Navigation: Regular Expressions >

Quick reference

 

 

 

 

This document contains a quick-reference summary of the regular expression syntax.

 

Quoting

 

  \x         where x is non-alphanumeric is a literal x

  \Q...\E    treat enclosed characters as literal

 

 

 

Characters

 

  \a         alarm, that is, the BEL character (hex 07)

  \cx        "control-x", where x is any ASCII character

  \e         escape (hex 1B)

  \f         form feed (hex 0C)

  \n         newline (hex 0A)

  \r         carriage return (hex 0D)

  \t         tab (hex 09)

  \0dd       character with octal code 0dd

  \ddd       character with octal code ddd, or backreference

  \o{ddd..}  character with octal code ddd..

  \xhh       character with hex code hh

  \x{hhh..}  character with hex code hhh..

 

 

Note that \0dd is always an octal code, and that \8 and \9 are the literal characters "8" and "9".

 

 

 

Character types

 

  .          any character except newline;

             in dotall mode, any character whatsoever

  \C         one data unit, even in UTF mode (best avoided)

  \d         a decimal digit

  \D         a character that is not a decimal digit

  \h         a horizontal white space character

  \H         a character that is not a horizontal white space character

  \N         a character that is not a newline

  \p{xx}     a character with the xx property

  \P{xx}     a character without the xx property

  \R         a newline sequence

  \s         a white space character

  \S         a character that is not a white space character

  \v         a vertical white space character

  \V         a character that is not a vertical white space character

  \w         a "word" character

  \W         a "non-word" character

  \X         a Unicode extended grapheme cluster

 

 

By default, \d, \s, and \w match only ASCII characters, even in UTF-8 mode or in the 16- bit and 32-bit libraries. However, if locale-specific matching is happening, \s and \w may also match characters with code points in the range 128-255. If the PCRE_UCP option is set, the behaviour of these escape sequences is changed to use Unicode properties and they match many more characters.

 

 

 

General category properties for \p and \P

 

  C          Other

  Cc         Control

  Cf         Format

  Cn         Unassigned

  Co         Private use

  Cs         Surrogate

 

  L          Letter

  Ll         Lower case letter

  Lm         Modifier letter

  Lo         Other letter

  Lt         Title case letter

  Lu         Upper case letter

  L&         Ll, Lu, or Lt

 

  M          Mark

  Mc         Spacing mark

  Me         Enclosing mark

  Mn         Non-spacing mark

 

  N          Number

  Nd         Decimal number

  Nl         Letter number

  No         Other number

 

  P          Punctuation

  Pc         Connector punctuation

  Pd         Dash punctuation

  Pe         Close punctuation

  Pf         Final punctuation

  Pi         Initial punctuation

  Po         Other punctuation

  Ps         Open punctuation

 

  S          Symbol

  Sc         Currency symbol

  Sk         Modifier symbol

  Sm         Mathematical symbol

  So         Other symbol

 

  Z          Separator

  Zl         Line separator

  Zp         Paragraph separator

  Zs         Space separator

 

 

 

PCRE special category properties for \p and \P

 

  Xan        Alphanumeric: union of properties L and N

  Xps        POSIX space: property Z or tab, NL, VT, FF, CR

  Xsp        Perl space: property Z or tab, NL, VT, FF, CR

  Xuc        Univerally-named character: one that can be

             represented by a Universal Character Name

  Xwd        Perl word: property Xan or underscore

 

 

Perl and POSIX space are now the same. Perl added VT to its space character set at release 5.18 and PCRE changed at release 8.34.

 

 

 

Script names for \p and \P

 

Arabic, Armenian, Avestan, Balinese, Bamum, Batak, Bengali, Bopomofo, Brahmi, Braille, Buginese, Buhid, Canadian_Aboriginal, Carian, Chakma, Cham, Cherokee, Common, Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyptian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Imperial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian, Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao, Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam, Mandaic, Meetei_Mayek, Meroitic_Cursive, Meroitic_Hieroglyphs, Miao, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic, Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya, Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Sharada, Shavian, Sinhala, Sora_Sompeng, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le, Tai_Tham, Tai_Viet, Takri, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Vai, Yi.

 

 

 

Character classes

 

  [...]       positive character class

  [^...]      negative character class

  [x-y]       range (can be used for hex characters)

  [[:xxx:]]   positive POSIX named set

  [[:^xxx:]]  negative POSIX named set

 

  alnum       alphanumeric

  alpha       alphabetic

  ascii       0-127

  blank       space or tab

  cntrl       control character

  digit       decimal digit

  graph       printing, excluding space

  lower       lower case letter

  print       printing, including space

  punct       printing, excluding alphanumeric

  space       white space

  upper       upper case letter

  word        same as \w

  xdigit      hexadecimal digit

 

In PCRE, POSIX character set names recognize only ASCII characters by default, but some of them use Unicode properties if PCRE_UCP is set. You can use \Q...\E inside a character class.

 

 

 

Quantifiers

 

  ?           0 or 1, greedy

  ?+          0 or 1, possessive

  ??          0 or 1, lazy

  *           0 or more, greedy

  *+          0 or more, possessive

  *?          0 or more, lazy

  +           1 or more, greedy

  ++          1 or more, possessive

  +?          1 or more, lazy

  {n}         exactly n

  {n,m}       at least n, no more than m, greedy

  {n,m}+      at least n, no more than m, possessive

  {n,m}?      at least n, no more than m, lazy

  {n,}        n or more, greedy

  {n,}+       n or more, possessive

  {n,}?       n or more, lazy

 

 

 

Anchors and simple assertions

 

  \b          word boundary

  \B          not a word boundary

  ^           start of subject

              also after internal newline in multiline mode

  \A          start of subject

  $           end of subject

              also before newline at end of subject

              also before internal newline in multiline mode

  \Z          end of subject

              also before newline at end of subject

  \z          end of subject

  \G          first matching position in subject

 

 

 

Match point reset

 

  \K          reset start of match

 

 

Alternation

 

  expr|expr|expr...

 

 

Capturing

 

  (...)           capturing group

  (?<name>...)    named capturing group (Perl)

  (?'name'...)    named capturing group (Perl)

  (?P<name>...)   named capturing group (Python)

  (?:...)         non-capturing group

  (?|...)         non-capturing group; reset group numbers for

                  capturing groups in each alternative

 

 

Atomic groups

 

  (?>...)         atomic, non-capturing group

 

 

Comment

 

  (?#....)        comment (not nestable)

 

 

Option settings

 

  (?i)            caseless

  (?J)            allow duplicate names

  (?m)            multiline

  (?s)            single line (dotall)

  (?U)            default ungreedy (lazy)

  (?x)            extended (ignore white space)

  (?-...)         unset option(s)

 

 

The following are recognized only at the start of a pattern or after one of the newline-setting options with similar syntax:

 

  (*LIMIT_MATCH=d)            set the match limit to d (decimal number)

  (*LIMIT_RECURSION=d)        set the recursion limit to d (decimal number)

  (*NO_START_OPT)             no start-match optimization (PCRE_NO_START_OPTIMIZE)

  (*UTF8)                     set UTF-8 mode: 8-bit library (PCRE_UTF8)

  (*UTF16)                    set UTF-16 mode: 16-bit library (PCRE_UTF16)

  (*UTF32)                    set UTF-32 mode: 32-bit library (PCRE_UTF32)

  (*UTF)                      set appropriate UTF mode for the library in use

  (*UCP)                      set PCRE_UCP (use Unicode properties for \d etc)

 

 

Note that LIMIT_MATCH and LIMIT_RECURSION can only reduce the value of the limits set by the caller of pcre_exec(), not increase them.

 

 

 

Lookahead and lookbehind assertions

 

  (?=...)         positive look ahead

  (?!...)         negative look ahead

  (?<=...)        positive look behind

  (?<!...)        negative look behind

 

 

Each top-level branch of a look behind must be of a fixed length.

 

 

 

Backreference

 

  \n              reference by number (can be ambiguous)

  \gn             reference by number

  \g{n}           reference by number

  \g{-n}          relative reference by number

  \k<name>        reference by name (Perl)

  \k'name'        reference by name (Perl)

  \g{name}        reference by name (Perl)

  \k{name}        reference by name (.NET)

  (?P=name)       reference by name (Python)

 

 

 

Subroutine references (possibly recursive)

 

  (?R)            recurse whole pattern

  (?n)            call subpattern by absolute number

  (?+n)           call subpattern by relative number

  (?-n)           call subpattern by relative number

  (?&name)        call subpattern by name (Perl)

  (?P>name)       call subpattern by name (Python)

  \g<name>        call subpattern by name (Oniguruma)

  \g'name'        call subpattern by name (Oniguruma)

  \g<n>           call subpattern by absolute number (Oniguruma)

  \g'n'           call subpattern by absolute number (Oniguruma)

  \g<+n>          call subpattern by relative number (PCRE extension)

  \g'+n'          call subpattern by relative number (PCRE extension)

  \g<-n>          call subpattern by relative number (PCRE extension)

  \g'-n'          call subpattern by relative number (PCRE extension)

 

 

 

Conditional patterns

 

  (?(condition)yes-pattern)

  (?(condition)yes-pattern|no-pattern)

 

  (?(n)...        absolute reference condition

  (?(+n)...       relative reference condition

  (?(-n)...       relative reference condition

  (?(<name>)...   named reference condition (Perl)

  (?('name')...   named reference condition (Perl)

  (?(name)...     named reference condition (PCRE)

  (?(R)...        overall recursion condition

  (?(Rn)...       specific group recursion condition

  (?(R&name)...   specific recursion condition

  (?(DEFINE)...   define subpattern for reference

  (?(assert)...   assertion condition

 

 

 

Backtracking control

 

The following act immediately they are reached:

 

  (*ACCEPT)       force successful match

  (*FAIL)         force backtrack; synonym (*F)

  (*MARK:NAME)    set name to be passed back; synonym (*:NAME)

 

 

The following act only when a subsequent match failure causes a backtrack to reach them. They all force a match failure, but they differ in what happens afterwards. Those that advance the start-of-match point do so only if the pattern is not anchored.

 

  (*COMMIT)       overall failure, no advance of starting point

  (*PRUNE)        advance to next starting character

  (*PRUNE:NAME)   equivalent to (*MARK:NAME)(*PRUNE)

  (*SKIP)         advance to current matching position

  (*SKIP:NAME)    advance to position corresponding to an earlier

                  (*MARK:NAME); if not found, the (*SKIP) is ignored

  (*THEN)         local failure, backtrack to next alternation

  (*THEN:NAME)    equivalent to (*MARK:NAME)(*THEN)

 

 

 

Newline conventions

 

These are recognized only at the very start of the pattern or after a (*BSR_...), (*UTF8), (*UTF16), (*UTF32) or (*UCP) option.

 

  (*CR)           carriage return only

  (*LF)           linefeed only

  (*CRLF)         carriage return followed by linefeed

  (*ANYCRLF)      all three of the above

  (*ANY)          any Unicode newline sequence

 

 

 

What \R matches

 

These are recognized only at the very start of the pattern or after a (*...) option that sets the newline convention or a UTF or UCP mode.

 

  (*BSR_ANYCRLF)  CR, LF, or CRLF

  (*BSR_UNICODE)  any Unicode newline sequence

 

 

 

Callouts

 

  (?C)      callout

  (?Cn)     callout with data n

 

 

 

 

 

Philip Hazel

University Computing Service

Cambridge CB2 3QH, England.

Last updated: 12 November 2013

Copyright © 1997-2013 University of Cambridge.

 

 

 

 

 

 

 

Copyright © 2022 Rickard Johansson