3 perlreref - Perl Regular Expressions Reference
7 This is a quick reference to Perl's regular expressions.
8 For full information see L<perlre> and L<perlop>, as well
9 as the L</"SEE ALSO"> section in this document.
13 C<=~> determines to which variable the regex is applied.
14 In its absence, $_ is used.
18 C<!~> determines to which variable the regex is applied,
19 and negates the result of the match; it returns
20 false if the match succeeds, and true if it fails.
24 C<m/pattern/msixpogc> searches a string for a pattern match,
25 applying the given options.
27 m Multiline mode - ^ and $ match internal lines
28 s match as a Single line - . matches \n
30 x eXtended legibility - free whitespace and comments
31 p Preserve a copy of the matched string -
32 ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be defined.
33 o compile pattern Once
34 g Global - all occurrences
35 c don't reset pos on failed matches when using /g
37 If 'pattern' is an empty string, the last I<successfully> matched
38 regex is used. Delimiters other than '/' may be used for both this
39 operator and the following ones. The leading C<m> can be omitted
40 if the delimiter is '/'.
42 C<qr/pattern/msixpo> lets you store a regex in a variable,
43 or pass one around. Modifiers as for C<m//>, and are stored
46 C<s/pattern/replacement/msixpogce> substitutes matches of
47 'pattern' with 'replacement'. Modifiers as for C<m//>,
50 e Evaluate 'replacement' as an expression
52 'e' may be specified multiple times. 'replacement' is interpreted
53 as a double quoted string unless a single-quote (C<'>) is the delimiter.
55 C<?pattern?> is like C<m/pattern/> but matches only once. No alternate
56 delimiters can be used. Must be reset with reset().
60 \ Escapes the character immediately following it
61 . Matches any single character except a newline (unless /s is
63 ^ Matches at the beginning of the string (or line, if /m is used)
64 $ Matches at the end of the string (or line, if /m is used)
65 * Matches the preceding element 0 or more times
66 + Matches the preceding element 1 or more times
67 ? Matches the preceding element 0 or 1 times
68 {...} Specifies a range of occurrences for the element preceding it
69 [...] Matches any one of the characters contained within the brackets
70 (...) Groups subexpressions for capturing to $1, $2...
71 (?:...) Groups subexpressions without capturing (cluster)
72 | Matches either the subexpression preceding or following it
73 \1, \2, \3 ... Matches the text from the Nth group
74 \g1 or \g{1}, \g2 ... Matches the text from the Nth group
75 \g-1 or \g{-1}, \g-2 ... Matches the text from the Nth previous group
76 \g{name} Named backreference
77 \k<name> Named backreference
78 \k'name' Named backreference
79 (?P=name) Named backreference (python syntax)
81 =head2 ESCAPE SEQUENCES
83 These work as in normal strings.
91 \037 Any octal ASCII value
92 \x7f Any hexadecimal ASCII value
93 \x{263a} A wide hexadecimal value
95 \N{name} A named character
96 \N{U+263D} A Unicode character by hex ordinal
98 \l Lowercase next character
99 \u Titlecase next character
100 \L Lowercase until \E
101 \U Uppercase until \E
102 \Q Disable pattern metacharacters until \E
105 For Titlecase, see L</Titlecase>.
107 This one works differently from normal strings:
109 \b An assertion, not backspace, except in a character class
111 =head2 CHARACTER CLASSES
113 [amy] Match 'a', 'm' or 'y'
114 [f-j] Dash specifies "range"
115 [f-j-] Dash escaped or at start or end means 'dash'
116 [^f-j] Caret indicates "match any character _except_ these"
118 The following sequences (except C<\N>) work within or without a character class.
119 The first six are locale aware, all are Unicode aware. See L<perllocale>
120 and L<perlunicode> for details.
125 \W A non-word character
126 \s A whitespace character
127 \S A non-whitespace character
128 \h An horizontal whitespace
129 \H A non horizontal whitespace
130 \N A non newline (when not followed by '{NAME}'; experimental;
131 not valid in a character class; equivalent to [^\n]; it's
132 like '.' without /s modifier)
133 \v A vertical whitespace
134 \V A non vertical whitespace
135 \R A generic newline (?>\v|\x0D\x0A)
137 \C Match a byte (with Unicode, '.' matches a character)
138 \pP Match P-named (Unicode) property
139 \p{...} Match Unicode property with name longer than 1 character
141 \P{...} Match lack of Unicode property with name longer than 1 char
142 \X Match Unicode extended grapheme cluster
144 POSIX character classes and their Unicode and Perl equivalents:
147 range range backslash
148 POSIX \p{...} \p{} sequence Description
149 -----------------------------------------------------------------------
150 alnum PosixAlnum Alnum Alpha plus Digit
151 alpha PosixAlpha Alpha Alphabetic characters
152 ascii ASCII Any ASCII character
153 blank PosixBlank Blank \h Horizontal whitespace;
154 full-range also written
155 as \p{HorizSpace} (GNU
157 cntrl PosixCntrl Cntrl Control characters
158 digit PosixDigit Digit \d Decimal digits
159 graph PosixGraph Graph Alnum plus Punct
160 lower PosixLower Lower Lowercase characters
161 print PosixPrint Print Graph plus Print, but not
163 punct PosixPunct Punct These aren't precisely
164 equivalent. See NOTE,
166 space PosixSpace Space [\s\cK] Whitespace
167 PerlSpace SpacePerl \s Perl's whitespace
169 upper PosixUpper Upper Uppercase characters
170 word PerlWord Word \w Alnum plus '_' (Perl
172 xdigit ASCII_Hex_Digit XDigit Hexadecimal digit,
176 NOTE on C<[[:punct:]]>, C<\p{PosixPunct}> and C<\p{Punct}>:
177 In the ASCII range, C<[[:punct:]]> and C<\p{PosixPunct}> match
178 C<[-!"#$%&'()*+,./:;<=E<gt>?@[\\\]^_`{|}~]> (although if a locale is in
179 effect, it could alter the behavior of C<[[:punct:]]>); and C<\p{Punct}>
180 matches C<[-!"#%&'()*,./:;?@[\\\]_{}]>. When matching a UTF-8 string,
181 C<[[:punct:]]> matches what it does in the ASCII range, plus what
182 C<\p{Punct}> matches. C<\p{Punct}> matches, anything that isn't a
183 control, an alphanumeric, a space, nor a symbol.
185 Within a character class:
187 POSIX traditional Unicode
188 [:digit:] \d \p{Digit}
189 [:^digit:] \D \P{Digit}
193 All are zero-width assertions.
195 ^ Match string start (or line, if /m is used)
196 $ Match string end (or line, if /m is used) or before newline
197 \b Match word boundary (between \w and \W)
198 \B Match except at word boundary (between \w and \w or \W and \W)
199 \A Match string start (regardless of /m)
200 \Z Match string end (before optional newline)
201 \z Match absolute string end
202 \G Match where previous m//g left off
203 \K Keep the stuff left of the \K, don't include it in $&
207 Quantifiers are greedy by default and match the B<longest> leftmost.
209 Maximal Minimal Possessive Allowed range
210 ------- ------- ---------- -------------
211 {n,m} {n,m}? {n,m}+ Must occur at least n times
212 but no more than m times
213 {n,} {n,}? {n,}+ Must occur at least n times
214 {n} {n}? {n}+ Must occur exactly n times
215 * *? *+ 0 or more times (same as {0,})
216 + +? ++ 1 or more times (same as {1,})
217 ? ?? ?+ 0 or 1 time (same as {0,1})
219 The possessive forms (new in Perl 5.10) prevent backtracking: what gets
220 matched by a pattern with a possessive quantifier will not be backtracked
221 into, even if that causes the whole match to fail.
223 There is no quantifier C<{,n}>. That's interpreted as a literal string.
225 =head2 EXTENDED CONSTRUCTS
228 (?:...) Groups subexpressions without capturing (cluster)
229 (?pimsx-imsx:...) Enable/disable option (as per m// modifiers)
230 (?=...) Zero-width positive lookahead assertion
231 (?!...) Zero-width negative lookahead assertion
232 (?<=...) Zero-width positive lookbehind assertion
233 (?<!...) Zero-width negative lookbehind assertion
234 (?>...) Grab what we can, prohibit backtracking
236 (?<name>...) Named capture
237 (?'name'...) Named capture
238 (?P<name>...) Named capture (python syntax)
239 (?{ code }) Embedded code, return value becomes $^R
240 (??{ code }) Dynamic regex, return value used as regex
241 (?N) Recurse into subpattern number N
242 (?-N), (?+N) Recurse into Nth previous/next subpattern
243 (?R), (?0) Recurse at the beginning of the whole pattern
244 (?&name) Recurse into a named subpattern
245 (?P>name) Recurse into a named subpattern (python syntax)
247 (?(cond)yes) Conditional expression, where "cond" can be:
248 (N) subpattern N has matched something
249 (<name>) named subpattern has matched something
250 ('name') named subpattern has matched something
251 (?{code}) code condition
252 (R) true if recursing
253 (RN) true if recursing into Nth subpattern
254 (R&name) true if recursing into named subpattern
255 (DEFINE) always false, no no-pattern allowed
259 $_ Default variable for operators to use
261 $` Everything prior to matched string
262 $& Entire matched string
263 $' Everything after to matched string
265 ${^PREMATCH} Everything prior to matched string
266 ${^MATCH} Entire matched string
267 ${^POSTMATCH} Everything after to matched string
269 The use of C<$`>, C<$&> or C<$'> will slow down B<all> regex use
270 within your program. Consult L<perlvar> for C<@->
271 to see equivalent expressions that won't cause slow down.
272 See also L<Devel::SawAmpersand>. Starting with Perl 5.10, you
273 can also use the equivalent variables C<${^PREMATCH}>, C<${^MATCH}>
274 and C<${^POSTMATCH}>, but for them to be defined, you have to
275 specify the C</p> (preserve) modifier on your regular expression.
277 $1, $2 ... hold the Xth captured expr
278 $+ Last parenthesized pattern match
279 $^N Holds the most recently closed capture
280 $^R Holds the result of the last (?{...}) expr
281 @- Offsets of starts of groups. $-[0] holds start of whole match
282 @+ Offsets of ends of groups. $+[0] holds end of whole match
283 %+ Named capture buffers
284 %- Named capture buffers, as array refs
286 Captured groups are numbered according to their I<opening> paren.
290 lc Lowercase a string
291 lcfirst Lowercase first char of a string
292 uc Uppercase a string
293 ucfirst Titlecase first char of a string
295 pos Return or set current match position
296 quotemeta Quote metacharacters
297 reset Reset ?pattern? status
298 study Analyze string for optimizing matching
300 split Use a regex to split a string into parts
302 The first four of these are like the escape sequences C<\L>, C<\l>,
303 C<\U>, and C<\u>. For Titlecase, see L</Titlecase>.
309 Unicode concept which most often is equal to uppercase, but for
310 certain characters like the German "sharp s" there is a difference.
314 Iain Truskett. Updated by the Perl 5 Porters.
316 This document may be distributed under the same terms as Perl itself.
324 L<perlretut> for a tutorial on regular expressions.
328 L<perlrequick> for a rapid tutorial.
332 L<perlre> for more details.
336 L<perlvar> for details on the variables.
340 L<perlop> for details on the operators.
344 L<perlfunc> for details on the functions.
348 L<perlfaq6> for FAQs on regular expressions.
352 L<perlrebackslash> for a reference on backslash sequences.
356 L<perlrecharclass> for a reference on character classes.
360 The L<re> module to alter behaviour and aid
365 L<perldebug/"Debugging regular expressions">
369 L<perluniintro>, L<perlunicode>, L<charnames> and L<perllocale>
370 for details on regexes and internationalisation.
374 I<Mastering Regular Expressions> by Jeffrey Friedl
375 (F<http://oreilly.com/catalog/9780596528126/>) for a thorough grounding and
376 reference on the topic.