} is the address: a regular expression.
@code{i} is the insert command.
@code{d} is the value to insert.
A line with @samp{d} is then inserted above the matched line:
@example
$ printf "%s\n" a b c | sed '/b/id'
a
d
b
c
@end example
@item /@var{regexp}/M
@itemx \%@var{regexp}%M
@cindex @value{SSEDEXT}, @code{M} modifier
The @code{M} modifier to regular-expression matching is a @value{SSED}
extension which directs @value{SSED} to match the regular expression
in @cite{multi-line} mode. The modifier causes @code{^} and @code{$} to
match respectively (in addition to the normal behavior) the empty string
after a newline, and the empty string before a newline. There are
special character sequences
@ifclear PERL
(@code{\`} and @code{\'})
@end ifclear
which always match the beginning or the end of the buffer.
In addition,
the period character does not match a new-line character in
multi-line mode.
@end table
@cindex regex addresses and pattern space
@cindex regex addresses and input lines
Regex addresses operate on the content of the current
pattern space. If the pattern space is changed (for example with @code{s///}
command) the regular expression matching will operate on the changed text.
In the following example, automatic printing is disabled with
@option{-n}. The @code{s/2/X/} command changes lines containing
@samp{2} to @samp{X}. The command @code{/[0-9]/p} matches
lines with digits and prints them.
Because the second line is changed before the @code{/[0-9]/} regex,
it will not match and will not be printed:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 3 | sed -n 's/2/X/ ; /[0-9]/p'
1
3
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@node Range Addresses
@section Range Addresses
@cindex Range of lines
@cindex Several lines, selecting
An address range can be specified by specifying two addresses
separated by a comma (@code{,}). An address range matches lines
starting from where the first address matches, and continues
until the second address matches (inclusively):
@example
$ seq 10 | sed -n '4,6p'
4
5
6
@end example
If the second address is a @var{regexp}, then checking for the
ending match will start with the line @emph{following} the
line which matched the first address: a range will always
span at least two lines (except of course if the input stream
ends).
@example
$ seq 10 | sed -n '4,/[0-9]/p'
4
5
@end example
If the second address is a @var{number} less than (or equal to)
the line matching the first address, then only the one line is
matched:
@example
$ seq 10 | sed -n '4,1p'
4
@end example
@anchor{Zero Address Regex Range}
@cindex Special addressing forms
@cindex Range with start address of zero
@cindex Zero, as range start address
@cindex @var{addr1},+N
@cindex @var{addr1},~N
@cindex GNU extensions, special two-address forms
@cindex GNU extensions, @code{0} address
@cindex GNU extensions, 0,@var{addr2} addressing
@cindex GNU extensions, @var{addr1},+@var{N} addressing
@cindex GNU extensions, @var{addr1},~@var{N} addressing
@value{SSED} also supports some special two-address forms; all these
are GNU extensions:
@table @code
@item 0,/@var{regexp}/
A line number of @code{0} can be used in an address specification like
@code{0,/@var{regexp}/} so that @command{sed} will try to match
@var{regexp} in the first input line too. In other words,
@code{0,/@var{regexp}/} is similar to @code{1,/@var{regexp}/},
except that if @var{addr2} matches the very first line of input the
@code{0,/@var{regexp}/} form will consider it to end the range, whereas
the @code{1,/@var{regexp}/} form will match the beginning of its range and
hence make the range span up to the @emph{second} occurrence of the
regular expression.
The following examples demonstrate the difference between starting
with address 1 and 0:
@example
$ seq 10 | sed -n '1,/[0-9]/p'
1
2
$ seq 10 | sed -n '0,/[0-9]/p'
1
@end example
@item @var{addr1},+@var{N}
Matches @var{addr1} and the @var{N} lines following @var{addr1}.
@example
$ seq 10 | sed -n '6,+2p'
6
7
8
@end example
@var{addr1} can be a line number or a regular expression.
@item @var{addr1},~@var{N}
Matches @var{addr1} and the lines following @var{addr1}
until the next line whose input line number is a multiple of @var{N}.
The following command prints starting at line 6, until the next line which
is a multiple of 4 (i.e. line 8):
@example
$ seq 10 | sed -n '6,~4p'
6
7
8
@end example
@var{addr1} can be a line number or a regular expression.
@end table
@node Zero Address
@section Zero Address
@cindex Zero Address
As a @value{SSED} extension, @code{0} address can be used in two cases:
@enumerate
@item
In a regex range addresses as @code{0,/@var{regexp}/}
(@pxref{Zero Address Regex Range}).
@item
With the @code{r} command, inserting a file before the first line
(@pxref{Adding a header to multiple files}).
@end enumerate
Note that these are the only places where the @code{0} address makes
sense; Commands which are given the @code{0} address in any
other way will give an error.
@node sed regular expressions
@chapter Regular Expressions: selecting text
@menu
* Regular Expressions Overview:: Overview of Regular expression in @command{sed}
* BRE vs ERE:: Basic (BRE) and extended (ERE) regular expression
syntax
* BRE syntax:: Overview of basic regular expression syntax
* ERE syntax:: Overview of extended regular expression syntax
* Character Classes and Bracket Expressions::
* regexp extensions:: Additional regular expression commands
* Back-references and Subexpressions:: Back-references and Subexpressions
* Escapes:: Specifying special characters
* Locale Considerations:: Multibyte characters and locale considerations
@end menu
@node Regular Expressions Overview
@section Overview of regular expression in @command{sed}
@c NOTE: Keep examples in the 'overview' section
@c neutral in regards to BRE/ERE - to ease understanding.
To know how to use @command{sed}, people should understand regular
expressions (@dfn{regexp} for short). A regular expression
is a pattern that is matched against a
subject string from left to right. Most characters are
@dfn{ordinary}: they stand for
themselves in a pattern, and match the corresponding characters.
Regular expressions in @command{sed} are specified between two
slashes.
The following command prints lines containing the string @samp{hello}:
@example
sed -n '/hello/p'
@end example
The above example is equivalent to this @command{grep} command:
@example
grep 'hello'
@end example
The power of regular expressions comes from the ability to include
alternatives and repetitions in the pattern. These are encoded in the
pattern by the use of @dfn{special characters}, which do not stand for
themselves but instead are interpreted in some special way.
The character @code{^} (caret) in a regular expression matches the
beginning of the line. The character @code{.} (dot) matches any single
character. The following @command{sed} command matches and prints
lines which start with the letter @samp{b}, followed by any single character,
followed by the letter @samp{d}:
@example
$ printf "%s\n" abode bad bed bit bid byte body | sed -n '/^b.d/p'
bad
bed
bid
body
@end example
The following sections explain the meaning and usage of special
characters in regular expressions.
@node BRE vs ERE
@section Basic (BRE) and extended (ERE) regular expression
Basic and extended regular expressions are two variations on the
syntax of the specified pattern. Basic Regular Expression (BRE) syntax is the
default in @command{sed} (and similarly in @command{grep}).
Use the POSIX-specified @option{-E} option (@option{-r},
@option{--regexp-extended}) to enable Extended Regular Expression (ERE) syntax.
In @value{SSED}, the only difference between basic and extended regular
expressions is in the behavior of a few special characters: @samp{?},
@samp{+}, parentheses, braces (@samp{@{@}}), and @samp{|}.
With basic (BRE) syntax, these characters do not have special meaning
unless prefixed with a backslash (@samp{\}); While with extended (ERE) syntax
it is reversed: these characters are special unless they are prefixed
with backslash (@samp{\}).
@multitable @columnfractions .28 .36 .35
@headitem Desired pattern
@tab Basic (BRE) Syntax
@tab Extended (ERE) Syntax
@item literal @samp{+} (plus sign)
@tab
@exampleindent 0
@codequoteundirected on
@codequotebacktick on
@example
$ echo 'a+b=c' > foo
$ sed -n '/a+b/p' foo
a+b=c
@end example
@codequotebacktick off
@codequoteundirected off
@tab
@exampleindent 0
@codequoteundirected on
@codequotebacktick on
@example
$ echo 'a+b=c' > foo
$ sed -E -n '/a\+b/p' foo
a+b=c
@end example
@codequotebacktick off
@codequoteundirected off
@item One or more @samp{a} characters followed by @samp{b}
(plus sign as special meta-character)
@tab
@exampleindent 0
@codequoteundirected on
@codequotebacktick on
@example
$ echo aab > foo
$ sed -n '/a\+b/p' foo
aab
@end example
@codequotebacktick off
@codequoteundirected off
@tab
@exampleindent 0
@codequoteundirected on
@codequotebacktick on
@example
$ echo aab > foo
$ sed -E -n '/a+b/p' foo
aab
@end example
@codequotebacktick off
@codequoteundirected off
@end multitable
@node BRE syntax
@section Overview of basic regular expression syntax
Here is a brief description
of regular expression syntax as used in @command{sed}.
@table @code
@item @var{char}
A single ordinary character matches itself.
@item *
@cindex GNU extensions, to basic regular expressions
Matches a sequence of zero or more instances of matches for the
preceding regular expression, which must be an ordinary character, a
special character preceded by @code{\}, a @code{.}, a grouped regexp
(see below), or a bracket expression. As a GNU extension, a
postfixed regular expression can also be followed by @code{*}; for
example, @code{a**} is equivalent to @code{a*}. POSIX
1003.1-2001 says that @code{*} stands for itself when it appears at
the start of a regular expression or subexpression, but many
non-GNU implementations do not support this and portable
scripts should instead use @code{\*} in these contexts.
@item .
Matches any character, including newline.
@item ^
Matches the null string at beginning of the pattern space, i.e. what
appears after the circumflex must appear at the beginning of the
pattern space.
In most scripts, pattern space is initialized to the content of each
line (@pxref{Execution Cycle, , How @code{sed} works}). So, it is a
useful simplification to think of @code{^#include} as matching only
lines where @samp{#include} is the first thing on the line---if there is
any preceding space, for example, the match fails. This simplification is
valid as long as the original content of pattern space is not modified,
for example with an @code{s} command.
@code{^} acts as a special character only at the beginning of the
regular expression or subexpression (that is, after @code{\(} or
@code{\|}). Portable scripts should avoid @code{^} at the beginning of
a subexpression, though, as POSIX allows implementations that
treat @code{^} as an ordinary character in that context.
@item $
It is the same as @code{^}, but refers to end of pattern space.
@code{$} also acts as a special character only at the end
of the regular expression or subexpression (that is, before @code{\)}
or @code{\|}), and its use at the end of a subexpression is not
portable.
@item [@var{list}]
@itemx [^@var{list}]
Matches any single character in @var{list}: for example,
@code{[aeiou]} matches all vowels. A list may include
sequences like @code{@var{char1}-@var{char2}}, which
matches any character between (inclusive) @var{char1}
and @var{char2}.
@xref{Character Classes and Bracket Expressions}.
@item \+
@cindex GNU extensions, to basic regular expressions
As @code{*}, but matches one or more. It is a GNU extension.
@item \?
@cindex GNU extensions, to basic regular expressions
As @code{*}, but only matches zero or one. It is a GNU extension.
@item \@{@var{i}\@}
As @code{*}, but matches exactly @var{i} sequences (@var{i} is a
decimal integer; for portability, keep it between 0 and 255
inclusive).
@item \@{@var{i},@var{j}\@}
Matches between @var{i} and @var{j}, inclusive, sequences.
@item \@{@var{i},\@}
Matches more than or equal to @var{i} sequences.
@item \(@var{regexp}\)
Groups the inner @var{regexp} as a whole, this is used to:
@itemize @bullet
@item
@cindex GNU extensions, to basic regular expressions
Apply postfix operators, like @code{\(abcd\)*}:
this will search for zero or more whole sequences
of @samp{abcd}, while @code{abcd*} would search
for @samp{abc} followed by zero or more occurrences
of @samp{d}. Note that support for @code{\(abcd\)*} is
required by POSIX 1003.1-2001, but many non-GNU
implementations do not support it and hence it is not universally
portable.
@item
Use back references (see below).
@end itemize
@item @var{regexp1}\|@var{regexp2}
@cindex GNU extensions, to basic regular expressions
Matches either @var{regexp1} or @var{regexp2}. Use
parentheses to use complex alternative regular expressions.
The matching process tries each alternative in turn, from
left to right, and the first one that succeeds is used.
It is a GNU extension.
@item @var{regexp1}@var{regexp2}
Matches the concatenation of @var{regexp1} and @var{regexp2}.
Concatenation binds more tightly than @code{\|}, @code{^}, and
@code{$}, but less tightly than the other regular expression
operators.
@item \@var{digit}
Matches the @var{digit}-th @code{\(@dots{}\)} parenthesized
subexpression in the regular expression. This is called a @dfn{back
reference}. Subexpressions are implicitly numbered by counting
occurrences of @code{\(} left-to-right.
@item \n
Matches the newline character.
@item \@var{char}
Matches @var{char}, where @var{char} is one of @code{$},
@code{*}, @code{.}, @code{[}, @code{\}, or @code{^}.
Note that the only C-like
backslash sequences that you can portably assume to be
interpreted are @code{\n} and @code{\\}; in particular
@code{\t} is not portable, and matches a @samp{t} under most
implementations of @command{sed}, rather than a tab character.
@end table
@cindex Greedy regular expression matching
Note that the regular expression matcher is greedy, i.e., matches
are attempted from left to right and, if two or more matches are
possible starting at the same character, it selects the longest.
@noindent
Examples:
@table @samp
@item abcdef
Matches @samp{abcdef}.
@item a*b
Matches zero or more @samp{a}s followed by a single
@samp{b}. For example, @samp{b} or @samp{aaaaab}.
@item a\?b
Matches @samp{b} or @samp{ab}.
@item a\+b\+
Matches one or more @samp{a}s followed by one or more
@samp{b}s: @samp{ab} is the shortest possible match, but
other examples are @samp{aaaab} or @samp{abbbbb} or
@samp{aaaaaabbbbbbb}.
@item .*
@itemx .\+
These two both match all the characters in a string;
however, the first matches every string (including the empty
string), while the second matches only strings containing
at least one character.
@item ^main.*(.*)
This matches a string starting with @samp{main},
followed by an opening and closing
parenthesis. The @samp{n}, @samp{(} and @samp{)} need not
be adjacent.
@item ^#
This matches a string beginning with @samp{#}.
@item \\$
This matches a string ending with a single backslash. The
regexp contains two backslashes for escaping.
@item \$
Instead, this matches a string consisting of a single dollar sign,
because it is escaped.
@item [a-zA-Z0-9]
In the C locale, this matches any ASCII letters or digits.
@item [^ @kbd{@key{TAB}}]\+
(Here @kbd{@key{TAB}} stands for a single tab character.)
This matches a string of one or more
characters, none of which is a space or a tab.
Usually this means a word.
@item ^\(.*\)\n\1$
This matches a string consisting of two equal substrings separated by
a newline.
@item .\@{9\@}A$
This matches nine characters followed by an @samp{A} at the end of a line.
@item ^.\@{15\@}A
This matches the start of a string that contains 16 characters,
the last of which is an @samp{A}.
@end table
@node ERE syntax
@section Overview of extended regular expression syntax
@cindex Extended regular expressions, syntax
The only difference between basic and extended regular expressions is in
the behavior of a few characters: @samp{?}, @samp{+}, parentheses,
braces (@samp{@{@}}), and @samp{|}. While basic regular expressions
require these to be escaped if you want them to behave as special
characters, when using extended regular expressions you must escape
them if you want them @emph{to match a literal character}. @samp{|}
is special here because @samp{\|} is a GNU extension -- standard
basic regular expressions do not provide its functionality.
@noindent
Examples:
@table @code
@item abc?
becomes @samp{abc\?} when using extended regular expressions. It matches
the literal string @samp{abc?}.
@item c\+
becomes @samp{c+} when using extended regular expressions. It matches
one or more @samp{c}s.
@item a\@{3,\@}
becomes @samp{a@{3,@}} when using extended regular expressions. It matches
three or more @samp{a}s.
@item \(abc\)\@{2,3\@}
becomes @samp{(abc)@{2,3@}} when using extended regular expressions. It
matches either @samp{abcabc} or @samp{abcabcabc}.
@item \(abc*\)\1
becomes @samp{(abc*)\1} when using extended regular expressions.
Backreferences must still be escaped when using extended regular
expressions.
@item a\|b
becomes @samp{a|b} when using extended regular expressions. It matches
@samp{a} or @samp{b}.
@end table
@node Character Classes and Bracket Expressions
@section Character Classes and Bracket Expressions
@c The 'character class' section is shamelessly copied from grep's manual.
@cindex bracket expression
@cindex character class
A @dfn{bracket expression} is a list of characters enclosed by @samp{[} and
@samp{]}.
It matches any single character in that list;
if the first character of the list is the caret @samp{^},
then it matches any character @strong{not} in the list.
For example, the following command replaces the strings
@samp{gray} or @samp{grey} with @samp{blue}:
@example
sed 's/gr[ae]y/blue/'
@end example
@c TODO: fix 'ref' to look good in both HTML and PDF
Bracket expressions can be used in both
@ref{BRE syntax,,basic} and @ref{ERE syntax,,extended}
regular expressions (that is, with or without the @option{-E}/@option{-r}
options).
@cindex range expression
Within a bracket expression, a @dfn{range expression} consists of two
characters separated by a hyphen.
It matches any single character that
sorts between the two characters, inclusive.
In the default C locale, the sorting sequence is the native character
order; for example, @samp{[a-d]} is equivalent to @samp{[abcd]}.
Finally, certain named classes of characters are predefined within
bracket expressions, as follows.
These named classes must be used @emph{inside} brackets
themselves. Correct usage:
@example
$ echo 1 | sed 's/[[:digit:]]/X/'
X
@end example
Incorrect usage is rejected by newer @command{sed} versions.
Older versions accepted it but treated it as a single bracket expression
(which is equivalent to @samp{[dgit:]},
that is, only the characters @var{d/g/i/t/:}):
@example
# current GNU sed versions - incorrect usage rejected
$ echo 1 | sed 's/[:digit:]/X/'
sed: character class syntax is [[:space:]], not [:space:]
# older GNU sed versions
$ echo 1 | sed 's/[:digit:]/X/'
1
@end example
@cindex classes of characters
@cindex character classes
@cindex named character classes
@table @samp
@item [:alnum:]
@opindex alnum @r{character class}
@cindex alphanumeric characters
Alphanumeric characters:
@samp{[:alpha:]} and @samp{[:digit:]}; in the @samp{C} locale and ASCII
character encoding, this is the same as @samp{[0-9A-Za-z]}.
@item [:alpha:]
@opindex alpha @r{character class}
@cindex alphabetic characters
Alphabetic characters:
@samp{[:lower:]} and @samp{[:upper:]}; in the @samp{C} locale and ASCII
character encoding, this is the same as @samp{[A-Za-z]}.
@item [:blank:]
@opindex blank @r{character class}
@cindex blank characters
Blank characters:
space and tab.
@item [:cntrl:]
@opindex cntrl @r{character class}
@cindex control characters
Control characters.
In ASCII, these characters have octal codes 000
through 037, and 177 (DEL).
In other character sets, these are
the equivalent characters, if any.
@item [:digit:]
@opindex digit @r{character class}
@cindex digit characters
@cindex numeric characters
Digits: @code{0 1 2 3 4 5 6 7 8 9}.
@item [:graph:]
@opindex graph @r{character class}
@cindex graphic characters
Graphical characters:
@samp{[:alnum:]} and @samp{[:punct:]}.
@item [:lower:]
@opindex lower @r{character class}
@cindex lower-case letters
Lower-case letters; in the @samp{C} locale and ASCII character
encoding, this is
@code{a b c d e f g h i j k l m n o p q r s t u v w x y z}.
@item [:print:]
@opindex print @r{character class}
@cindex printable characters
Printable characters:
@samp{[:alnum:]}, @samp{[:punct:]}, and space.
@item [:punct:]
@opindex punct @r{character class}
@cindex punctuation characters
Punctuation characters; in the @samp{C} locale and ASCII character
encoding, this is
@code{!@: " # $ % & ' ( ) * + , - .@: / : ; < = > ?@: @@ [ \ ] ^ _ ` @{ | @} ~}.
@item [:space:]
@opindex space @r{character class}
@cindex space characters
@cindex whitespace characters
Space characters: in the @samp{C} locale, this is
tab, newline, vertical tab, form feed, carriage return, and space.
@item [:upper:]
@opindex upper @r{character class}
@cindex upper-case letters
Upper-case letters: in the @samp{C} locale and ASCII character
encoding, this is
@code{A B C D E F G H I J K L M N O P Q R S T U V W X Y Z}.
@item [:xdigit:]
@opindex xdigit @r{character class}
@cindex xdigit class
@cindex hexadecimal digits
Hexadecimal digits:
@code{0 1 2 3 4 5 6 7 8 9 A B C D E F a b c d e f}.
@end table
Note that the brackets in these class names are
part of the symbolic names, and must be included in addition to
the brackets delimiting the bracket expression.
Most meta-characters lose their special meaning inside bracket expressions:
@table @samp
@item ]
ends the bracket expression if it's not the first list item.
So, if you want to make the @samp{]} character a list item,
you must put it first.
@item -
represents the range if it's not first or last in a list or the ending point
of a range.
@item ^
represents the characters not in the list.
If you want to make the @samp{^}
character a list item, place it anywhere but first.
@end table
TODO: incorporate this paragraph (copied verbatim from BRE section).
@cindex @code{POSIXLY_CORRECT} behavior, bracket expressions
The characters @code{$}, @code{*}, @code{.}, @code{[}, and @code{\}
are normally not special within @var{list}. For example, @code{[\*]}
matches either @samp{\} or @samp{*}, because the @code{\} is not
special here. However, strings like @code{[.ch.]}, @code{[=a=]}, and
@code{[:space:]} are special within @var{list} and represent collating
symbols, equivalence classes, and character classes, respectively, and
@code{[} is therefore special within @var{list} when it is followed by
@code{.}, @code{=}, or @code{:}. Also, when not in
@env{POSIXLY_CORRECT} mode, special escapes like @code{\n} and
@code{\t} are recognized within @var{list}. @xref{Escapes}.
@c ********
@c TODO: improve explanation about collation classes and equivalence classes
@c perhaps dedicate a section to Locales ??
@table @samp
@item [.
represents the open collating symbol.
@item .]
represents the close collating symbol.
@item [=
represents the open equivalence class.
@item =]
represents the close equivalence class.
@item [:
represents the open character class symbol, and should be followed by a
valid character class name.
@item :]
represents the close character class symbol.
@end table
@node regexp extensions
@section regular expression extensions
The following sequences have special meaning inside regular expressions
(used in @ref{Regexp Addresses,,addresses} and the @code{s} command).
These can be used in both
@ref{BRE syntax,,basic} and @ref{ERE syntax,,extended}
regular expressions (that is, with or without the @option{-E}/@option{-r}
options).
@table @code
@item \w
Matches any ``word'' character. A ``word'' character is any
letter or digit or the underscore character.
@example
$ echo "abc %-= def." | sed 's/\w/X/g'
XXX %-= XXX.
@end example
@item \W
Matches any ``non-word'' character.
@example
$ echo "abc %-= def." | sed 's/\W/X/g'
abcXXXXXdefX
@end example
@item \b
Matches a word boundary; that is it matches if the character
to the left is a ``word'' character and the character to the
right is a ``non-word'' character, or vice-versa.
@example
$ echo "abc %-= def." | sed 's/\b/X/g'
XabcX %-= XdefX.
@end example
@item \B
Matches everywhere but on a word boundary; that is it matches
if the character to the left and the character to the right
are either both ``word'' characters or both ``non-word''
characters.
@example
$ echo "abc %-= def." | sed 's/\B/X/g'
aXbXc X%X-X=X dXeXf.X
@end example
@item \s
Matches whitespace characters (spaces and tabs).
Newlines embedded in the pattern/hold spaces will also match:
@example
$ echo "abc %-= def." | sed 's/\s/X/g'
abcX%-=Xdef.
@end example
@item \S
Matches non-whitespace characters.
@example
$ echo "abc %-= def." | sed 's/\S/X/g'
XXX XXX XXXX
@end example
@item \<
Matches the beginning of a word.
@example
$ echo "abc %-= def." | sed 's/\</X/g'
Xabc %-= Xdef.
@end example
@item \>
Matches the end of a word.
@example
$ echo "abc %-= def." | sed 's/\>/X/g'
abcX %-= defX.
@end example
@item \`
Matches only at the start of pattern space. This is different
from @code{^} in multi-line mode.
Compare the following two examples:
@example
$ printf "a\nb\nc\n" | sed 'N;N;s/^/X/gm'
Xa
Xb
Xc
$ printf "a\nb\nc\n" | sed 'N;N;s/\`/X/gm'
Xa
b
c
@end example
@item \'
Matches only at the end of pattern space. This is different
from @code{$} in multi-line mode.
@end table
@node Back-references and Subexpressions
@section Back-references and Subexpressions
@cindex subexpression
@cindex back-reference
@dfn{back-references} are regular expression commands which refer to a
previous part of the matched regular expression. Back-references are
specified with backslash and a single digit (e.g. @samp{\1}). The
part of the regular expression they refer to is called a
@dfn{subexpression}, and is designated with parentheses.
Back-references and subexpressions are used in two cases: in the
regular expression search pattern, and in the @var{replacement} part
of the @command{s} command (@pxref{Regexp Addresses,,Regular
Expression Addresses} and @ref{The "s" Command}).
In a regular expression pattern, back-references are used to match
the same content as a previously matched subexpression. In the
following example, the subexpression is @samp{.} - any single
character (being surrounded by parentheses makes it a
subexpression). The back-reference @samp{\1} asks to match the same
content (same character) as the sub-expression.
The command below matches words starting with any character,
followed by the letter @samp{o}, followed by the same character as the
first.
@example
$ sed -E -n '/^(.)o\1$/p' /usr/share/dict/words
bob
mom
non
pop
sos
tot
wow
@end example
Multiple subexpressions are automatically numbered from
left-to-right. This command searches for 6-letter
palindromes (the first three letters are 3 subexpressions,
followed by 3 back-references in reverse order):
@example
$ sed -E -n '/^(.)(.)(.)\3\2\1$/p' /usr/share/dict/words
redder
@end example
In the @command{s} command, back-references can be
used in the @var{replacement} part to refer back to subexpressions in
the @var{regexp} part.
The following example uses two subexpressions in the regular
expression to match two space-separated words. The back-references in
the @var{replacement} part prints the words in a different order:
@example
$ echo "James Bond" | sed -E 's/(.*) (.*)/The name is \2, \1 \2./'
The name is Bond, James Bond.
@end example
When used with alternation, if the group does not participate in the
match then the back-reference makes the whole match fail. For
example, @samp{a(.)|b\1} will not match @samp{ba}. When multiple
regular expressions are given with @option{-e} or from a file
(@samp{-f @var{file}}), back-references are local to each expression.
@node Escapes
@section Escape Sequences - specifying special characters
@cindex GNU extensions, special escapes
Until this chapter, we have only encountered escapes of the form
@samp{\^}, which tell @command{sed} not to interpret the circumflex
as a special character, but rather to take it literally. For
example, @samp{\*} matches a single asterisk rather than zero
or more backslashes.
@cindex @code{POSIXLY_CORRECT} behavior, escapes
This chapter introduces another kind of escape@footnote{All
the escapes introduced here are GNU
extensions, with the exception of @code{\n}. In basic regular
expression mode, setting @code{POSIXLY_CORRECT} disables them inside
bracket expressions.}---that
is, escapes that are applied to a character or sequence of characters
that ordinarily are taken literally, and that @command{sed} replaces
with a special character. This provides a way
of encoding non-printable characters in patterns in a visible manner.
There is no restriction on the appearance of non-printing characters
in a @command{sed} script but when a script is being prepared in the
shell or by text editing, it is usually easier to use one of
the following escape sequences than the binary character it
represents:
The list of these escapes is:
@table @code
@item \a
Produces or matches a @sc{bel} character, that is an ``alert'' (@sc{ascii} 7).
@item \f
Produces or matches a form feed (@sc{ascii} 12).
@item \n
Produces or matches a newline (@sc{ascii} 10).
@item \r
Produces or matches a carriage return (@sc{ascii} 13).
@item \t
Produces or matches a horizontal tab (@sc{ascii} 9).
@item \v
Produces or matches a so called ``vertical tab'' (@sc{ascii} 11).
@item \c@var{x}
Produces or matches @kbd{@sc{Control}-@var{x}}, where @var{x} is
any character. The precise effect of @samp{\c@var{x}} is as follows:
if @var{x} is a lower case letter, it is converted to upper case.
Then bit 6 of the character (hex 40) is inverted. Thus @samp{\cz} becomes
hex 1A, but @samp{\c@{} becomes hex 3B, while @samp{\c;} becomes hex 7B.
@item \d@var{xxx}
Produces or matches a character whose decimal @sc{ascii} value is @var{xxx}.
@item \o@var{xxx}
Produces or matches a character whose octal @sc{ascii} value is @var{xxx}.
@item \x@var{xx}
Produces or matches a character whose hexadecimal @sc{ascii} value is @var{xx}.
@end table
@samp{\b} (backspace) was omitted because of the conflict with
the existing ``word boundary'' meaning.
@subsection Escaping Precedence
@value{SSED} processes escape sequences @emph{before} passing
the text onto the regular-expression matching of the @command{s///} command
and Address matching. Thus the following two commands are equivalent
(@samp{0x5e} is the hexadecimal @sc{ascii} value of the character @samp{^}):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo 'a^c' | sed 's/^/b/'
ba^c
$ echo 'a^c' | sed 's/\x5e/b/'
ba^c
@end group
@end example
@codequoteundirected off
@codequotebacktick off
As are the following (@samp{0x5b},@samp{0x5d} are the hexadecimal
@sc{ascii} values of @samp{[},@samp{]}, respectively):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo abc | sed 's/[a]/x/'
Xbc
$ echo abc | sed 's/\x5ba\x5d/x/'
Xbc
@end group
@end example
@codequoteundirected off
@codequotebacktick off
However it is recommended to avoid such special characters
due to unexpected edge-cases. For example, the following
are not equivalent:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ echo 'a^c' | sed 's/\^/b/'
abc
$ echo 'a^c' | sed 's/\\\x5e/b/'
a^c
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@c also: this fails in different places:
@c $ sed 's/[//'
@c sed: -e expression #1, char 5: unterminated `s' command
@c $ sed 's/\x5b//'
@c sed: -e expression #1, char 8: Invalid regular expression
@c
@c which is OK but confusing to explain why (the first
@c fails in compile.c:snarf_char_class while the second
@c is passed to the regex engine and then fails).
@node Locale Considerations
@section Multibyte characters and Locale Considerations
@value{SSED} processes valid multibyte characters in multibyte locales
(e.g. @code{UTF-8}). @footnote{Some regexp edge-cases depends on the
operating system and libc implementation. The examples shown are known
to work as-expected on GNU/Linux systems using glibc.}
@noindent The following example uses the Greek letter Capital Sigma
(@value{ucsigma},
Unicode code point @code{0x03A3}). In a @code{UTF-8} locale,
@command{sed} correctly processes the Sigma as one character despite
it being 2 octets (bytes):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ locale | grep LANG
LANG=en_US.UTF-8
$ printf 'a\u03A3b'
a@value{ucsigma}b
$ printf 'a\u03A3b' | sed 's/./X/g'
XXX
$ printf 'a\u03A3b' | od -tx1 -An
61 ce a3 62
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@noindent
To force @command{sed} to process octets separately, use the @code{C} locale
(also known as the @code{POSIX} locale):
@codequoteundirected on
@codequotebacktick on
@example
$ printf 'a\u03A3b' | LC_ALL=C sed 's/./X/g'
XXXX
@end example
@codequoteundirected off
@codequotebacktick off
@subsection Invalid multibyte characters
@command{sed}'s regular expressions @emph{do not} match
invalid multibyte sequences in a multibyte locale.
@noindent
In the following examples, the ascii value @code{0xCE} is
an incomplete multibyte character (shown here as @value{unicodeFFFD}).
The regular expression @samp{.} does not match it:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf 'a\xCEb\n'
a@value{unicodeFFFD}e
$ printf 'a\xCEb\n' | sed 's/./X/g'
X@value{unicodeFFFD}X
$ printf 'a\xCEc\n' | sed 's/./X/g' | od -tx1c -An
58 ce 58 0a
X X \n
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@noindent Similarly, the 'catch-all' regular expression @samp{.*} does not
match the entire line:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf 'a\xCEc\n' | sed 's/.*//' | od -tx1c -An
ce 63 0a
c \n
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@noindent
@value{SSED} offers the special @command{z} command to clear the
current pattern space regardless of invalid multibyte characters
(i.e. it works like @code{s/.*//} but also removes invalid multibyte
characters):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf 'a\xCEc\n' | sed 'z' | od -tx1c -An
0a
\n
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@noindent Alternatively, force the @code{C} locale to process
each octet separately (every octet is a valid character in the @code{C}
locale):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf 'a\xCEc\n' | LC_ALL=C sed 's/.*//' | od -tx1c -An
0a
\n
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@command{sed}'s inability to process invalid multibyte characters
can be used to detect such invalid sequences in a file.
In the following examples, the @code{\xCE\xCE} is an invalid
multibyte sequence, while @code{\xCE\A3} is a valid multibyte sequence
(of the Greek Sigma character).
@noindent
The following @command{sed} program removes all valid
characters using @code{s/.//g}. Any content left in the pattern space
(the invalid characters) are added to the hold space using the
@code{H} command. On the last line (@code{$}), the hold space is retrieved
(@code{x}), newlines are removed (@code{s/\n//g}), and any remaining
octets are printed unambiguously (@code{l}). Thus, any invalid
multibyte sequences are printed as octal values:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf 'ab\nc\n\xCE\xCEde\n\xCE\xA3f\n' > invalid.txt
$ cat invalid.txt
ab
c
@value{unicodeFFFD}@value{unicodeFFFD}de
@value{ucsigma}f
$ sed -n 's/.//g ; H ; $@{x;s/\n//g;l@}' invalid.txt
\316\316$
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@noindent With a few more commands, @command{sed} can print
the exact line number corresponding to each invalid characters (line 3).
These characters can then be removed by forcing the @code{C} locale
and using octal escape sequences:
@codequoteundirected on
@codequotebacktick on
@example
$ sed -n 's/.//g;=;l' invalid.txt | paste - - | awk '$2!="$"'
3 \316\316$
$ LC_ALL=C sed '3s/\o316\o316//' invalid.txt > fixed.txt
@end example
@codequoteundirected off
@codequotebacktick off
@subsection Upper/Lower case conversion
@value{SSED}'s substitute command (@code{s}) supports upper/lower
case conversions using @code{\U},@code{\L} codes.
These conversions support multibyte characters:
@codequoteundirected on
@codequotebacktick on
@example
$ printf 'ABC\u03a3\n'
ABC@value{ucsigma}
$ printf 'ABC\u03a3\n' | sed 's/.*/\L&/'
abc@value{lcsigma}
@end example
@codequoteundirected off
@codequotebacktick off
@noindent
@xref{The "s" Command}.
@subsection Multibyte regexp character classes
@c TODO: fix following paragraphs (copied verbatim from 'bracket
@c expression' section).
In other locales, the sorting sequence is not specified, and
@samp{[a-d]} might be equivalent to @samp{[abcd]} or to
@samp{[aBbCcDd]}, or it might fail to match any character, or the set of
characters that it matches might even be erratic.
To obtain the traditional interpretation
of bracket expressions, you can use the @samp{C} locale by setting the
@env{LC_ALL} environment variable to the value @samp{C}.
@example
# TODO: is there any real-world system/locale where 'A'
# is replaced by '-' ?
$ echo A | sed 's/[a-z]/-/'
A
@end example
Their interpretation depends on the @env{LC_CTYPE} locale;
for example, @samp{[[:alnum:]]} means the character class of numbers and letters
in the current locale.
TODO: show example of collation
@codequoteundirected on
@codequotebacktick on
@example
# TODO: this works on glibc systems, not on musl-libc/freebsd/macosx.
$ printf 'cliché\n' | LC_ALL=fr_FR.utf8 sed 's/[[=e=]]/X/g'
clichX
@end example
@codequoteundirected off
@codequotebacktick off
@node advanced sed
@chapter Advanced @command{sed}: cycles and buffers
@menu
* Execution Cycle:: How @command{sed} works
* Hold and Pattern Buffers::
* Multiline techniques:: Using D,G,H,N,P to process multiple lines
* Branching and flow control::
@end menu
@node Execution Cycle
@section How @command{sed} Works
@cindex Buffer spaces, pattern and hold
@cindex Spaces, pattern and hold
@cindex Pattern space, definition
@cindex Hold space, definition
@command{sed} maintains two data buffers: the active @emph{pattern} space,
and the auxiliary @emph{hold} space. Both are initially empty.
@command{sed} operates by performing the following cycle on each
line of input: first, @command{sed} reads one line from the input
stream, removes any trailing newline, and places it in the pattern space.
Then commands are executed; each command can have an address associated
to it: addresses are a kind of condition code, and a command is only
executed if the condition is verified before the command is to be
executed.
When the end of the script is reached, unless the @option{-n} option
is in use, the contents of pattern space are printed out to the output
stream, adding back the trailing newline if it was removed.@footnote{Actually,
if @command{sed} prints a line without the terminating newline, it will
nevertheless print the missing newline as soon as more text is sent to
the same output stream, which gives the ``least expected surprise''
even though it does not make commands like @samp{sed -n p} exactly
identical to @command{cat}.} Then the next cycle starts for the next
input line.
Unless special commands (like @samp{D}) are used, the pattern space is
deleted between two cycles. The hold space, on the other hand, keeps
its data between cycles (see commands @samp{h}, @samp{H}, @samp{x},
@samp{g}, @samp{G} to move data between both buffers).
@node Hold and Pattern Buffers
@section Hold and Pattern Buffers
TODO
@node Multiline techniques
@section Multiline techniques - using D,G,H,N,P to process multiple lines
Multiple lines can be processed as one buffer using the
@code{D},@code{G},@code{H},@code{N},@code{P}. They are similar to
their lowercase counterparts (@code{d},@code{g},
@code{h},@code{n},@code{p}), except that these commands append or
subtract data while respecting embedded newlines - allowing adding and
removing lines from the pattern and hold spaces.
They operate as follows:
@table @code
@item D
@emph{deletes} line from the pattern space until the first newline,
and restarts the cycle.
@item G
@emph{appends} line from the hold space to the pattern space, with a
newline before it.
@item H
@emph{appends} line from the pattern space to the hold space, with a
newline before it.
@item N
@emph{appends} line from the input file to the pattern space.
@item P
@emph{prints} line from the pattern space until the first newline.
@end table
The following example illustrates the operation of @code{N} and
@code{D} commands:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 6 | sed -n 'N;l;D'
1\n2$
2\n3$
3\n4$
4\n5$
5\n6$
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@enumerate
@item
@command{sed} starts by reading the first line into the pattern space
(i.e. @samp{1}).
@item
At the beginning of every cycle, the @code{N}
command appends a newline and the next line to the pattern space
(i.e. @samp{1}, @samp{\n}, @samp{2} in the first cycle).
@item
The @code{l} command prints the content of the pattern space
unambiguously.
@item
The @code{D} command then removes the content of pattern
space up to the first newline (leaving @samp{2} at the end of
the first cycle).
@item
At the next cycle the @code{N} command appends a
newline and the next input line to the pattern space
(e.g. @samp{2}, @samp{\n}, @samp{3}).
@end enumerate
@cindex processing paragraphs
@cindex paragraphs, processing
A common technique to process blocks of text such as paragraphs
(instead of line-by-line) is using the following construct:
@codequoteundirected on
@codequotebacktick on
@example
sed '/./@{H;$!d@} ; x ; s/REGEXP/REPLACEMENT/'
@end example
@codequoteundirected off
@codequotebacktick off
@enumerate
@item
The first expression, @code{/./@{H;$!d@}} operates on all non-empty lines,
and adds the current line (in the pattern space) to the hold space.
On all lines except the last, the pattern space is deleted and the cycle is
restarted.
@item
The other expressions @code{x} and @code{s} are executed only on empty
lines (i.e. paragraph separators). The @code{x} command fetches the
accumulated lines from the hold space back to the pattern space. The
@code{s///} command then operates on all the text in the paragraph
(including the embedded newlines).
@end enumerate
The following example demonstrates this technique:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ cat input.txt
a a a aa aaa
aaaa aaaa aa
aaaa aaa aaa
bbbb bbb bbb
bb bb bbb bb
bbbbbbbb bbb
ccc ccc cccc
cccc ccccc c
cc cc cc cc
$ sed '/./@{H;$!d@} ; x ; s/^/\nSTART-->/ ; s/$/\n<--END/' input.txt
START-->
a a a aa aaa
aaaa aaaa aa
aaaa aaa aaa
<--END
START-->
bbbb bbb bbb
bb bb bbb bb
bbbbbbbb bbb
<--END
START-->
ccc ccc cccc
cccc ccccc c
cc cc cc cc
<--END
@end group
@end example
@codequoteundirected off
@codequotebacktick off
For more annotated examples, @pxref{Text search across multiple lines}
and @ref{Line length adjustment}.
@node Branching and flow control
@section Branching and Flow Control
The branching commands @code{b}, @code{t}, and @code{T} enable
changing the flow of @command{sed} programs.
By default, @command{sed} reads an input line into the pattern buffer,
then continues to processes all commands in order.
Commands without addresses affect all lines.
Commands with addresses affect only matching lines.
@xref{Execution Cycle} and @ref{Addresses overview}.
@command{sed} does not support a typical @code{if/then} construct.
Instead, some commands can be used as conditionals or to change the
default flow control:
@table @code
@item d
delete (clears) the current pattern space,
and restart the program cycle without processing the rest of the commands
and without printing the pattern space.
@item D
delete the contents of the pattern space @emph{up to the first newline},
and restart the program cycle without processing the rest of
the commands and without printing the pattern space.
@item [addr]X
@itemx [addr]@{ X ; X ; X @}
@item /regexp/X
@item /regexp/@{ X ; X ; X @}
Addresses and regular expressions can be used as an @code{if/then}
conditional: If @var{[addr]} matches the current pattern space,
execute the command(s).
For example: The command @code{/^#/d} means:
@emph{if} the current pattern matches the regular expression @code{^#} (a line
starting with a hash), @emph{then} execute the @code{d} command:
delete the line without printing it, and restart the program cycle
immediately.
@item b
branch unconditionally (that is: always jump to a label, skipping
or repeating other commands, without restarting a new cycle). Combined
with an address, the branch can be conditionally executed on matched
lines.
@item t
branch conditionally (that is: jump to a label) @emph{only if} a
@code{s///} command has succeeded since the last input line was read
or another conditional branch was taken.
@item T
similar but opposite to the @code{t} command: branch only if
there has been @emph{no} successful substitutions since the last
input line was read.
@end table
The following two @command{sed} programs are equivalent. The first
(contrived) example uses the @code{b} command to skip the @code{s///}
command on lines containing @samp{1}. The second example uses an
address with negation (@samp{!}) to perform substitution only on
desired lines. The @code{y///} command is still executed on all
lines:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ printf '%s\n' a1 a2 a3 | sed -E '/1/bx ; s/a/z/ ; :x ; y/123/456/'
a4
z5
z6
$ printf '%s\n' a1 a2 a3 | sed -E '/1/!s/a/z/ ; y/123/456/'
a4
z5
z6
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@subsection Branching and Cycles
@cindex labels
@cindex omitting labels
@cindex cycle, restarting
@cindex restarting a cycle
The @code{b},@code{t} and @code{T} commands can be followed by a label
(typically a single letter). Labels are defined with a colon followed by
one or more letters (e.g. @samp{:x}). If the label is omitted the
branch commands restart the cycle. Note the difference between
branching to a label and restarting the cycle: when a cycle is
restarted, @command{sed} first prints the current content of the
pattern space, then reads the next input line into the pattern space;
Jumping to a label (even if it is at the beginning of the program)
does not print the pattern space and does not read the next input line.
The following program is a no-op. The @code{b} command (the only command
in the program) does not have a label, and thus simply restarts the cycle.
On each cycle, the pattern space is printed and the next input line is read:
@example
@group
$ seq 3 | sed b
1
2
3
@end group
@end example
@cindex infinite loop, branching
@cindex branching, infinite loop
The following example is an infinite-loop - it doesn't terminate and
doesn't print anything. The @code{b} command jumps to the @samp{x}
label, and a new cycle is never started:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 3 | sed ':x ; bx'
# The above command requires gnu sed (which supports additional
# commands following a label, without a newline). A portable equivalent:
# sed -e ':x' -e bx
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@cindex branching and n, N
@cindex n, and branching
@cindex N, and branching
Branching is often complemented with the @code{n} or @code{N} commands:
both commands read the next input line into the pattern space without waiting
for the cycle to restart. Before reading the next input line, @code{n}
prints the current pattern space then empties it, while @code{N}
appends a newline and the next input line to the pattern space.
Consider the following two examples:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ seq 3 | sed ':x ; n ; bx'
1
2
3
$ seq 3 | sed ':x ; N ; bx'
1
2
3
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@itemize
@item
Both examples do not inf-loop, despite never starting a new cycle.
@item
In the first example, the @code{n} commands first prints the content
of the pattern space, empties the pattern space then reads the next
input line.
@item
In the second example, the @code{N} commands appends the next input
line to the pattern space (with a newline). Lines are accumulated in
the pattern space until there are no more input lines to read, then
the @code{N} command terminates the @command{sed} program. When the
program terminates, the end-of-cycle actions are performed, and the
entire pattern space is printed.
@item
The second example requires @value{SSED},
because it uses the non-POSIX-standard behavior of @code{N}.
See the ``@code{N} command on the last line'' paragraph
in @ref{Reporting Bugs}.
@item
To further examine the difference between the two examples,
try the following commands:
@codequoteundirected on
@codequotebacktick on
@example
@group
printf '%s\n' aa bb cc dd | sed ':x ; n ; = ; bx'
printf '%s\n' aa bb cc dd | sed ':x ; N ; = ; bx'
printf '%s\n' aa bb cc dd | sed ':x ; n ; s/\n/***/ ; bx'
printf '%s\n' aa bb cc dd | sed ':x ; N ; s/\n/***/ ; bx'
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@end itemize
@subsection Branching example: joining lines
@cindex joining lines with branching
@cindex branching, joining lines
@cindex quoted-printable lines, joining
@cindex joining quoted-printable lines
@cindex t, joining lines with
@cindex b, joining lines with
@cindex b, versus t
@cindex t, versus b
As a real-world example of using branching, consider the case of
@uref{https://en.wikipedia.org/wiki/Quoted-printable,quoted-printable} files,
typically used to encode email messages.
In these files long lines are split and marked with a @dfn{soft line break}
consisting of a single @samp{=} character at the end of the line:
@example
@group
$ cat jaques.txt
All the wor=
ld's a stag=
e,
And all the=
men and wo=
men merely =
players:
They have t=
heir exits =
and their e=
ntrances;
And one man=
in his tim=
e plays man=
y parts.
@end group
@end example
The following program uses an address match @samp{/=$/} as a
conditional: If the current pattern space ends with a @samp{=}, it
reads the next input line using @code{N}, replaces all @samp{=}
characters which are followed by a newline, and unconditionally
branches (@code{b}) to the beginning of the program without restarting
a new cycle. If the pattern space does not ends with @samp{=}, the
default action is performed: the pattern space is printed and a new
cycle is started:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ sed ':x ; /=$/ @{ N ; s/=\n//g ; bx @}' jaques.txt
All the world's a stage,
And all the men and women merely players:
They have their exits and their entrances;
And one man in his time plays many parts.
@end group
@end example
@codequoteundirected off
@codequotebacktick off
Here's an alternative program with a slightly different approach: On
all lines except the last, @code{N} appends the line to the pattern
space. A substitution command then removes soft line breaks
(@samp{=} at the end of a line, i.e. followed by a newline) by replacing
them with an empty string.
@emph{if} the substitution was successful (meaning the pattern space contained
a line which should be joined), The conditional branch command @code{t} jumps
to the beginning of the program without completing or restarting the cycle.
If the substitution failed (meaning there were no soft line breaks),
The @code{t} command will @emph{not} branch. Then, @code{P} will
print the pattern space content until the first newline, and @code{D}
will delete the pattern space content until the first new line.
(To learn more about @code{N}, @code{P} and @code{D} commands
@pxref{Multiline techniques}).
@codequoteundirected on
@codequotebacktick on
@example
@group
$ sed ':x ; $!N ; s/=\n// ; tx ; P ; D' jaques.txt
All the world's a stage,
And all the men and women merely players:
They have their exits and their entrances;
And one man in his time plays many parts.
@end group
@end example
@codequoteundirected off
@codequotebacktick off
For more line-joining examples @pxref{Joining lines}.
@node Examples
@chapter Some Sample Scripts
Here are some @command{sed} scripts to guide you in the art of mastering
@command{sed}.
@menu
Useful one-liners:
* Joining lines::
Some exotic examples:
* Centering lines::
* Increment a number::
* Rename files to lower case::
* Print bash environment::
* Reverse chars of lines::
* Text search across multiple lines::
* Line length adjustment::
* Adding a header to multiple files::
Emulating standard utilities:
* tac:: Reverse lines of files
* cat -n:: Numbering lines
* cat -b:: Numbering non-blank lines
* wc -c:: Counting chars
* wc -w:: Counting words
* wc -l:: Counting lines
* head:: Printing the first lines
* tail:: Printing the last lines
* uniq:: Make duplicate lines unique
* uniq -d:: Print duplicated lines of input
* uniq -u:: Remove all duplicated lines
* cat -s:: Squeezing blank lines
@end menu
@node Joining lines
@section Joining lines
This section uses @code{N}, @code{D} and @code{P} commands to process
multiple lines, and the @code{b} and @code{t} commands for branching.
@xref{Multiline techniques} and @ref{Branching and flow control}.
Join specific lines (e.g. if lines 2 and 3 need to be joined):
@codequoteundirected on
@codequotebacktick on
@example
$ cat lines.txt
hello
hel
lo
hello
$ sed '2@{N;s/\n//;@}' lines.txt
hello
hello
hello
@end example
@codequoteundirected off
@codequotebacktick off
Join backslash-continued lines:
@codequoteundirected on
@codequotebacktick on
@example
$ cat 1.txt
this \
is \
a \
long \
line
and another \
line
$ sed -e ':x /\\$/ @{ N; s/\\\n//g ; bx @}' 1.txt
this is a long line
and another line
#TODO: The above requires gnu sed.
# non-gnu seds need newlines after ':' and 'b'
@end example
@codequoteundirected off
@codequotebacktick off
Join lines that start with whitespace (e.g SMTP headers):
@codequoteundirected on
@codequotebacktick on
@example
@group
$ cat 2.txt
Subject: Hello
World
Content-Type: multipart/alternative;
boundary=94eb2c190cc6370f06054535da6a
Date: Tue, 3 Jan 2017 19:41:16 +0000 (GMT)
Authentication-Results: mx.gnu.org;
dkim=pass header.i=@@gnu.org;
spf=pass
Message-ID: <abcdef@@gnu.org>
From: John Doe <jdoe@@gnu.org>
To: Jane Smith <jsmith@@gnu.org>
$ sed -E ':a ; $!N ; s/\n\s+/ / ; ta ; P ; D' 2.txt
Subject: Hello World
Content-Type: multipart/alternative; boundary=94eb2c190cc6370f06054535da6a
Date: Tue, 3 Jan 2017 19:41:16 +0000 (GMT)
Authentication-Results: mx.gnu.org; dkim=pass header.i=@@gnu.org; spf=pass
Message-ID: <abcdef@@gnu.org>
From: John Doe <jdoe@@gnu.org>
To: Jane Smith <jsmith@@gnu.org>
# A portable (non-gnu) variation:
# sed -e :a -e '$!N;s/\n */ /;ta' -e 'P;D'
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@node Centering lines
@section Centering Lines
This script centers all lines of a file on a 80 columns width.
To change that width, the number in @code{\@{@dots{}\@}} must be
replaced, and the number of added spaces also must be changed.
Note how the buffer commands are used to separate parts in
the regular expressions to be matched---this is a common
technique.
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
@group
# Put 80 spaces in the buffer
1 @{
x
s/^$/ /
s/^.*$/&&&&&&&&/
x
@}
@end group
@group
# delete leading and trailing spaces
y/@kbd{@key{TAB}}/ /
s/^ *//
s/ *$//
@end group
@group
# add a newline and 80 spaces to end of line
G
@end group
@group
# keep first 81 chars (80 + a newline)
s/^\(.\@{81\@}\).*$/\1/
@end group
@group
# \2 matches half of the spaces, which are moved to the beginning
s/^\(.*\)\n\(.*\)\2/\2\1/
@end group
@end example
@c end---------------------------------------------
@node Increment a number
@section Increment a Number
This script is one of a few that demonstrate how to do arithmetic
in @command{sed}. This is indeed possible,@footnote{@command{sed} guru Greg
Ubben wrote an implementation of the @command{dc} @sc{rpn} calculator!
It is distributed together with sed.} but must be done manually.
To increment one number you just add 1 to last digit, replacing
it by the following digit. There is one exception: when the digit
is a nine the previous digits must be also incremented until you
don't have a nine.
This solution by Bruno Haible is very clever and smart because
it uses a single buffer; if you don't have this limitation, the
algorithm used in @ref{cat -n, Numbering lines}, is faster.
It works by replacing trailing nines with an underscore, then
using multiple @code{s} commands to increment the last digit,
and then again substituting underscores with zeros.
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
/[^0-9]/ d
@group
# replace all trailing 9s by _ (any other character except digits, could
# be used)
:d
s/9\(_*\)$/_\1/
td
@end group
@group
# incr last digit only. The first line adds a most-significant
# digit of 1 if we have to add a digit.
@end group
@group
s/^\(_*\)$/1\1/; tn
s/8\(_*\)$/9\1/; tn
s/7\(_*\)$/8\1/; tn
s/6\(_*\)$/7\1/; tn
s/5\(_*\)$/6\1/; tn
s/4\(_*\)$/5\1/; tn
s/3\(_*\)$/4\1/; tn
s/2\(_*\)$/3\1/; tn
s/1\(_*\)$/2\1/; tn
s/0\(_*\)$/1\1/; tn
@end group
@group
:n
y/_/0/
@end group
@end example
@c end---------------------------------------------
@node Rename files to lower case
@section Rename Files to Lower Case
This is a pretty strange use of @command{sed}. We transform text, and
transform it to be shell commands, then just feed them to shell.
Don't worry, even worse hacks are done when using @command{sed}; I have
seen a script converting the output of @command{date} into a @command{bc}
program!
The main body of this is the @command{sed} script, which remaps the name
from lower to upper (or vice-versa) and even checks out
if the remapped name is the same as the original name.
Note how the script is parameterized using shell
variables and proper quoting.
@c start-------------------------------------------
@example
@group
#! /bin/sh
# rename files to lower/upper case...
#
# usage:
# move-to-lower *
# move-to-upper *
# or
# move-to-lower -R .
# move-to-upper -R .
#
@end group
@group
help()
@{
cat << eof
Usage: $0 [-n] [-r] [-h] files...
@end group
@group
-n do nothing, only see what would be done
-R recursive (use find)
-h this message
files files to remap to lower case
@end group
@group
Examples:
$0 -n * (see if everything is ok, then...)
$0 *
@end group
$0 -R .
@group
eof
@}
@end group
@group
apply_cmd='sh'
finder='echo "$@@" | tr " " "\n"'
files_only=
@end group
@group
while :
do
case "$1" in
-n) apply_cmd='cat' ;;
-R) finder='find "$@@" -type f';;
-h) help ; exit 1 ;;
*) break ;;
esac
shift
done
@end group
@group
if [ -z "$1" ]; then
echo Usage: $0 [-h] [-n] [-r] files...
exit 1
fi
@end group
@group
LOWER='abcdefghijklmnopqrstuvwxyz'
UPPER='ABCDEFGHIJKLMNOPQRSTUVWXYZ'
@end group
@group
case `basename $0` in
*upper*) TO=$UPPER; FROM=$LOWER ;;
*) FROM=$UPPER; TO=$LOWER ;;
esac
@end group
eval $finder | sed -n '
@group
# remove all trailing slashes
s/\/*$//
@end group
@group
# add ./ if there is no path, only a filename
/\//! s/^/.\//
@end group
@group
# save path+filename
h
@end group
@group
# remove path
s/.*\///
@end group
@group
# do conversion only on filename
y/'$FROM'/'$TO'/
@end group
@group
# now line contains original path+file, while
# hold space contains the new filename
x
@end group
@group
# add converted file name to line, which now contains
# path/file-name\nconverted-file-name
G
@end group
@group
# check if converted file name is equal to original file name,
# if it is, do not print anything
/^.*\/\(.*\)\n\1/b
@end group
@group
# escape special characters for the shell
s/["$`\\]/\\&/g
@end group
@group
# now, transform path/fromfile\n, into
# mv path/fromfile path/tofile and print it
s/^\(.*\/\)\(.*\)\n\(.*\)$/mv "\1\2" "\1\3"/p
@end group
' | $apply_cmd
@end example
@c end---------------------------------------------
@node Print bash environment
@section Print @command{bash} Environment
This script strips the definition of the shell functions
from the output of the @command{set} Bourne-shell command.
@c start-------------------------------------------
@example
#!/bin/sh
@group
set | sed -n '
:x
@end group
@group
@ifinfo
# if no occurrence of "=()" print and load next line
@end ifinfo
@ifnotinfo
# if no occurrence of @samp{=()} print and load next line
@end ifnotinfo
/=()/! @{ p; b; @}
/ () $/! @{ p; b; @}
@end group
@group
# possible start of functions section
# save the line in case this is a var like FOO="() "
h
@end group
@group
# if the next line has a brace, we quit because
# nothing comes after functions
n
/^@{/ q
@end group
@group
# print the old line
x; p
@end group
@group
# work on the new line now
x; bx
'
@end group
@end example
@c end---------------------------------------------
@node Reverse chars of lines
@section Reverse Characters of Lines
This script can be used to reverse the position of characters
in lines. The technique moves two characters at a time, hence
it is faster than more intuitive implementations.
Note the @code{tx} command before the definition of the label.
This is often needed to reset the flag that is tested by
the @code{t} command.
Imaginative readers will find uses for this script. An example
is reversing the output of @command{banner}.@footnote{This requires
another script to pad the output of banner; for example
@example
#! /bin/sh
banner -w $1 $2 $3 $4 |
sed -e :a -e '/^.\@{0,'$1'\@}$/ @{ s/$/ /; ba; @}' |
~/sedscripts/reverseline.sed
@end example
}
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
/../! b
@group
# Reverse a line. Begin embedding the line between two newlines
s/^.*$/\
&\
/
@end group
@group
# Move first character at the end. The regexp matches until
# there are zero or one characters between the markers
tx
:x
s/\(\n.\)\(.*\)\(.\n\)/\3\2\1/
tx
@end group
@group
# Remove the newline markers
s/\n//g
@end group
@end example
@c end---------------------------------------------
@node Text search across multiple lines
@section Text search across multiple lines
This section uses @code{N} and @code{D} commands to search for
consecutive words spanning multiple lines. @xref{Multiline techniques}.
These examples deal with finding doubled occurrences of words in a document.
Finding doubled words in a single line is easy using GNU @command{grep}
and similarly with @value{SSED}:
@c NOTE: in all examples, 'the@ the' is used to prevent
@c 'make syntax-check' from complaining about double words.
@codequoteundirected on
@codequotebacktick on
@example
@group
$ cat two-cities-dup1.txt
It was the best of times,
it was the worst of times,
it was the@ the age of wisdom,
it was the age of foolishness,
$ grep -E '\b(\w+)\s+\1\b' two-cities-dup1.txt
it was the@ the age of wisdom,
$ grep -n -E '\b(\w+)\s+\1\b' two-cities-dup1.txt
3:it was the@ the age of wisdom,
$ sed -En '/\b(\w+)\s+\1\b/p' two-cities-dup1.txt
it was the@ the age of wisdom,
$ sed -En '/\b(\w+)\s+\1\b/@{=;p@}' two-cities-dup1.txt
3
it was the@ the age of wisdom,
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@itemize @bullet
@item
The regular expression @samp{\b\w+\s+} searches for word-boundary (@samp{\b}),
followed by one-or-more word-characters (@samp{\w+}), followed by whitespace
(@samp{\s+}). @xref{regexp extensions}.
@item
Adding parentheses around the @samp{(\w+)} expression creates a subexpression.
The regular expression pattern @samp{(PATTERN)\s+\1} defines a subexpression
(in the parentheses) followed by a back-reference, separated by whitespace.
A successful match means the @var{PATTERN} was repeated twice in succession.
@xref{Back-references and Subexpressions}.
@item
The word-boundery expression (@samp{\b}) at both ends ensures partial
words are not matched (e.g. @samp{the then} is not a desired match).
@c Thanks to Jim for pointing this out in
@c https://lists.gnu.org/archive/html/sed-devel/2016-12/msg00041.html
@item
The @option{-E} option enables extended regular expression syntax, alleviating
the need to add backslashes before the parenthesis. @xref{ERE syntax}.
@end itemize
When the doubled word span two lines the above regular expression
will not find them as @command{grep} and @command{sed} operate line-by-line.
By using @command{N} and @command{D} commands, @command{sed} can apply
regular expressions on multiple lines (that is, multiple lines are stored
in the pattern space, and the regular expression works on it):
@c NOTE: use 'the@*the' instead of a real new line to prevent
@c 'make syntax-check' to complain about doubled-words.
@codequoteundirected on
@codequotebacktick on
@example
$ cat two-cities-dup2.txt
It was the best of times, it was the
worst of times, it was the@*the age of wisdom,
it was the age of foolishness,
$ sed -En '@{N; /\b(\w+)\s+\1\b/@{=;p@} ; D@}' two-cities-dup2.txt
3
worst of times, it was the@*the age of wisdom,
@end example
@codequoteundirected off
@codequotebacktick off
@itemize @bullet
@item
The @command{N} command appends the next line to the pattern space
(thus ensuring it contains two consecutive lines in every cycle).
@item
The regular expression uses @samp{\s+} for word separator which matches
both spaces and newlines.
@item
The regular expression matches, the entire pattern space is printed
with @command{p}. No lines are printed by default due to the @option{-n} option.
@item
The @command{D} removes the first line from the pattern space (up until the
first newline), readying it for the next cycle.
@end itemize
See the GNU @command{coreutils} manual for an alternative solution using
@command{tr -s} and @command{uniq} at
@c NOTE: cheating and keeping the URL line shorter than 80 characters
@c by using 'gnu.org' and '/s/'.
@url{https://gnu.org/s/coreutils/manual/html_node/Squeezing-and-deleting.html}.
@node Line length adjustment
@section Line length adjustment
This section uses @code{N} and @code{P} commands to read and write
lines, and the @code{b} command for branching.
@xref{Multiline techniques} and @ref{Branching and flow control}.
This (somewhat contrived) example deal with formatting and wrapping
lines of text of the following input file:
@example
@group
$ cat two-cities-mix.txt
It was the best of times, it was
the worst of times, it
was the age of
wisdom,
it
was
the age
of foolishness,
@end group
@end example
@exdent The following sed program wraps lines at 40 characters:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ cat wrap40.sed
# outer loop
:x
# Append a newline followed by the next input line to the pattern buffer
N
# Remove all newlines from the pattern buffer
s/\n/ /g
# Inner loop
:y
# Add a newline after the first 40 characters
s/(.@{40,40@})/\1\n/
# If there is a newline in the pattern buffer
# (i.e. the previous substitution added a newline)
/\n/ @{
# There are newlines in the pattern buffer -
# print the content until the first newline.
P
# Remove the printed characters and the first newline
s/.*\n//
# branch to label 'y' - repeat inner loop
by
@}
# No newlines in the pattern buffer - Branch to label 'x' (outer loop)
# and read the next input line
bx
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@exdent The wrapped output:
@codequoteundirected on
@codequotebacktick on
@example
@group
$ sed -E -f wrap40.sed two-cities-mix.txt
It was the best of times, it was the wor
st of times, it was the age of wisdom, i
t was the age of foolishness,
@end group
@end example
@codequoteundirected off
@codequotebacktick off
@node Adding a header to multiple files
@section Adding a header to multiple files
@value{SSED} can be used to safely modify multiple files at once.
@exdent Add a single line to the beginning of source code files:
@codequoteundirected on
@codequotebacktick on
@example
sed -i '1i/* Copyright (C) FOO BAR */' *.c
@end example
@codequoteundirected off
@codequotebacktick off
@exdent Adding a few lines is possible using @samp{\n} in the text:
@codequoteundirected on
@codequotebacktick on
@example
sed -i '1i/*\n * Copyright (C) FOO BAR\n * Created by Jane Doe\n */' *.c
@end example
@codequoteundirected off
@codequotebacktick off
To add multiple lines from another file, use @code{0rFILE}.
A typical use case is adding a license notice header to all files:
@codequoteundirected on
@codequotebacktick on
@example
## Create the header file:
$ cat<<'EOF'>LIC.TXT
/*
Copyright (C) 1989-2021 FOO BAR
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; If not, see <https://www.gnu.org/licenses/>.
*/
EOF
## Add the file at the beginning of all source code files:
$ sed -i '0rLIC.TXT' *.cpp *.h
@end example
@codequoteundirected off
@codequotebacktick off
With script files (e.g. @file{.sh},@file{.py},@file{.pl} files)
the license notice typically appears @emph{after} the first line (the
'shebang' @samp{#!} line). The @code{1rFILE} command will add @file{FILE}
@emph{after} the first line:
@codequoteundirected on
@codequotebacktick on
@example
## Create the header file:
$ cat<<'EOF'>LIC.TXT
##
## Copyright (C) 1989-2021 FOO BAR
##
## This program is free software; you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation; either version 3, or (at your option)
## any later version.
##
## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
## GNU General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with this program; If not, see <https://www.gnu.org/licenses/>.
##
##
EOF
## Add the file at the beginning of all source code files:
$ sed -i '1rLIC.TXT' *.py *.sh
@end example
@codequoteundirected off
@codequotebacktick off
The above @command{sed} commands can be combined with @command{find}
to locate files in all subdirectories, @command{xargs} to run additional
commands on selected files and @command{grep} to filter out files that already
contain a copyright notice:
@codequoteundirected on
@codequotebacktick on
@example
find \( -iname '*.cpp' -o -iname '*.c' -o -iname '*.h' \) \
| xargs grep -Li copyright \
| xargs -r sed -i '0rLIC.TXT'
@end example
@codequoteundirected off
@codequotebacktick off
@exdent Or a slightly safe version (handling files with spaces and newlines):
@codequoteundirected on
@codequotebacktick on
@example
find \( -iname '*.cpp' -o -iname '*.c' -o -iname '*.h' \) -print0 \
| xargs -0 grep -Z -Li copyright \
| xargs -0 -r sed -i '0rLIC.TXT'
@end example
@codequoteundirected off
@codequotebacktick off
Note: using the @code{0} address with @code{r} command requires @value{SSED}
version 4.9 or later. @xref{Zero Address}.
@node tac
@section Reverse Lines of Files
This one begins a series of totally useless (yet interesting)
scripts emulating various Unix commands. This, in particular,
is a @command{tac} workalike.
Note that on implementations other than GNU @command{sed}
this script might easily overflow internal buffers.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
# reverse all lines of input, i.e. first line became last, ...
@group
# from the second line, the buffer (which contains all previous lines)
# is *appended* to current line, so, the order will be reversed
1! G
@end group
@group
# on the last line we're done -- print everything
$ p
@end group
@group
# store everything on the buffer again
h
@end group
@end example
@c end---------------------------------------------
@node cat -n
@section Numbering Lines
This script replaces @samp{cat -n}; in fact it formats its output
exactly like GNU @command{cat} does.
Of course this is completely useless and for two reasons: first,
because somebody else did it in C, second, because the following
Bourne-shell script could be used for the same purpose and would
be much faster:
@c start-------------------------------------------
@example
@group
#! /bin/sh
sed -e "=" $@@ | sed -e '
s/^/ /
N
s/^ *\(......\)\n/\1 /
'
@end group
@end example
@c end---------------------------------------------
It uses @command{sed} to print the line number, then groups lines two
by two using @code{N}. Of course, this script does not teach as much as
the one presented below.
The algorithm used for incrementing uses both buffers, so the line
is printed as soon as possible and then discarded. The number
is split so that changing digits go in a buffer and unchanged ones go
in the other; the changed digits are modified in a single step
(using a @code{y} command). The line number for the next line
is then composed and stored in the hold space, to be used in the
next iteration.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
# Prime the pump on the first line
x
/^$/ s/^.*$/1/
@end group
@group
# Add the correct line number before the pattern
G
h
@end group
@group
# Format it and print it
s/^/ /
s/^ *\(......\)\n/\1 /p
@end group
@group
# Get the line number from hold space; add a zero
# if we're going to add a digit on the next line
g
s/\n.*$//
/^9*$/ s/^/0/
@end group
@group
# separate changing/unchanged digits with an x
s/.9*$/x&/
@end group
@group
# keep changing digits in hold space
h
s/^.*x//
y/0123456789/1234567890/
x
@end group
@group
# keep unchanged digits in pattern space
s/x.*$//
@end group
@group
# compose the new number, remove the newline implicitly added by G
G
s/\n//
h
@end group
@end example
@c end---------------------------------------------
@node cat -b
@section Numbering Non-blank Lines
Emulating @samp{cat -b} is almost the same as @samp{cat -n}---we only
have to select which lines are to be numbered and which are not.
The part that is common to this script and the previous one is
not commented to show how important it is to comment @command{sed}
scripts properly...
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
/^$/ @{
p
b
@}
@end group
@group
# Same as cat -n from now
x
/^$/ s/^.*$/1/
G
h
s/^/ /
s/^ *\(......\)\n/\1 /p
x
s/\n.*$//
/^9*$/ s/^/0/
s/.9*$/x&/
h
s/^.*x//
y/0123456789/1234567890/
x
s/x.*$//
G
s/\n//
h
@end group
@end example
@c end---------------------------------------------
@node wc -c
@section Counting Characters
This script shows another way to do arithmetic with @command{sed}.
In this case we have to add possibly large numbers, so implementing
this by successive increments would not be feasible (and possibly
even more complicated to contrive than this script).
The approach is to map numbers to letters, kind of an abacus
implemented with @command{sed}. @samp{a}s are units, @samp{b}s are
tens and so on: we simply add the number of characters
on the current line as units, and then propagate the carry
to tens, hundreds, and so on.
As usual, running totals are kept in hold space.
On the last line, we convert the abacus form back to decimal.
For the sake of variety, this is done with a loop rather than
with some 80 @code{s} commands@footnote{Some implementations
have a limit of 199 commands per script}: first we
convert units, removing @samp{a}s from the number; then we
rotate letters so that tens become @samp{a}s, and so on
until no more letters remain.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
# Add n+1 a's to hold space (+1 is for the newline)
s/./a/g
H
x
s/\n/a/
@end group
@group
# Do the carry. The t's and b's are not necessary,
# but they do speed up the thing
t a
: a; s/aaaaaaaaaa/b/g; t b; b done
: b; s/bbbbbbbbbb/c/g; t c; b done
: c; s/cccccccccc/d/g; t d; b done
: d; s/dddddddddd/e/g; t e; b done
: e; s/eeeeeeeeee/f/g; t f; b done
: f; s/ffffffffff/g/g; t g; b done
: g; s/gggggggggg/h/g; t h; b done
: h; s/hhhhhhhhhh//g
@end group
@group
: done
$! @{
h
b
@}
@end group
# On the last line, convert back to decimal
@group
: loop
/a/! s/[b-h]*/&0/
s/aaaaaaaaa/9/
s/aaaaaaaa/8/
s/aaaaaaa/7/
s/aaaaaa/6/
s/aaaaa/5/
s/aaaa/4/
s/aaa/3/
s/aa/2/
s/a/1/
@end group
@group
: next
y/bcdefgh/abcdefg/
/[a-h]/ b loop
p
@end group
@end example
@c end---------------------------------------------
@node wc -w
@section Counting Words
This script is almost the same as the previous one, once each
of the words on the line is converted to a single @samp{a}
(in the previous script each letter was changed to an @samp{a}).
It is interesting that real @command{wc} programs have optimized
loops for @samp{wc -c}, so they are much slower at counting
words rather than characters. This script's bottleneck,
instead, is arithmetic, and hence the word-counting one
is faster (it has to manage smaller numbers).
Again, the common parts are not commented to show the importance
of commenting @command{sed} scripts.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
# Convert words to a's
s/[ @kbd{@key{TAB}}][ @kbd{@key{TAB}}]*/ /g
s/^/ /
s/ [^ ][^ ]*/a /g
s/ //g
@end group
@group
# Append them to hold space
H
x
s/\n//
@end group
@group
# From here on it is the same as in wc -c.
/aaaaaaaaaa/! bx; s/aaaaaaaaaa/b/g
/bbbbbbbbbb/! bx; s/bbbbbbbbbb/c/g
/cccccccccc/! bx; s/cccccccccc/d/g
/dddddddddd/! bx; s/dddddddddd/e/g
/eeeeeeeeee/! bx; s/eeeeeeeeee/f/g
/ffffffffff/! bx; s/ffffffffff/g/g
/gggggggggg/! bx; s/gggggggggg/h/g
s/hhhhhhhhhh//g
:x
$! @{ h; b; @}
:y
/a/! s/[b-h]*/&0/
s/aaaaaaaaa/9/
s/aaaaaaaa/8/
s/aaaaaaa/7/
s/aaaaaa/6/
s/aaaaa/5/
s/aaaa/4/
s/aaa/3/
s/aa/2/
s/a/1/
y/bcdefgh/abcdefg/
/[a-h]/ by
p
@end group
@end example
@c end---------------------------------------------
@node wc -l
@section Counting Lines
No strange things are done now, because @command{sed} gives us
@samp{wc -l} functionality for free!!! Look:
@c start-------------------------------------------
@example
@group
#!/usr/bin/sed -nf
$=
@end group
@end example
@c end---------------------------------------------
@node head
@section Printing the First Lines
This script is probably the simplest useful @command{sed} script.
It displays the first 10 lines of input; the number of displayed
lines is right before the @code{q} command.
@c start-------------------------------------------
@example
@group
#!/usr/bin/sed -f
10q
@end group
@end example
@c end---------------------------------------------
@node tail
@section Printing the Last Lines
Printing the last @var{n} lines rather than the first is more complex
but indeed possible. @var{n} is encoded in the second line, before
the bang character.
This script is similar to the @command{tac} script in that it keeps the
final output in the hold space and prints it at the end:
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
1! @{; H; g; @}
1,10 !s/[^\n]*\n//
$p
h
@end group
@end example
@c end---------------------------------------------
Mainly, the scripts keeps a window of 10 lines and slides it
by adding a line and deleting the oldest (the substitution command
on the second line works like a @code{D} command but does not
restart the loop).
The ``sliding window'' technique is a very powerful way to write
efficient and complex @command{sed} scripts, because commands like
@code{P} would require a lot of work if implemented manually.
To introduce the technique, which is fully demonstrated in the
rest of this chapter and is based on the @code{N}, @code{P}
and @code{D} commands, here is an implementation of @command{tail}
using a simple ``sliding window.''
This looks complicated but in fact the working is the same as
the last script: after we have kicked in the appropriate number
of lines, however, we stop using the hold space to keep inter-line
state, and instead use @code{N} and @code{D} to slide pattern
space by one line:
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
@group
1h
2,10 @{; H; g; @}
$q
1,9d
N
D
@end group
@end example
@c end---------------------------------------------
Note how the first, second and fourth line are inactive after
the first ten lines of input. After that, all the script does
is: exiting on the last line of input, appending the next input
line to pattern space, and removing the first line.
@node uniq
@section Make Duplicate Lines Unique
This is an example of the art of using the @code{N}, @code{P}
and @code{D} commands, probably the most difficult to master.
@c start-------------------------------------------
@example
@group
#!/usr/bin/sed -f
h
@end group
@group
:b
# On the last line, print and exit
$b
N
/^\(.*\)\n\1$/ @{
# The two lines are identical. Undo the effect of
# the n command.
g
bb
@}
@end group
@group
# If the @code{N} command had added the last line, print and exit
$b
@end group
@group
# The lines are different; print the first and go
# back working on the second.
P
D
@end group
@end example
@c end---------------------------------------------
As you can see, we maintain a 2-line window using @code{P} and @code{D}.
This technique is often used in advanced @command{sed} scripts.
@node uniq -d
@section Print Duplicated Lines of Input
This script prints only duplicated lines, like @samp{uniq -d}.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
$b
N
/^\(.*\)\n\1$/ @{
# Print the first of the duplicated lines
s/.*\n//
p
@end group
@group
# Loop until we get a different line
:b
$b
N
/^\(.*\)\n\1$/ @{
s/.*\n//
bb
@}
@}
@end group
@group
# The last line cannot be followed by duplicates
$b
@end group
@group
# Found a different one. Leave it alone in the pattern space
# and go back to the top, hunting its duplicates
D
@end group
@end example
@c end---------------------------------------------
@node uniq -u
@section Remove All Duplicated Lines
This script prints only unique lines, like @samp{uniq -u}.
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
@group
# Search for a duplicate line --- until that, print what you find.
$b
N
/^\(.*\)\n\1$/ ! @{
P
D
@}
@end group
@group
:c
# Got two equal lines in pattern space. At the
# end of the file we simply exit
$d
@end group
@group
# Else, we keep reading lines with @code{N} until we
# find a different one
s/.*\n//
N
/^\(.*\)\n\1$/ @{
bc
@}
@end group
@group
# Remove the last instance of the duplicate line
# and go back to the top
D
@end group
@end example
@c end---------------------------------------------
@node cat -s
@section Squeezing Blank Lines
As a final example, here are three scripts, of increasing complexity
and speed, that implement the same function as @samp{cat -s}, that is
squeezing blank lines.
The first leaves a blank line at the beginning and end if there are
some already.
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
@group
# on empty lines, join with next
# Note there is a star in the regexp
:x
/^\n*$/ @{
N
bx
@}
@end group
@group
# now, squeeze all '\n', this can be also done by:
# s/^\(\n\)*/\1/
s/\n*/\
/
@end group
@end example
@c end---------------------------------------------
This one is a bit more complex and removes all empty lines
at the beginning. It does leave a single blank line at end
if one was there.
@c start-------------------------------------------
@example
#!/usr/bin/sed -f
@group
# delete all leading empty lines
1,/^./@{
/./!d
@}
@end group
@group
# on an empty line we remove it and all the following
# empty lines, but one
:x
/./!@{
N
s/^\n$//
tx
@}
@end group
@end example
@c end---------------------------------------------
This removes leading and trailing blank lines. It is also the
fastest. Note that loops are completely done with @code{n} and
@code{b}, without relying on @command{sed} to restart the
script automatically at the end of a line.
@c start-------------------------------------------
@example
#!/usr/bin/sed -nf
@group
# delete all (leading) blanks
/./!d
@end group
@group
# get here: so there is a non empty
:x
# print it
p
# get next
n
# got chars? print it again, etc...
/./bx
@end group
@group
# no, don't have chars: got an empty line
:z
# get next, if last line we finish here so no trailing
# empty lines are written
n
# also empty? then ignore it, and get next... this will
# remove ALL empty lines
/./!bz
@end group
@group
# all empty lines were deleted/ignored, but we have a non empty. As
# what we want to do is to squeeze, insert a blank line artificially
i\
@end group
bx
@end example
@c end---------------------------------------------
@node Limitations
@chapter @value{SSED}'s Limitations and Non-limitations
@cindex GNU extensions, unlimited line length
@cindex Portability, line length limitations
For those who want to write portable @command{sed} scripts,
be aware that some implementations have been known to
limit line lengths (for the pattern and hold spaces)
to be no more than 4000 bytes.
The @sc{posix} standard specifies that conforming @command{sed}
implementations shall support at least 8192 byte line lengths.
@value{SSED} has no built-in limit on line length;
as long as it can @code{malloc()} more (virtual) memory,
you can feed or construct lines as long as you like.
However, recursion is used to handle subpatterns and indefinite
repetition. This means that the available stack space may limit
the size of the buffer that can be processed by certain patterns.
@node Other Resources
@chapter Other Resources for Learning About @command{sed}
For up to date information about @value{SSED} please
visit @uref{https://www.gnu.org/software/sed/}.
Send general questions and suggestions to @email{sed-devel@@gnu.org}.
Visit the mailing list archives for past discussions at
@uref{https://lists.gnu.org/archive/html/sed-devel/}.
@cindex Additional reading about @command{sed}
The following resources provide information about @command{sed}
(both @value{SSED} and other variations). Note these not maintained by
@value{SSED} developers.
@itemize @bullet
@item
sed @code{$HOME}: @uref{http://sed.sf.net}
@item
sed FAQ: @uref{http://sed.sf.net/sedfaq.html}
@item
seder's grabbag: @uref{http://sed.sf.net/grabbag}
@item
The @code{sed-users} mailing list maintained by Sven Guckes:
@uref{http://groups.yahoo.com/group/sed-users/}
(note this is @emph{not} the @value{SSED} mailing list).
@end itemize
@node Reporting Bugs
@chapter Reporting Bugs
@cindex Bugs, reporting
Email bug reports to @email{bug-sed@@gnu.org}.
Also, please include the output of @samp{sed --version} in the body
of your report if at all possible.
Please do not send a bug report like this:
@example
@i{@i{@r{while building frobme-1.3.4}}}
$ configure
@error{} sed: file sedscr line 1: Unknown option to 's'
@end example
If @value{SSED} doesn't configure your favorite package, take a
few extra minutes to identify the specific problem and make a stand-alone
test case. Unlike other programs such as C compilers, making such test
cases for @command{sed} is quite simple.
A stand-alone test case includes all the data necessary to perform the
test, and the specific invocation of @command{sed} that causes the problem.
The smaller a stand-alone test case is, the better. A test case should
not involve something as far removed from @command{sed} as ``try to configure
frobme-1.3.4''. Yes, that is in principle enough information to look
for the bug, but that is not a very practical prospect.
Here are a few commonly reported bugs that are not bugs.
@table @asis
@anchor{N_command_last_line}
@item @code{N} command on the last line
@cindex Portability, @code{N} command on the last line
@cindex Non-bugs, @code{N} command on the last line
Most versions of @command{sed} exit without printing anything when
the @command{N} command is issued on the last line of a file.
@value{SSED} prints pattern space before exiting unless of course
the @command{-n} command switch has been specified. This choice is
by design.
Default behavior (gnu extension, non-POSIX conforming):
@example
$ seq 3 | sed N
1
2
3
@end example
@noindent
To force POSIX-conforming behavior:
@example
$ seq 3 | sed --posix N
1
2
@end example
For example, the behavior of
@example
sed N foo bar
@end example
@noindent
would depend on whether foo has an even or an odd number of
lines@footnote{which is the actual ``bug'' that prompted the
change in behavior}. Or, when writing a script to read the
next few lines following a pattern match, traditional
implementations of @code{sed} would force you to write
something like
@example
/foo/@{ $!N; $!N; $!N; $!N; $!N; $!N; $!N; $!N; $!N @}
@end example
@noindent
instead of just
@example
/foo/@{ N;N;N;N;N;N;N;N;N; @}
@end example
@cindex @code{POSIXLY_CORRECT} behavior, @code{N} command
In any case, the simplest workaround is to use @code{$d;N} in
scripts that rely on the traditional behavior, or to set
the @code{POSIXLY_CORRECT} variable to a non-empty value.
@item Regex syntax clashes (problems with backslashes)
@cindex GNU extensions, to basic regular expressions
@cindex Non-bugs, regex syntax clashes
@command{sed} uses the @sc{posix} basic regular expression syntax. According to
the standard, the meaning of some escape sequences is undefined in
this syntax; notable in the case of @command{sed} are @code{\|},
@code{\+}, @code{\?}, @code{\`}, @code{\'}, @code{\<},
@code{\>}, @code{\b}, @code{\B}, @code{\w}, and @code{\W}.
As in all GNU programs that use @sc{posix} basic regular
expressions, @command{sed} interprets these escape sequences as special
characters. So, @code{x\+} matches one or more occurrences of @samp{x}.
@code{abc\|def} matches either @samp{abc} or @samp{def}.
This syntax may cause problems when running scripts written for other
@command{sed}s. Some @command{sed} programs have been written with the
assumption that @code{\|} and @code{\+} match the literal characters
@code{|} and @code{+}. Such scripts must be modified by removing the
spurious backslashes if they are to be used with modern implementations
of @command{sed}, like
GNU @command{sed}.
On the other hand, some scripts use s|abc\|def||g to remove occurrences
of @emph{either} @code{abc} or @code{def}. While this worked until
@command{sed} 4.0.x, newer versions interpret this as removing the
string @code{abc|def}. This is again undefined behavior according to
POSIX, and this interpretation is arguably more robust: older
@command{sed}s, for example, required that the regex matcher parsed
@code{\/} as @code{/} in the common case of escaping a slash, which is
again undefined behavior; the new behavior avoids this, and this is good
because the regex matcher is only partially under our control.
@cindex GNU extensions, special escapes
In addition, this version of @command{sed} supports several escape characters
(some of which are multi-character) to insert non-printable characters
in scripts (@code{\a}, @code{\c}, @code{\d}, @code{\o}, @code{\r},
@code{\t}, @code{\v}, @code{\x}). These can cause similar problems
with scripts written for other @command{sed}s.
@item @option{-i} clobbers read-only files
@cindex In-place editing
@cindex @value{SSEDEXT}, in-place editing
@cindex Non-bugs, in-place editing
In short, @samp{sed -i} will let you delete the contents of
a read-only file, and in general the @option{-i} option
(@pxref{Invoking sed, , Invocation}) lets you clobber
protected files. This is not a bug, but rather a consequence
of how the Unix file system works.
The permissions on a file say what can happen to the data
in that file, while the permissions on a directory say what can
happen to the list of files in that directory. @samp{sed -i}
will not ever open for writing a file that is already on disk.
Rather, it will work on a temporary file that is finally renamed
to the original name: if you rename or delete files, you're actually
modifying the contents of the directory, so the operation depends on
the permissions of the directory, not of the file. For this same
reason, @command{sed} does not let you use @option{-i} on a writable file
in a read-only directory, and will break hard or symbolic links when
@option{-i} is used on such a file.
@item @code{0a} does not work (gives an error)
@cindex @code{0} address
@cindex GNU extensions, @code{0} address
@cindex Non-bugs, @code{0} address
There is no line 0. 0 is a special address that is only used to treat
addresses like @code{0,/@var{RE}/} as active when the script starts: if
you write @code{1,/abc/d} and the first line includes the string @samp{abc},
then that match would be ignored because address ranges must span at least
two lines (barring the end of the file); but what you probably wanted is
to delete every line up to the first one including @samp{abc}, and this
is obtained with @code{0,/abc/d}.
@ifclear PERL
@item @code{[a-z]} is case insensitive
@cindex Non-bugs, localization-related
You are encountering problems with locales. POSIX mandates that @code{[a-z]}
uses the current locale's collation order -- in C parlance, that means using
@code{strcoll(3)} instead of @code{strcmp(3)}. Some locales have a
case-insensitive collation order, others don't.
Another problem is that @code{[a-z]} tries to use collation symbols.
This only happens if you are on the GNU system, using
GNU libc's regular expression matcher instead of compiling the
one supplied with GNU sed. In a Danish locale, for example,
the regular expression @code{^[a-z]$} matches the string @samp{aa},
because this is a single collating symbol that comes after @samp{a}
and before @samp{b}; @samp{ll} behaves similarly in Spanish
locales, or @samp{ij} in Dutch locales.
To work around these problems, which may cause bugs in shell scripts, set
the @env{LC_COLLATE} and @env{LC_CTYPE} environment variables to @samp{C}.
@item @code{s/.*//} does not clear pattern space
@cindex Non-bugs, localization-related
@cindex @value{SSEDEXT}, emptying pattern space
@cindex Emptying pattern space
This happens if your input stream includes invalid multibyte
sequences. @sc{posix} mandates that such sequences
are @emph{not} matched by @samp{.}, so that @samp{s/.*//} will not clear
pattern space as you would expect. In fact, there is no way to clear
sed's buffers in the middle of the script in most multibyte locales
(including UTF-8 locales). For this reason, @value{SSED} provides a `z'
command (for `zap') as an extension.
To work around these problems, which may cause bugs in shell scripts, set
the @env{LC_COLLATE} and @env{LC_CTYPE} environment variables to @samp{C}.
@end ifclear
@end table
@page
@node GNU Free Documentation License
@appendix GNU Free Documentation License
@include fdl.texi
@page
@node Concept Index
@unnumbered Concept Index
This is a general index of all issues discussed in this manual, with the
exception of the @command{sed} commands and command-line options.
@printindex cp
@page
@node Command and Option Index
@unnumbered Command and Option Index
This is an alphabetical list of all @command{sed} commands and command-line
options.
@printindex fn
@contents
@bye
@c XXX FIXME: the term "cycle" is never defined...