Searches the input files (or STDIN if no files are named, or if file is - )
for lines containing a match to the pattern.

-G --basic-regexp pattern is a Basic Regular Expression (default).
                                                  ?, +, {, |, ( and ) have no special meaning

-E --extended-regexp pattern is an Extended Regular Expression.
egrep is the same as grep -E.

-F --fixed-strings pattern is a list of fixed strings, separated by newlines, any of which is to be matched.
fgrep is the same as grep -F.

-P --perl-regexp Interpret pattern as a Perl regular expression.

controling matching
-v --invert-match	Invert the sense of matching to select non-matching lines.
-i --ignore-case
-e pattern --regexp=pattern	Use pattern as the matching pattern; useful to protect patterns beginning with - from being intrepreted as options.
-f file --file=file	Obtain patterns from file, one per line. An empty file contains zero patterns and matches nothing.
-w --word-regexp	matches must form whole words. The matching substring must either be at the beginning of the line, or preceded by a non-alphanumeric character and must be either at the end of the line or followed by a non-alphanumeric character. Alphanumeric characters are letters, digits, and the underscore.
-x --line-regexp	exactly match the whole line.
-y	Obsolete synonym for -i (--ignore-case).

controlling output
-o --only-matching	Show only the part of a line that matches pattern.
--color[=when] -colour[=when]	Display matches with screen attributes as per the GREP_COLOR environment variable. Foreground: black 30, green 32, yellow 33, blue 34, purple 35, cyan 36, white 37, red 31 (default) Background: black 40, green 42, yellow 43, blue 44, purple 45, underscore 4, [blink] 5, inverse 7 codes can be combined using ; as in export GREP_COLOR="33;44" when may be never, always, or auto Codes are not output when inapropriate (like going to a file or pipe) unless when=always grep --color=always birds logfile | more -r
-Bnum --before-context=num	Output num lines of context before lines containing the pattern then the lines containing pattern.
-Anum --after-context=num	Output lines containing the pattern and num lines of context after them. Places a line containing -- between contiguous groups of matches. example:    grep -A2 d  grep -B2 d  input  output a b c   b c d d d e f g e f   … -- -- aa bb cc   bb cc dd dd dd ee ff gg ee ff
-c --count	only output count of matches (or -cv non-matches) for each input file Prefixed by 'filename :' if multiple files, even if count is 0 grep -c images *html |grep -v 0$†      With the -v, --invert-match, count non-matching lines.
-n --line-number	Prefix each line with the line number
-b --byte-offset	display the byte offset within the input file before each line
-u --unix-byte-offsets	Use Unix-style byte offsets. report byte offsets i.e. with CR characters stripped off. only for MS-DOS and MS-Windows.
-q --quiet --silent	Quiet; do not write to standard output. Errors like "Permission denied" are written to STDERR and return code is set to 2. Exit with zero status as soon as a match is found. If no match is found $? is set to 1.
-s --no-messages	Suppress error messages about nonexistent or unaccessable files normally sent to STDERR.
-H --with-filename	filename is prepended to each match when multiple files are searched.
-h --no-filename	don't
-l --files-with-matches	outputs the only the filename of files containing a match. Once a match is found grep proceeds to the next file.
-L --files-without-match	outputs the filenname with no match
-Z --null	terminate filename with NULL . Used when file names contain newlines!. Used with commands find -print0, perl -0, sort -z, and xargs -0.
--label=label	Displays input actually coming from standard input as input coming from file label. This is useful for tools like zgrep, e.g. gzip -cd foo.gz |grep --label=foo something

Restricting Processing
-C num --context=num	limit output to num lines
-m num --max-count= num	stop reading a file after num matching lines. If the input is standard input from a regular file, and num matching lines are output, the standard input is positioned to just after the last matching line. This enables a calling process to resume a search. When grep stops after num matching lines, it outputs any trailing context lines. When used with -c or --count , grep does not output a count greater than num. When used with -v or --invert-match , grep stops after outputting num non-matching lines.

Directory processing
-R -r --recursive	Read all files under each directory, recursively; equivalent to -d recurse .
-d action --directories=action	If an input is a directory, use action to process it. recurse read all files under each directory, recursively; equivalent to -r .
--include=pattern	Recurse in directories only searching file matching pattern.
--exclude=pattern	Recurse in directories skip file matching pattern.

Special handling
--binary‑files=type	If the first few bytes of a file indicate that the file contains binary data, outputs either a one-line message saying that a binary file matches, or no message if there is no match. binary the default and without-match, grep assumes that a binary file does not match; this is equivalent to -I (Ignore). text processe a binary file as if it were ascii text; equivalent to -a . Outputing control characters to the terminal can set attributes on the terminal making it unreadable. If this happens try STTY SANE
-a --text	Process a binary file as if it were ASCII text equivalent to ‑‑binary‑files=text.
-I	Ignore a binary file (treat it as if it did not contain matching data ) equivalent to ‑‑binary‑files=without‑match.
-U --binary	Treat the file(s) as binary. By default, under MS-DOS and MS-Windows, grep guesses the file type by looking at the contents of the first 32KB read from the file. If grep decides the file is a text file, it strips the CR characters from the original file contents (to make regular expressions with ^ and $ work correctly). Specifying -U causes all files to be read and passed to the matching mechanism verbatim; if the file is a text file with CR/LF pairs at the end of each line, this will cause some regular expressions to fail. only effective under MS-DOS and MS-Windows unless used with -b .

--help
-V --version	Display the version to standard error.
--mmap	use the mmap system call to read input, instead of read which may yield better performance. This can cause undefined behavior if an input file shrinks, or if an I/O error occurs.
-D action --devices=action	If an input file is a device, FIFO or socket, use action to process it. The default action is read, which means that devices are read as ordinary files. If action is skip, devices are silently skipped.
--line-buffered	can have a performance penality.

Basic Regular Expressions: ?, +, {, |, ( and ) have no special meaning * is a meta character.

REGULAR EXPRESSIONS A pattern that describes a set of strings, constructed by using various operators to combine smaller expressions. Grep understands two different versions of regular expression syntax: basic and extended. In GNU grep, there is no difference in available functionality using either syntax. The simplest regular expression matchs a single character. examples: a matches a , R matches R Regular expresions can be combined. AR matches AR bracket expressions list of characters enclosed by [ and ] matches any single character in the list. Most metacharacters have no special meaning inside lists.  To include ] place it first in the list. ][)(  To include ^ place it anywhere but first.  To include - place it last. For example: [123456789] matches any single digit but not 0. nonmatching list begins with a ^  (caret) and causes the expression to match any character not in the list. For example: ^[abcdefghijklmnopqrstuvwxyz] matches a single non-lowercase character. range expression consists of two characters separated by - (hyphen) and matches any single character that sorts between the two characters, inclusive, using the locale's collating sequence and character set. For example: [3-7] is equivalent to [34567]. Many locales sort characters in dictionary order, in these locales [a-d] is typically equivalent [aBbCcDd] not [abcd]. To obtain the traditional interpretation of bracket expressions, use the C locale by setting the LC_ALL environment variable to the value C. named classes : [:alnum:], [:alpha:], [:cntrl:], [:digit:], [:graph:], [:lower:], [:print:], [:punct:], [:space:], [:upper:], and [:xdigit:]. Brackets in these class names are part of the symbolic names, and must be included in addition to any brackets that may be delimiting a bracket expression. [[:digit:]] is equivalent to [0123456789] [[:alnum:]] is frequently equivalent to [0-9A-Za-z] depending on the C locale and the ASCII character encoding, whereas the [:alnum:] is independent of locale and character set. Using ANSI-C quoting from the bash for [TAB] use grep $'\tA\t' to find an A surronded by tabs. .  (period) matches any single character. ^  (caret) matches the beginning of a line $  (dollar sign) matches the end of a line. \w (word) is a synonym for [[:alnum:]] and \W is a synonym for [^[:alnum:]]. \> matches the empty string at the beginning of a word. \< matches the empty string at the end of a word. \b matches the empty string at the edge of a word, \B matches the empty string provided it's not at the edge of a word. repetition operators follow a regular expression the preceding item ? matched at most once † Use for stingy vs greedy matching. placing a question mark after any of the greedy quantifiers, they can be made to choose the smallest quantity for the first try. So if you say /.*?foo/, the .*? * will be matched zero or more times † Match is greedy attempting to match as much as possible. for example /.*xx/ matches everything in the line:asdfbxx asdfxx asdfxx + will be matched one or more times † {n} is matched exactly n times The braces may need to be escaped as in z\{1,3\} z zz or zzz {n,} is matched n or more times {n,m} is matched at least n times, but not more than m times. /[0-9]\{2,3\} bytes/ matches 99 bytes and 999 bytes as does /[[:digit:]]\{2,3\} bytes/ A metacharacter ( examples: * , . , [) can be treated as normal character by preceding it with a \  (backslash) Two regular expressions may be concatenated as in grep "aa""bb". Regular expressions joined by | match any string matching either expression. † Repetition takes precedence over concatenation, which in turn takes precedence over alternation. A subexpression enclosed in parentheses overrides precedence rules. backreference \n, where n is a single digit, matches the substring previously matched by the nth parenthesized subexpression of the regular expression.

Environment Variables

GREP_COLOR Specifies the color for highlighting.

A locale LC_xxx is specified by examining:
LC_ALL, LC_xxx, LANG, in order.
The first of these variables that is set specifies the locale.
For example, if LC_ALL is not set, but LC_MESSAGES is set to pt_BR, then BrazilianPortuguese is used.
The C locale is used if none of these are set, or if the locale catalog is not installed, or if grep was not compiled with national language support (NLS).
LC_ALL, LC_COLLATE,  LANG collating sequence used to interpret range expressions like [a-z].
LC_ALL, LC_CTYPE,    LANG type of characters, e.g., which characters are whitespace.
LC_ALL, LC_MESSAGES, LANG language for messages. The default C locale uses American English messages.

POSIXLY_CORRECT If set, grep behaves as POSIX.2 requires; otherwise, grep behaves more like other GNU programs.

• requires that options that follow file names must be treated as file names; by default, such options are permuted to the front of the operand list and are treated as options.
• unrecognized options be diagnosed as "illegal".

_N_GNU_nonoption_argv_flags_ (N is grep's numeric process ID.)
If the i^th character of this environment variable's value is 1,
do not consider the i^th operand to be an option, even if it appears to be one.
A shell can put this variable in the environment for each command it runs, specifying which operands are the results of file name wildcard expansion and therefore should not be treated as options.
Only with the GNU C library, and only when POSIXLY_CORRECT is not set.

Diagnostics
0 if selected lines are found and
1 no lines selected.
2 an error occurred like: no permission or file not found.

Large repetition counts in the {n,m} construct may cause grep to use lots of memory. Certain obscure regular expressions require exponential time and space.

Backreferences are very slow, and require exponential time.

Current "official" GNU Doc

See also egrep, fgrep, sed, sh, attributes, environ, largefile, regex, regexp, XPG4

GNU Project 2002/01/22 GREP(1)