1 PCREGREP(1) PCREGREP(1)
5 pcregrep - a grep with Perl-compatible regular expressions.
9 pcregrep [options] [long options] [pattern] [path1 path2 ...]
14 pcregrep searches files for character patterns, in the same way as
15 other grep commands do, but it uses the PCRE regular expression library
16 to support patterns that are compatible with the regular expressions of
17 Perl 5. See pcrepattern(3) for a full description of syntax and seman-
18 tics of the regular expressions that PCRE supports.
20 Patterns, whether supplied on the command line or in a separate file,
21 are given without delimiters. For example:
23 pcregrep Thursday /etc/motd
25 If you attempt to use delimiters (for example, by surrounding a pattern
26 with slashes, as is common in Perl scripts), they are interpreted as
27 part of the pattern. Quotes can of course be used to delimit patterns
28 on the command line because they are interpreted by the shell, and
29 indeed they are required if a pattern contains white space or shell
32 The first argument that follows any option settings is treated as the
33 single pattern to be matched when neither -e nor -f is present. Con-
34 versely, when one or both of these options are used to specify pat-
35 terns, all arguments are treated as path names. At least one of -e, -f,
36 or an argument pattern must be provided.
38 If no files are specified, pcregrep reads the standard input. The stan-
39 dard input can also be referenced by a name consisting of a single
42 pcregrep some-pattern /file1 - /file3
44 By default, each line that matches a pattern is copied to the standard
45 output, and if there is more than one file, the file name is output at
46 the start of each line, followed by a colon. However, there are options
47 that can change how pcregrep behaves. In particular, the -M option
48 makes it possible to search for patterns that span line boundaries.
49 What defines a line boundary is controlled by the -N (--newline)
52 Patterns are limited to 8K or BUFSIZ characters, whichever is the
53 greater. BUFSIZ is defined in <stdio.h>. When there is more than one
54 pattern (specified by the use of -e and/or -f), each pattern is applied
55 to each line in the order in which they are defined, except that all
56 the -e patterns are tried before the -f patterns.
58 By default, as soon as one pattern matches (or fails to match when -v
59 is used), no further patterns are considered. However, if --colour (or
60 --color) is used to colour the matching substrings, or if --only-match-
61 ing, --file-offsets, or --line-offsets is used to output only the part
62 of the line that matched (either shown literally, or as an offset),
63 scanning resumes immediately following the match, so that further
64 matches on the same line can be found. If there are multiple patterns,
65 they are all tried on the remainder of the line, but patterns that fol-
66 low the one that matched are not tried on the earlier part of the line.
68 This is the same behaviour as GNU grep, but it does mean that the order
69 in which multiple patterns are specified can affect the output when one
70 of the above options is used.
72 Patterns that can match an empty string are accepted, but empty string
73 matches are never recognized. An example is the pattern
74 "(super)?(man)?", in which all components are optional. This pattern
75 finds all occurrences of both "super" and "man"; the output differs
76 from matching with "super|man" when only the matching substrings are
79 If the LC_ALL or LC_CTYPE environment variable is set, pcregrep uses
80 the value to set a locale when calling the PCRE library. The --locale
81 option can be used to override this.
84 SUPPORT FOR COMPRESSED FILES
86 It is possible to compile pcregrep so that it uses libz or libbz2 to
87 read files whose names end in .gz or .bz2, respectively. You can find
88 out whether your binary has support for one or both of these file types
89 by running it with the --help option. If the appropriate support is not
90 present, files are treated as plain text. The standard input is always
96 The order in which some of the options appear can affect the output.
97 For example, both the -h and -l options affect the printing of file
98 names. Whichever comes later in the command line will be the one that
101 -- This terminate the list of options. It is useful if the next
102 item on the command line starts with a hyphen but is not an
103 option. This allows for the processing of patterns and file-
104 names that start with hyphens.
106 -A number, --after-context=number
107 Output number lines of context after each matching line. If
108 filenames and/or line numbers are being output, a hyphen sep-
109 arator is used instead of a colon for the context lines. A
110 line containing "--" is output between each group of lines,
111 unless they are in fact contiguous in the input file. The
112 value of number is expected to be relatively small. However,
113 pcregrep guarantees to have up to 8K of following text avail-
114 able for context output.
116 -B number, --before-context=number
117 Output number lines of context before each matching line. If
118 filenames and/or line numbers are being output, a hyphen sep-
119 arator is used instead of a colon for the context lines. A
120 line containing "--" is output between each group of lines,
121 unless they are in fact contiguous in the input file. The
122 value of number is expected to be relatively small. However,
123 pcregrep guarantees to have up to 8K of preceding text avail-
124 able for context output.
126 -C number, --context=number
127 Output number lines of context both before and after each
128 matching line. This is equivalent to setting both -A and -B
132 Do not output individual lines from the files that are being
133 scanned; instead output the number of lines that would other-
134 wise have been shown. If no lines are selected, the number
135 zero is output. If several files are are being scanned, a
136 count is output for each of them. However, if the --files-
137 with-matches option is also used, only those files whose
138 counts are greater than zero are listed. When -c is used, the
139 -A, -B, and -C options are ignored.
142 If this option is given without any data, it is equivalent to
143 "--colour=auto". If data is required, it must be given in
144 the same shell item, separated by an equals sign.
146 --colour=value, --color=value
147 This option specifies under what circumstances the parts of a
148 line that matched a pattern should be coloured in the output.
149 By default, the output is not coloured. The value (which is
150 optional, see above) may be "never", "always", or "auto". In
151 the latter case, colouring happens only if the standard out-
152 put is connected to a terminal. More resources are used when
153 colouring is enabled, because pcregrep has to search for all
154 possible matches in a line, not just one, in order to colour
157 The colour that is used can be specified by setting the envi-
158 ronment variable PCREGREP_COLOUR or PCREGREP_COLOR. The value
159 of this variable should be a string of two numbers, separated
160 by a semicolon. They are copied directly into the control
161 string for setting colour on a terminal, so it is your
162 responsibility to ensure that they make sense. If neither of
163 the environment variables is set, the default is "1;31",
166 -D action, --devices=action
167 If an input path is not a regular file or a directory,
168 "action" specifies how it is to be processed. Valid values
169 are "read" (the default) or "skip" (silently skip the path).
171 -d action, --directories=action
172 If an input path is a directory, "action" specifies how it is
173 to be processed. Valid values are "read" (the default),
174 "recurse" (equivalent to the -r option), or "skip" (silently
175 skip the path). In the default case, directories are read as
176 if they were ordinary files. In some operating systems the
177 effect of reading a directory like this is an immediate end-
180 -e pattern, --regex=pattern, --regexp=pattern
181 Specify a pattern to be matched. This option can be used mul-
182 tiple times in order to specify several patterns. It can also
183 be used as a way of specifying a single pattern that starts
184 with a hyphen. When -e is used, no argument pattern is taken
185 from the command line; all arguments are treated as file
186 names. There is an overall maximum of 100 patterns. They are
187 applied to each line in the order in which they are defined
188 until one matches (or fails to match if -v is used). If -f is
189 used with -e, the command line patterns are matched first,
190 followed by the patterns from the file, independent of the
191 order in which these options are specified. Note that multi-
192 ple use of -e is not the same as a single pattern with alter-
193 natives. For example, X|Y finds the first character in a line
194 that is X or Y, whereas if the two patterns are given sepa-
195 rately, pcregrep finds X if it is present, even if it follows
196 Y in the line. It finds Y only if there is no X in the line.
197 This really matters only if you are using -o to show the
198 part(s) of the line that matched.
201 When pcregrep is searching the files in a directory as a con-
202 sequence of the -r (recursive search) option, any regular
203 files whose names match the pattern are excluded. Subdirecto-
204 ries are not excluded by this option; they are searched
205 recursively, subject to the --exclude-dir and --include_dir
206 options. The pattern is a PCRE regular expression, and is
207 matched against the final component of the file name (not the
208 entire path). If a file name matches both --include and
209 --exclude, it is excluded. There is no short form for this
212 --exclude-dir=pattern
213 When pcregrep is searching the contents of a directory as a
214 consequence of the -r (recursive search) option, any subdi-
215 rectories whose names match the pattern are excluded. (Note
216 that the --exclude option does not affect subdirectories.)
217 The pattern is a PCRE regular expression, and is matched
218 against the final component of the name (not the entire
219 path). If a subdirectory name matches both --include-dir and
220 --exclude-dir, it is excluded. There is no short form for
224 Interpret each pattern as a list of fixed strings, separated
225 by newlines, instead of as a regular expression. The -w
226 (match as a word) and -x (match whole line) options can be
227 used with -F. They apply to each of the fixed strings. A line
228 is selected if any of the fixed strings are found in it (sub-
229 ject to -w or -x, if present).
231 -f filename, --file=filename
232 Read a number of patterns from the file, one per line, and
233 match them against each line of input. A data line is output
234 if any of the patterns match it. The filename can be given as
235 "-" to refer to the standard input. When -f is used, patterns
236 specified on the command line using -e may also be present;
237 they are tested before the file's patterns. However, no other
238 pattern is taken from the command line; all arguments are
239 treated as file names. There is an overall maximum of 100
240 patterns. Trailing white space is removed from each line, and
241 blank lines are ignored. An empty file contains no patterns
242 and therefore matches nothing. See also the comments about
243 multiple patterns versus a single pattern with alternatives
244 in the description of -e above.
247 Instead of showing lines or parts of lines that match, show
248 each match as an offset from the start of the file and a
249 length, separated by a comma. In this mode, no context is
250 shown. That is, the -A, -B, and -C options are ignored. If
251 there is more than one match in a line, each of them is shown
252 separately. This option is mutually exclusive with --line-
253 offsets and --only-matching.
256 Force the inclusion of the filename at the start of output
257 lines when searching a single file. By default, the filename
258 is not shown in this case. For matching lines, the filename
259 is followed by a colon; for context lines, a hyphen separator
260 is used. If a line number is also being output, it follows
264 Suppress the output filenames when searching multiple files.
265 By default, filenames are shown when multiple files are
266 searched. For matching lines, the filename is followed by a
267 colon; for context lines, a hyphen separator is used. If a
268 line number is also being output, it follows the file name.
270 --help Output a help message, giving brief details of the command
271 options and file type support, and then exit.
274 Ignore upper/lower case distinctions during comparisons.
277 When pcregrep is searching the files in a directory as a con-
278 sequence of the -r (recursive search) option, only those reg-
279 ular files whose names match the pattern are included. Subdi-
280 rectories are always included and searched recursively, sub-
281 ject to the --include-dir and --exclude-dir options. The pat-
282 tern is a PCRE regular expression, and is matched against the
283 final component of the file name (not the entire path). If a
284 file name matches both --include and --exclude, it is
285 excluded. There is no short form for this option.
287 --include-dir=pattern
288 When pcregrep is searching the contents of a directory as a
289 consequence of the -r (recursive search) option, only those
290 subdirectories whose names match the pattern are included.
291 (Note that the --include option does not affect subdirecto-
292 ries.) The pattern is a PCRE regular expression, and is
293 matched against the final component of the name (not the
294 entire path). If a subdirectory name matches both --include-
295 dir and --exclude-dir, it is excluded. There is no short form
298 -L, --files-without-match
299 Instead of outputting lines from the files, just output the
300 names of the files that do not contain any lines that would
301 have been output. Each file name is output once, on a sepa-
304 -l, --files-with-matches
305 Instead of outputting lines from the files, just output the
306 names of the files containing lines that would have been out-
307 put. Each file name is output once, on a separate line.
308 Searching normally stops as soon as a matching line is found
309 in a file. However, if the -c (count) option is also used,
310 matching continues in order to obtain the correct count, and
311 those files that have at least one match are listed along
312 with their counts. Using this option with -c is a way of sup-
313 pressing the listing of files with no matches.
316 This option supplies a name to be used for the standard input
317 when file names are being output. If not supplied, "(standard
318 input)" is used. There is no short form for this option.
321 When this option is given, input is read and processed line
322 by line, and the output is flushed after each write. By
323 default, input is read in large chunks, unless pcregrep can
324 determine that it is reading from a terminal (which is cur-
325 rently possible only in Unix environments). Output to termi-
326 nal is normally automatically flushed by the operating sys-
327 tem. This option can be useful when the input or output is
328 attached to a pipe and you do not want pcregrep to buffer up
329 large amounts of data. However, its use will affect perfor-
330 mance, and the -M (multiline) option ceases to work.
333 Instead of showing lines or parts of lines that match, show
334 each match as a line number, the offset from the start of the
335 line, and a length. The line number is terminated by a colon
336 (as usual; see the -n option), and the offset and length are
337 separated by a comma. In this mode, no context is shown.
338 That is, the -A, -B, and -C options are ignored. If there is
339 more than one match in a line, each of them is shown sepa-
340 rately. This option is mutually exclusive with --file-offsets
344 This option specifies a locale to be used for pattern match-
345 ing. It overrides the value in the LC_ALL or LC_CTYPE envi-
346 ronment variables. If no locale is specified, the PCRE
347 library's default (usually the "C" locale) is used. There is
348 no short form for this option.
351 Processing some regular expression patterns can require a
352 very large amount of memory, leading in some cases to a pro-
353 gram crash if not enough is available. Other patterns may
354 take a very long time to search for all possible matching
355 strings. The pcre_exec() function that is called by pcregrep
356 to do the matching has two parameters that can limit the
357 resources that it uses.
359 The --match-limit option provides a means of limiting
360 resource usage when processing patterns that are not going to
361 match, but which have a very large number of possibilities in
362 their search trees. The classic example is a pattern that
363 uses nested unlimited repeats. Internally, PCRE uses a func-
364 tion called match() which it calls repeatedly (sometimes
365 recursively). The limit set by --match-limit is imposed on
366 the number of times this function is called during a match,
367 which has the effect of limiting the amount of backtracking
370 The --recursion-limit option is similar to --match-limit, but
371 instead of limiting the total number of times that match() is
372 called, it limits the depth of recursive calls, which in turn
373 limits the amount of memory that can be used. The recursion
374 depth is a smaller number than the total number of calls,
375 because not all calls to match() are recursive. This limit is
376 of use only if it is set smaller than --match-limit.
378 There are no short forms for these options. The default set-
379 tings are specified when the PCRE library is compiled, with
380 the default default being 10 million.
383 Allow patterns to match more than one line. When this option
384 is given, patterns may usefully contain literal newline char-
385 acters and internal occurrences of ^ and $ characters. The
386 output for any one match may consist of more than one line.
387 When this option is set, the PCRE library is called in "mul-
388 tiline" mode. There is a limit to the number of lines that
389 can be matched, imposed by the way that pcregrep buffers the
390 input file as it scans it. However, pcregrep ensures that at
391 least 8K characters or the rest of the document (whichever is
392 the shorter) are available for forward matching, and simi-
393 larly the previous 8K characters (or all the previous charac-
394 ters, if fewer than 8K) are guaranteed to be available for
395 lookbehind assertions. This option does not work when input
396 is read line by line (see --line-buffered.)
398 -N newline-type, --newline=newline-type
399 The PCRE library supports five different conventions for
400 indicating the ends of lines. They are the single-character
401 sequences CR (carriage return) and LF (linefeed), the two-
402 character sequence CRLF, an "anycrlf" convention, which rec-
403 ognizes any of the preceding three types, and an "any" con-
404 vention, in which any Unicode line ending sequence is assumed
405 to end a line. The Unicode sequences are the three just men-
406 tioned, plus VT (vertical tab, U+000B), FF (formfeed,
407 U+000C), NEL (next line, U+0085), LS (line separator,
408 U+2028), and PS (paragraph separator, U+2029).
410 When the PCRE library is built, a default line-ending
411 sequence is specified. This is normally the standard
412 sequence for the operating system. Unless otherwise specified
413 by this option, pcregrep uses the library's default. The
414 possible values for this option are CR, LF, CRLF, ANYCRLF, or
415 ANY. This makes it possible to use pcregrep on files that
416 have come from other environments without having to modify
417 their line endings. If the data that is being scanned does
418 not agree with the convention set by this option, pcregrep
419 may behave in strange ways.
422 Precede each output line by its line number in the file, fol-
423 lowed by a colon for matching lines or a hyphen for context
424 lines. If the filename is also being output, it precedes the
425 line number. This option is forced if --line-offsets is used.
428 Show only the part of the line that matched a pattern instead
429 of the whole line. In this mode, no context is shown. That
430 is, the -A, -B, and -C options are ignored. If there is more
431 than one match in a line, each of them is shown separately.
432 If -o is combined with -v (invert the sense of the match to
433 find non-matching lines), no output is generated, but the
434 return code is set appropriately. If the matched portion of
435 the line is empty, nothing is output unless the file name or
436 line number are being printed, in which case they are shown
437 on an otherwise empty line. This option is mutually exclusive
438 with --file-offsets and --line-offsets.
440 -onumber, --only-matching=number
441 Show only the part of the line that matched the capturing
442 parentheses of the given number. Up to 32 capturing parenthe-
443 ses are supported. Because these options can be given without
444 an argument (see above), if an argument is present, it must
445 be given in the same shell item, for example, -o3 or --only-
446 matching=2. The comments given for the non-argument case
447 above also apply to this case. If the specified capturing
448 parentheses do not exist in the pattern, or were not set in
449 the match, nothing is output unless the file name or line
450 number are being printed.
453 Work quietly, that is, display nothing except error messages.
454 The exit status indicates whether or not any matches were
458 If any given path is a directory, recursively scan the files
459 it contains, taking note of any --include and --exclude set-
460 tings. By default, a directory is read as a normal file; in
461 some operating systems this gives an immediate end-of-file.
462 This option is a shorthand for setting the -d option to
465 --recursion-limit=number
466 See --match-limit above.
469 Suppress error messages about non-existent or unreadable
470 files. Such files are quietly skipped. However, the return
471 code is still 2, even if matches were found in other files.
474 Operate in UTF-8 mode. This option is available only if PCRE
475 has been compiled with UTF-8 support. Both patterns and sub-
476 ject lines must be valid strings of UTF-8 characters.
479 Write the version numbers of pcregrep and the PCRE library
480 that is being used to the standard error stream.
483 Invert the sense of the match, so that lines which do not
484 match any of the patterns are the ones that are found.
486 -w, --word-regex, --word-regexp
487 Force the patterns to match only whole words. This is equiva-
488 lent to having \b at the start and end of the pattern.
490 -x, --line-regex, --line-regexp
491 Force the patterns to be anchored (each must start matching
492 at the beginning of a line) and in addition, require them to
493 match entire lines. This is equivalent to having ^ and $
494 characters at the start and end of each alternative branch in
498 ENVIRONMENT VARIABLES
500 The environment variables LC_ALL and LC_CTYPE are examined, in that
501 order, for a locale. The first one that is set is used. This can be
502 overridden by the --locale option. If no locale is set, the PCRE
503 library's default (usually the "C" locale) is used.
508 The -N (--newline) option allows pcregrep to scan files with different
509 newline conventions from the default. However, the setting of this
510 option does not affect the way in which pcregrep writes information to
511 the standard error and output streams. It uses the string "\n" in C
512 printf() calls to indicate newlines, relying on the C I/O library to
513 convert this to an appropriate sequence if the output is sent to a
517 OPTIONS COMPATIBILITY
519 Many of the short and long forms of pcregrep's options are the same as
520 in the GNU grep program (version 2.5.4). Any long option of the form
521 --xxx-regexp (GNU terminology) is also available as --xxx-regex (PCRE
522 terminology). However, the --file-offsets, --include-dir, --line-off-
523 sets, --locale, --match-limit, -M, --multiline, -N, --newline, --recur-
524 sion-limit, -u, and --utf-8 options are specific to pcregrep, as is the
525 use of the --only-matching option with a capturing parentheses number.
527 Although most of the common options work the same way, a few are dif-
528 ferent in pcregrep. For example, the --include option's argument is a
529 glob for GNU grep, but a regular expression for pcregrep. If both the
530 -c and -l options are given, GNU grep lists only file names, without
531 counts, but pcregrep gives the counts.
536 There are four different ways in which an option with data can be spec-
537 ified. If a short form option is used, the data may follow immedi-
538 ately, or (with one exception) in the next command line item. For exam-
544 The exception is the -o option, which may appear with or without data.
545 Because of this, if data is present, it must follow immediately in the
546 same item, for example -o3.
548 If a long form option is used, the data may appear in the same command
549 line item, separated by an equals character, or (with two exceptions)
550 it may appear in the next command line item. For example:
555 Note, however, that if you want to supply a file name beginning with ~
556 as data in a shell command, and have the shell expand ~ to a home
557 directory, you must separate the file name from the option, because the
558 shell does not treat ~ specially unless it is at the start of an item.
560 The exceptions to the above are the --colour (or --color) and --only-
561 matching options, for which the data is optional. If one of these
562 options does have data, it must be given in the first form, using an
563 equals character. Otherwise pcregrep will assume that it has no data.
568 It is possible to supply a regular expression that takes a very long
569 time to fail to match certain lines. Such patterns normally involve
570 nested indefinite repeats, for example: (a+)*\d when matched against a
571 line of a's with no final digit. The PCRE matching function has a
572 resource limit that causes it to abort in these circumstances. If this
573 happens, pcregrep outputs an error message and the line that caused the
574 problem to the standard error stream. If there are more than 20 such
575 errors, pcregrep gives up.
577 The --match-limit option of pcregrep can be used to set the overall
578 resource limit; there is a second option called --recursion-limit that
579 sets a limit on the amount of memory (usually stack) that is used (see
580 the discussion of these options above).
585 Exit status is 0 if any matches were found, 1 if no matches were found,
586 and 2 for syntax errors and non-existent or inacessible files (even if
587 matches were found in other files) or too many matching errors. Using
588 the -s option to suppress error messages about inaccessble files does
589 not affect the return code.
594 pcrepattern(3), pcretest(1).
600 University Computing Service
601 Cambridge CB2 3QH, England.
606 Last updated: 16 November 2010
607 Copyright (c) 1997-2010 University of Cambridge.