1PERLRE(1) Perl Programmers Reference Guide PERLRE(1)
2
6 perlre - Perl regular expressions
7
9 This page describes the syntax of regular expressions in Perl.
10 If you haven't used regular expressions before, a tutorial introduction
12 is available in perlretut. If you know just a little about them, a
13 quick-start introduction is available in perlrequick.
14 Except for "The Basics" section, this page assumes you are familiar
16 with regular expression basics, like what is a "pattern", what does it
17 look like, and how it is basically used. For a reference on how they
18 are used, plus various examples of the same, see discussions of "m//",
19 "s///", "qr//" and "??" in "Regexp Quote-Like Operators" in perlop.
20 New in v5.22, "use re 'strict'" applies stricter rules than otherwise
22 when compiling regular expression patterns. It can find things that,
23 while legal, may not be what you intended.
24 The Basics
26 Regular expressions are strings with the very particular syntax and
27 meaning described in this document and auxiliary documents referred to
28 by this one. The strings are called "patterns". Patterns are used to
29 determine if some other string, called the "target", has (or doesn't
30 have) the characteristics specified by the pattern. We call this
31 "matching" the target string against the pattern. Usually the match is
32 done by having the target be the first operand, and the pattern be the
33 second operand, of one of the two binary operators "=~" and "!~",
34 listed in "Binding Operators" in perlop; and the pattern will have been
35 converted from an ordinary string by one of the operators in "Regexp
36 Quote-Like Operators" in perlop, like so:
37 $foo =~ m/abc/
39 This evaluates to true if and only if the string in the variable $foo
41 contains somewhere in it, the sequence of characters "a", "b", then
42 "c". (The "=~ m", or match operator, is described in
43 "m/PATTERN/msixpodualngc" in perlop.)
44 Patterns that aren't already stored in some variable must be
46 delimitted, at both ends, by delimitter characters. These are often,
47 as in the example above, forward slashes, and the typical way a pattern
48 is written in documentation is with those slashes. In most cases, the
49 delimitter is the same character, fore and aft, but there are a few
50 cases where a character looks like it has a mirror-image mate, where
51 the opening version is the beginning delimiter, and the closing one is
52 the ending delimiter, like
53 $foo =~ m<abc>
55 Most times, the pattern is evaluated in double-quotish context, but it
57 is possible to choose delimiters to force single-quotish, like
58 $foo =~ m'abc'
60 If the pattern contains its delimiter within it, that delimiter must be
62 escaped. Prefixing it with a backslash (e.g., "/foo\/bar/") serves
63 this purpose.
64 Any single character in a pattern matches that same character in the
66 target string, unless the character is a metacharacter with a special
67 meaning described in this document. A sequence of non-metacharacters
68 matches the same sequence in the target string, as we saw above with
69 "m/abc/".
70 Only a few characters (all of them being ASCII punctuation characters)
72 are metacharacters. The most commonly used one is a dot ".", which
73 normally matches almost any character (including a dot itself).
74 You can cause characters that normally function as metacharacters to be
76 interpreted literally by prefixing them with a "\", just like the
77 pattern's delimiter must be escaped if it also occurs within the
78 pattern. Thus, "\." matches just a literal dot, "." instead of its
79 normal meaning. This means that the backslash is also a metacharacter,
80 so "\\" matches a single "\". And a sequence that contains an escaped
81 metacharacter matches the same sequence (but without the escape) in the
82 target string. So, the pattern "/blur\\fl/" would match any target
83 string that contains the sequence "blur\fl".
84 The metacharacter "|" is used to match one thing or another. Thus
86 $foo =~ m/this|that/
88 is TRUE if and only if $foo contains either the sequence "this" or the
90 sequence "that". Like all metacharacters, prefixing the "|" with a
91 backslash makes it match the plain punctuation character; in its case,
92 the VERTICAL LINE.
93 $foo =~ m/this\|that/
95 is TRUE if and only if $foo contains the sequence "this|that".
97 You aren't limited to just a single "|".
99 $foo =~ m/fee|fie|foe|fum/
101 is TRUE if and only if $foo contains any of those 4 sequences from the
103 children's story "Jack and the Beanstalk".
104 As you can see, the "|" binds less tightly than a sequence of ordinary
106 characters. We can override this by using the grouping metacharacters,
107 the parentheses "(" and ")".
108 $foo =~ m/th(is|at) thing/
110 is TRUE if and only if $foo contains either the sequence "this thing"
112 or the sequence "that thing". The portions of the string that match
113 the portions of the pattern enclosed in parentheses are normally made
114 available separately for use later in the pattern, substitution, or
115 program. This is called "capturing", and it can get complicated. See
116 "Capture groups".
117 The first alternative includes everything from the last pattern
119 delimiter ("(", "(?:" (described later), etc. or the beginning of the
120 pattern) up to the first "|", and the last alternative contains
121 everything from the last "|" to the next closing pattern delimiter.
122 That's why it's common practice to include alternatives in parentheses:
123 to minimize confusion about where they start and end.
124 Alternatives are tried from left to right, so the first alternative
126 found for which the entire expression matches, is the one that is
127 chosen. This means that alternatives are not necessarily greedy. For
128 example: when matching "foo|foot" against "barefoot", only the "foo"
129 part will match, as that is the first alternative tried, and it
130 successfully matches the target string. (This might not seem important,
131 but it is important when you are capturing matched text using
132 parentheses.)
133 Besides taking away the special meaning of a metacharacter, a prefixed
135 backslash changes some letter and digit characters away from matching
136 just themselves to instead have special meaning. These are called
137 "escape sequences", and all such are described in perlrebackslash. A
138 backslash sequence (of a letter or digit) that doesn't currently have
139 special meaning to Perl will raise a warning if warnings are enabled,
140 as those are reserved for potential future use.
141 One such sequence is "\b", which matches a boundary of some sort.
143 "\b{wb}" and a few others give specialized types of boundaries. (They
144 are all described in detail starting at "\b{}, \b, \B{}, \B" in
145 perlrebackslash.) Note that these don't match characters, but the
146 zero-width spaces between characters. They are an example of a zero-
147 width assertion. Consider again,
148 $foo =~ m/fee|fie|foe|fum/
150 It evaluates to TRUE if, besides those 4 words, any of the sequences
152 "feed", "field", "Defoe", "fume", and many others are in $foo. By
153 judicious use of "\b" (or better (because it is designed to handle
154 natural language) "\b{wb}"), we can make sure that only the Giant's
155 words are matched:
156 $foo =~ m/\b(fee|fie|foe|fum)\b/
158 $foo =~ m/\b{wb}(fee|fie|foe|fum)\b{wb}/
159 The final example shows that the characters "{" and "}" are
161 metacharacters.
162 Another use for escape sequences is to specify characters that cannot
164 (or which you prefer not to) be written literally. These are described
165 in detail in "Character Escapes" in perlrebackslash, but the next three
166 paragraphs briefly describe some of them.
167 Various control characters can be written in C language style: "\n"
169 matches a newline, "\t" a tab, "\r" a carriage return, "\f" a form
170 feed, etc.
171 More generally, "\nnn", where nnn is a string of three octal digits,
173 matches the character whose native code point is nnn. You can easily
174 run into trouble if you don't have exactly three digits. So always use
175 three, or since Perl 5.14, you can use "\o{...}" to specify any number
176 of octal digits.
177 Similarly, "\xnn", where nn are hexadecimal digits, matches the
179 character whose native ordinal is nn. Again, not using exactly two
180 digits is a recipe for disaster, but you can use "\x{...}" to specify
181 any number of hex digits.
182 Besides being a metacharacter, the "." is an example of a "character
184 class", something that can match any single character of a given set of
185 them. In its case, the set is just about all possible characters.
186 Perl predefines several character classes besides the "."; there is a
187 separate reference page about just these, perlrecharclass.
188 You can define your own custom character classes, by putting into your
190 pattern in the appropriate place(s), a list of all the characters you
191 want in the set. You do this by enclosing the list within "[]" bracket
192 characters. These are called "bracketed character classes" when we are
193 being precise, but often the word "bracketed" is dropped. (Dropping it
194 usually doesn't cause confusion.) This means that the "[" character is
195 another metacharacter. It doesn't match anything just by itself; it is
196 used only to tell Perl that what follows it is a bracketed character
197 class. If you want to match a literal left square bracket, you must
198 escape it, like "\[". The matching "]" is also a metacharacter; again
199 it doesn't match anything by itself, but just marks the end of your
200 custom class to Perl. It is an example of a "sometimes metacharacter".
201 It isn't a metacharacter if there is no corresponding "[", and matches
202 its literal self:
203 print "]" =~ /]/; # prints 1
205 The list of characters within the character class gives the set of
207 characters matched by the class. "[abc]" matches a single "a" or "b"
208 or "c". But if the first character after the "[" is "^", the class
209 instead matches any character not in the list. Within a list, the "-"
210 character specifies a range of characters, so that "a-z" represents all
211 characters between "a" and "z", inclusive. If you want either "-" or
212 "]" itself to be a member of a class, put it at the start of the list
213 (possibly after a "^"), or escape it with a backslash. "-" is also
214 taken literally when it is at the end of the list, just before the
215 closing "]". (The following all specify the same class of three
216 characters: "[-az]", "[az-]", and "[a\-z]". All are different from
217 "[a-z]", which specifies a class containing twenty-six characters, even
218 on EBCDIC-based character sets.)
219 There is lots more to bracketed character classes; full details are in
221 "Bracketed Character Classes" in perlrecharclass.
222 Metacharacters
224 "The Basics" introduced some of the metacharacters. This section gives
226 them all. Most of them have the same meaning as in the egrep command.
227 Only the "\" is always a metacharacter. The others are metacharacters
229 just sometimes. The following tables lists all of them, summarizes
230 their use, and gives the contexts where they are metacharacters.
231 Outside those contexts or if prefixed by a "\", they match their
232 corresponding punctuation character. In some cases, their meaning
233 varies depending on various pattern modifiers that alter the default
234 behaviors. See "Modifiers".
235 PURPOSE WHERE
237 \ Escape the next character Always, except when
238 escaped by another \
239 ^ Match the beginning of the string Not in []
240 (or line, if /m is used)
241 ^ Complement the [] class At the beginning of []
242 . Match any single character except newline Not in []
243 (under /s, includes newline)
244 $ Match the end of the string Not in [], but can
245 (or before newline at the end of the mean interpolate a
246 string; or before any newline if /m is scalar
247 used)
248 | Alternation Not in []
249 () Grouping Not in []
250 [ Start Bracketed Character class Not in []
251 ] End Bracketed Character class Only in [], and
252 not first
253 * Matches the preceding element 0 or more Not in []
254 times
255 + Matches the preceding element 1 or more Not in []
256 times
257 ? Matches the preceding element 0 or 1 Not in []
258 times
259 { Starts a sequence that gives number(s) Not in []
260 of times the preceding element can be
261 matched
262 { when following certain escape sequences
263 starts a modifier to the meaning of the
264 sequence
265 } End sequence started by {
266 - Indicates a range Only in [] interior
267 # Beginning of comment, extends to line end Only with /x modifier
268 Notice that most of the metacharacters lose their special meaning when
270 they occur in a bracketed character class, except "^" has a different
271 meaning when it is at the beginning of such a class. And "-" and "]"
272 are metacharacters only at restricted positions within bracketed
273 character classes; while "}" is a metacharacter only when closing a
274 special construct started by "{".
275 In double-quotish context, as is usually the case, you need to be
277 careful about "$" and the non-metacharacter "@". Those could
278 interpolate variables, which may or may not be what you intended.
279 These rules were designed for compactness of expression, rather than
281 legibility and maintainability. The "/x and /xx" pattern modifiers
282 allow you to insert white space to improve readability. And use of
283 "re 'strict'" adds extra checking to catch some typos that might
284 silently compile into something unintended.
285 By default, the "^" character is guaranteed to match only the beginning
287 of the string, the "$" character only the end (or before the newline at
288 the end), and Perl does certain optimizations with the assumption that
289 the string contains only one line. Embedded newlines will not be
290 matched by "^" or "$". You may, however, wish to treat a string as a
291 multi-line buffer, such that the "^" will match after any newline
292 within the string (except if the newline is the last character in the
293 string), and "$" will match before any newline. At the cost of a
294 little more overhead, you can do this by using the "/m" modifier on the
295 pattern match operator. (Older programs did this by setting $*, but
296 this option was removed in perl 5.10.)
297 To simplify multi-line substitutions, the "." character never matches a
299 newline unless you use the "/s" modifier, which in effect tells Perl to
300 pretend the string is a single line--even if it isn't.
301 Modifiers
303 Overview
304 The default behavior for matching can be changed, using various
306 modifiers. Modifiers that relate to the interpretation of the pattern
307 are listed just below. Modifiers that alter the way a pattern is used
308 by Perl are detailed in "Regexp Quote-Like Operators" in perlop and
309 "Gory details of parsing quoted constructs" in perlop. Modifiers can
310 be added dynamically; see "Extended Patterns" below.
311 "m" Treat the string being matched against as multiple lines. That is,
313 change "^" and "$" from matching the start of the string's first
314 line and the end of its last line to matching the start and end of
315 each line within the string.
316 "s" Treat the string as single line. That is, change "." to match any
318 character whatsoever, even a newline, which normally it would not
319 match.
320 Used together, as "/ms", they let the "." match any character
322 whatsoever, while still allowing "^" and "$" to match,
323 respectively, just after and just before newlines within the
324 string.
325 "i" Do case-insensitive pattern matching. For example, "A" will match
327 "a" under "/i".
328 If locale matching rules are in effect, the case map is taken from
330 the current locale for code points less than 255, and from Unicode
331 rules for larger code points. However, matches that would cross
332 the Unicode rules/non-Unicode rules boundary (ords 255/256) will
333 not succeed, unless the locale is a UTF-8 one. See perllocale.
334 There are a number of Unicode characters that match a sequence of
336 multiple characters under "/i". For example, "LATIN SMALL LIGATURE
337 FI" should match the sequence "fi". Perl is not currently able to
338 do this when the multiple characters are in the pattern and are
339 split between groupings, or when one or more are quantified. Thus
340 "\N{LATIN SMALL LIGATURE FI}" =~ /fi/i; # Matches
342 "\N{LATIN SMALL LIGATURE FI}" =~ /[fi][fi]/i; # Doesn't match!
343 "\N{LATIN SMALL LIGATURE FI}" =~ /fi*/i; # Doesn't match!
344 # The below doesn't match, and it isn't clear what $1 and $2 would
346 # be even if it did!!
347 "\N{LATIN SMALL LIGATURE FI}" =~ /(f)(i)/i; # Doesn't match!
348 Perl doesn't match multiple characters in a bracketed character
350 class unless the character that maps to them is explicitly
351 mentioned, and it doesn't match them at all if the character class
352 is inverted, which otherwise could be highly confusing. See
353 "Bracketed Character Classes" in perlrecharclass, and "Negation" in
354 perlrecharclass.
355 "x" and "xx"
357 Extend your pattern's legibility by permitting whitespace and
358 comments. Details in "/x and /xx"
359 "p" Preserve the string matched such that "${^PREMATCH}", "${^MATCH}",
361 and "${^POSTMATCH}" are available for use after matching.
362 In Perl 5.20 and higher this is ignored. Due to a new copy-on-write
364 mechanism, "${^PREMATCH}", "${^MATCH}", and "${^POSTMATCH}" will be
365 available after the match regardless of the modifier.
366 "a", "d", "l", and "u"
368 These modifiers, all new in 5.14, affect which character-set rules
369 (Unicode, etc.) are used, as described below in "Character set
370 modifiers".
371 "n" Prevent the grouping metacharacters "()" from capturing. This
373 modifier, new in 5.22, will stop $1, $2, etc... from being filled
374 in.
375 "hello" =~ /(hi|hello)/; # $1 is "hello"
377 "hello" =~ /(hi|hello)/n; # $1 is undef
378 This is equivalent to putting "?:" at the beginning of every
380 capturing group:
381 "hello" =~ /(?:hi|hello)/; # $1 is undef
383 "/n" can be negated on a per-group basis. Alternatively, named
385 captures may still be used.
386 "hello" =~ /(?-n:(hi|hello))/n; # $1 is "hello"
388 "hello" =~ /(?<greet>hi|hello)/n; # $1 is "hello", $+{greet} is
389 # "hello"
390 Other Modifiers
392 There are a number of flags that can be found at the end of regular
393 expression constructs that are not generic regular expression
394 flags, but apply to the operation being performed, like matching or
395 substitution ("m//" or "s///" respectively).
396 Flags described further in "Using regular expressions in Perl" in
398 perlretut are:
399 c - keep the current position during repeated matching
401 g - globally match the pattern repeatedly in the string
402 Substitution-specific modifiers described in
404 "s/PATTERN/REPLACEMENT/msixpodualngcer" in perlop are:
405 e - evaluate the right-hand side as an expression
407 ee - evaluate the right side as a string then eval the result
408 o - pretend to optimize your code, but actually introduce bugs
409 r - perform non-destructive substitution and return the new value
410 Regular expression modifiers are usually written in documentation as
412 e.g., "the "/x" modifier", even though the delimiter in question might
413 not really be a slash. The modifiers "/imnsxadlup" may also be
414 embedded within the regular expression itself using the "(?...)"
415 construct, see "Extended Patterns" below.
416 Details on some modifiers
418 Some of the modifiers require more explanation than given in the
420 "Overview" above.
421 "/x" and "/xx"
423 A single "/x" tells the regular expression parser to ignore most
425 whitespace that is neither backslashed nor within a bracketed character
426 class. You can use this to break up your regular expression into more
427 readable parts. Also, the "#" character is treated as a metacharacter
428 introducing a comment that runs up to the pattern's closing delimiter,
429 or to the end of the current line if the pattern extends onto the next
430 line. Hence, this is very much like an ordinary Perl code comment.
431 (You can include the closing delimiter within the comment only if you
432 precede it with a backslash, so be careful!)
433 Use of "/x" means that if you want real whitespace or "#" characters in
435 the pattern (outside a bracketed character class, which is unaffected
436 by "/x"), then you'll either have to escape them (using backslashes or
437 "\Q...\E") or encode them using octal, hex, or "\N{}" or "\p{name=...}"
438 escapes. It is ineffective to try to continue a comment onto the next
439 line by escaping the "\n" with a backslash or "\Q".
440 You can use "(?#text)" to create a comment that ends earlier than the
442 end of the current line, but "text" also can't contain the closing
443 delimiter unless escaped with a backslash.
444 A common pitfall is to forget that "#" characters begin a comment under
446 "/x" and are not matched literally. Just keep that in mind when trying
447 to puzzle out why a particular "/x" pattern isn't working as expected.
448 Starting in Perl v5.26, if the modifier has a second "x" within it, it
450 does everything that a single "/x" does, but additionally non-
451 backslashed SPACE and TAB characters within bracketed character classes
452 are also generally ignored, and hence can be added to make the classes
453 more readable.
454 / [d-e g-i 3-7]/xx
456 /[ ! @ " # $ % ^ & * () = ? <> ' ]/xx
457 may be easier to grasp than the squashed equivalents
459 /[d-eg-i3-7]/
461 /[!@"#$%^&*()=?<>']/
462 Taken together, these features go a long way towards making Perl's
464 regular expressions more readable. Here's an example:
465 # Delete (most) C comments.
467 $program =~ s {
468 /\* # Match the opening delimiter.
469 .*? # Match a minimal number of characters.
470 \*/ # Match the closing delimiter.
471 } []gsx;
472 Note that anything inside a "\Q...\E" stays unaffected by "/x". And
474 note that "/x" doesn't affect space interpretation within a single
475 multi-character construct. For example "(?:...)" can't have a space
476 between the "(", "?", and ":". Within any delimiters for such a
477 construct, allowed spaces are not affected by "/x", and depend on the
478 construct. For example, all constructs using curly braces as
479 delimiters, such as "\x{...}" can have blanks within but adjacent to
480 the braces, but not elsewhere, and no non-blank space characters. An
481 exception are Unicode properties which follow Unicode rules, for which
482 see "Properties accessible through \p{} and \P{}" in perluniprops.
483 The set of characters that are deemed whitespace are those that Unicode
485 calls "Pattern White Space", namely:
486 U+0009 CHARACTER TABULATION
488 U+000A LINE FEED
489 U+000B LINE TABULATION
490 U+000C FORM FEED
491 U+000D CARRIAGE RETURN
492 U+0020 SPACE
493 U+0085 NEXT LINE
494 U+200E LEFT-TO-RIGHT MARK
495 U+200F RIGHT-TO-LEFT MARK
496 U+2028 LINE SEPARATOR
497 U+2029 PARAGRAPH SEPARATOR
498 Character set modifiers
500 "/d", "/u", "/a", and "/l", available starting in 5.14, are called the
502 character set modifiers; they affect the character set rules used for
503 the regular expression.
504 The "/d", "/u", and "/l" modifiers are not likely to be of much use to
506 you, and so you need not worry about them very much. They exist for
507 Perl's internal use, so that complex regular expression data structures
508 can be automatically serialized and later exactly reconstituted,
509 including all their nuances. But, since Perl can't keep a secret, and
510 there may be rare instances where they are useful, they are documented
511 here.
512 The "/a" modifier, on the other hand, may be useful. Its purpose is to
514 allow code that is to work mostly on ASCII data to not have to concern
515 itself with Unicode.
516 Briefly, "/l" sets the character set to that of whatever Locale is in
518 effect at the time of the execution of the pattern match.
519 "/u" sets the character set to Unicode.
521 "/a" also sets the character set to Unicode, BUT adds several
523 restrictions for ASCII-safe matching.
524 "/d" is the old, problematic, pre-5.14 Default character set behavior.
526 Its only use is to force that old behavior.
527 At any given time, exactly one of these modifiers is in effect. Their
529 existence allows Perl to keep the originally compiled behavior of a
530 regular expression, regardless of what rules are in effect when it is
531 actually executed. And if it is interpolated into a larger regex, the
532 original's rules continue to apply to it, and don't affect the other
533 parts.
534 The "/l" and "/u" modifiers are automatically selected for regular
536 expressions compiled within the scope of various pragmas, and we
537 recommend that in general, you use those pragmas instead of specifying
538 these modifiers explicitly. For one thing, the modifiers affect only
539 pattern matching, and do not extend to even any replacement done,
540 whereas using the pragmas gives consistent results for all appropriate
541 operations within their scopes. For example,
542 s/foo/\Ubar/il
544 will match "foo" using the locale's rules for case-insensitive
546 matching, but the "/l" does not affect how the "\U" operates. Most
547 likely you want both of them to use locale rules. To do this, instead
548 compile the regular expression within the scope of "use locale". This
549 both implicitly adds the "/l", and applies locale rules to the "\U".
550 The lesson is to "use locale", and not "/l" explicitly.
551 Similarly, it would be better to use "use feature 'unicode_strings'"
553 instead of,
554 s/foo/\Lbar/iu
556 to get Unicode rules, as the "\L" in the former (but not necessarily
558 the latter) would also use Unicode rules.
559 More detail on each of the modifiers follows. Most likely you don't
561 need to know this detail for "/l", "/u", and "/d", and can skip ahead
562 to /a.
563 /l
565 means to use the current locale's rules (see perllocale) when pattern
567 matching. For example, "\w" will match the "word" characters of that
568 locale, and "/i" case-insensitive matching will match according to the
569 locale's case folding rules. The locale used will be the one in effect
570 at the time of execution of the pattern match. This may not be the
571 same as the compilation-time locale, and can differ from one match to
572 another if there is an intervening call of the setlocale() function.
573 Prior to v5.20, Perl did not support multi-byte locales. Starting
575 then, UTF-8 locales are supported. No other multi byte locales are
576 ever likely to be supported. However, in all locales, one can have
577 code points above 255 and these will always be treated as Unicode no
578 matter what locale is in effect.
579 Under Unicode rules, there are a few case-insensitive matches that
581 cross the 255/256 boundary. Except for UTF-8 locales in Perls v5.20
582 and later, these are disallowed under "/l". For example, 0xFF (on
583 ASCII platforms) does not caselessly match the character at 0x178,
584 "LATIN CAPITAL LETTER Y WITH DIAERESIS", because 0xFF may not be "LATIN
585 SMALL LETTER Y WITH DIAERESIS" in the current locale, and Perl has no
586 way of knowing if that character even exists in the locale, much less
587 what code point it is.
588 In a UTF-8 locale in v5.20 and later, the only visible difference
590 between locale and non-locale in regular expressions should be tainting
591 (see perlsec).
592 This modifier may be specified to be the default by "use locale", but
594 see "Which character set modifier is in effect?".
595 /u
597 means to use Unicode rules when pattern matching. On ASCII platforms,
599 this means that the code points between 128 and 255 take on their
600 Latin-1 (ISO-8859-1) meanings (which are the same as Unicode's).
601 (Otherwise Perl considers their meanings to be undefined.) Thus, under
602 this modifier, the ASCII platform effectively becomes a Unicode
603 platform; and hence, for example, "\w" will match any of the more than
604 100_000 word characters in Unicode.
605 Unlike most locales, which are specific to a language and country pair,
607 Unicode classifies all the characters that are letters somewhere in the
608 world as "\w". For example, your locale might not think that "LATIN
609 SMALL LETTER ETH" is a letter (unless you happen to speak Icelandic),
610 but Unicode does. Similarly, all the characters that are decimal
611 digits somewhere in the world will match "\d"; this is hundreds, not
612 10, possible matches. And some of those digits look like some of the
613 10 ASCII digits, but mean a different number, so a human could easily
614 think a number is a different quantity than it really is. For example,
615 "BENGALI DIGIT FOUR" (U+09EA) looks very much like an "ASCII DIGIT
616 EIGHT" (U+0038), and "LEPCHA DIGIT SIX" (U+1C46) looks very much like
617 an "ASCII DIGIT FIVE" (U+0035). And, "\d+", may match strings of
618 digits that are a mixture from different writing systems, creating a
619 security issue. A fraudulent website, for example, could display the
620 price of something using U+1C46, and it would appear to the user that
621 something cost 500 units, but it really costs 600. A browser that
622 enforced script runs ("Script Runs") would prevent that fraudulent
623 display. "num()" in Unicode::UCD can also be used to sort this out.
624 Or the "/a" modifier can be used to force "\d" to match just the ASCII
625 0 through 9.
626 Also, under this modifier, case-insensitive matching works on the full
628 set of Unicode characters. The "KELVIN SIGN", for example matches the
629 letters "k" and "K"; and "LATIN SMALL LIGATURE FF" matches the sequence
630 "ff", which, if you're not prepared, might make it look like a
631 hexadecimal constant, presenting another potential security issue. See
632 <https://unicode.org/reports/tr36> for a detailed discussion of Unicode
633 security issues.
634 This modifier may be specified to be the default by "use feature
636 'unicode_strings", "use locale ':not_characters'", or "use 5.012" (or
637 higher), but see "Which character set modifier is in effect?".
638 /d
640 This modifier means to use the "Default" native rules of the platform
642 except when there is cause to use Unicode rules instead, as follows:
643 1. the target string is encoded in UTF-8; or
645 2. the pattern is encoded in UTF-8; or
647 3. the pattern explicitly mentions a code point that is above 255 (say
649 by "\x{100}"); or
650 4. the pattern uses a Unicode name ("\N{...}"); or
652 5. the pattern uses a Unicode property ("\p{...}" or "\P{...}"); or
654 6. the pattern uses a Unicode break ("\b{...}" or "\B{...}"); or
656 7. the pattern uses "(?[ ])"
658 8. the pattern uses "(*script_run: ...)"
660 Another mnemonic for this modifier is "Depends", as the rules actually
662 used depend on various things, and as a result you can get unexpected
663 results. See "The "Unicode Bug"" in perlunicode. The Unicode Bug has
664 become rather infamous, leading to yet other (without swearing) names
665 for this modifier, "Dicey" and "Dodgy".
666 Unless the pattern or string are encoded in UTF-8, only ASCII
668 characters can match positively.
669 Here are some examples of how that works on an ASCII platform:
671 $str = "\xDF"; # $str is not in UTF-8 format.
673 $str =~ /^\w/; # No match, as $str isn't in UTF-8 format.
674 $str .= "\x{0e0b}"; # Now $str is in UTF-8 format.
675 $str =~ /^\w/; # Match! $str is now in UTF-8 format.
676 chop $str;
677 $str =~ /^\w/; # Still a match! $str remains in UTF-8 format.
678 This modifier is automatically selected by default when none of the
680 others are, so yet another name for it is "Default".
681 Because of the unexpected behaviors associated with this modifier, you
683 probably should only explicitly use it to maintain weird backward
684 compatibilities.
685 /a (and /aa)
687 This modifier stands for ASCII-restrict (or ASCII-safe). This modifier
689 may be doubled-up to increase its effect.
690 When it appears singly, it causes the sequences "\d", "\s", "\w", and
692 the Posix character classes to match only in the ASCII range. They
693 thus revert to their pre-5.6, pre-Unicode meanings. Under "/a", "\d"
694 always means precisely the digits "0" to "9"; "\s" means the five
695 characters "[ \f\n\r\t]", and starting in Perl v5.18, the vertical tab;
696 "\w" means the 63 characters "[A-Za-z0-9_]"; and likewise, all the
697 Posix classes such as "[[:print:]]" match only the appropriate ASCII-
698 range characters.
699 This modifier is useful for people who only incidentally use Unicode,
701 and who do not wish to be burdened with its complexities and security
702 concerns.
703 With "/a", one can write "\d" with confidence that it will only match
705 ASCII characters, and should the need arise to match beyond ASCII, you
706 can instead use "\p{Digit}" (or "\p{Word}" for "\w"). There are
707 similar "\p{...}" constructs that can match beyond ASCII both white
708 space (see "Whitespace" in perlrecharclass), and Posix classes (see
709 "POSIX Character Classes" in perlrecharclass). Thus, this modifier
710 doesn't mean you can't use Unicode, it means that to get Unicode
711 matching you must explicitly use a construct ("\p{}", "\P{}") that
712 signals Unicode.
713 As you would expect, this modifier causes, for example, "\D" to mean
715 the same thing as "[^0-9]"; in fact, all non-ASCII characters match
716 "\D", "\S", and "\W". "\b" still means to match at the boundary
717 between "\w" and "\W", using the "/a" definitions of them (similarly
718 for "\B").
719 Otherwise, "/a" behaves like the "/u" modifier, in that case-
721 insensitive matching uses Unicode rules; for example, "k" will match
722 the Unicode "\N{KELVIN SIGN}" under "/i" matching, and code points in
723 the Latin1 range, above ASCII will have Unicode rules when it comes to
724 case-insensitive matching.
725 To forbid ASCII/non-ASCII matches (like "k" with "\N{KELVIN SIGN}"),
727 specify the "a" twice, for example "/aai" or "/aia". (The first
728 occurrence of "a" restricts the "\d", etc., and the second occurrence
729 adds the "/i" restrictions.) But, note that code points outside the
730 ASCII range will use Unicode rules for "/i" matching, so the modifier
731 doesn't really restrict things to just ASCII; it just forbids the
732 intermixing of ASCII and non-ASCII.
733 To summarize, this modifier provides protection for applications that
735 don't wish to be exposed to all of Unicode. Specifying it twice gives
736 added protection.
737 This modifier may be specified to be the default by "use re '/a'" or
739 "use re '/aa'". If you do so, you may actually have occasion to use
740 the "/u" modifier explicitly if there are a few regular expressions
741 where you do want full Unicode rules (but even here, it's best if
742 everything were under feature "unicode_strings", along with the "use re
743 '/aa'"). Also see "Which character set modifier is in effect?".
744 Which character set modifier is in effect?
746 Which of these modifiers is in effect at any given point in a regular
748 expression depends on a fairly complex set of interactions. These have
749 been designed so that in general you don't have to worry about it, but
750 this section gives the gory details. As explained below in "Extended
751 Patterns" it is possible to explicitly specify modifiers that apply
752 only to portions of a regular expression. The innermost always has
753 priority over any outer ones, and one applying to the whole expression
754 has priority over any of the default settings that are described in the
755 remainder of this section.
756 The "use re '/foo'" pragma can be used to set default modifiers
758 (including these) for regular expressions compiled within its scope.
759 This pragma has precedence over the other pragmas listed below that
760 also change the defaults.
761 Otherwise, "use locale" sets the default modifier to "/l"; and "use
763 feature 'unicode_strings", or "use 5.012" (or higher) set the default
764 to "/u" when not in the same scope as either "use locale" or "use
765 bytes". ("use locale ':not_characters'" also sets the default to "/u",
766 overriding any plain "use locale".) Unlike the mechanisms mentioned
767 above, these affect operations besides regular expressions pattern
768 matching, and so give more consistent results with other operators,
769 including using "\U", "\l", etc. in substitution replacements.
770 If none of the above apply, for backwards compatibility reasons, the
772 "/d" modifier is the one in effect by default. As this can lead to
773 unexpected results, it is best to specify which other rule set should
774 be used.
775 Character set modifier behavior prior to Perl 5.14
777 Prior to 5.14, there were no explicit modifiers, but "/l" was implied
779 for regexes compiled within the scope of "use locale", and "/d" was
780 implied otherwise. However, interpolating a regex into a larger regex
781 would ignore the original compilation in favor of whatever was in
782 effect at the time of the second compilation. There were a number of
783 inconsistencies (bugs) with the "/d" modifier, where Unicode rules
784 would be used when inappropriate, and vice versa. "\p{}" did not imply
785 Unicode rules, and neither did all occurrences of "\N{}", until 5.12.
786 Regular Expressions
788 Quantifiers
789 Quantifiers are used when a particular portion of a pattern needs to
791 match a certain number (or numbers) of times. If there isn't a
792 quantifier the number of times to match is exactly one. The following
793 standard quantifiers are recognized:
794 * Match 0 or more times
796 + Match 1 or more times
797 ? Match 1 or 0 times
798 {n} Match exactly n times
799 {n,} Match at least n times
800 {,n} Match at most n times
801 {n,m} Match at least n but not more than m times
802 (If a non-escaped curly bracket occurs in a context other than one of
804 the quantifiers listed above, where it does not form part of a
805 backslashed sequence like "\x{...}", it is either a fatal syntax error,
806 or treated as a regular character, generally with a deprecation warning
807 raised. To escape it, you can precede it with a backslash ("\{") or
808 enclose it within square brackets ("[{]"). This change will allow for
809 future syntax extensions (like making the lower bound of a quantifier
810 optional), and better error checking of quantifiers).
811 The "*" quantifier is equivalent to "{0,}", the "+" quantifier to
813 "{1,}", and the "?" quantifier to "{0,1}". n and m are limited to non-
814 negative integral values less than a preset limit defined when perl is
815 built. This is usually 65534 on the most common platforms. The actual
816 limit can be seen in the error message generated by code such as this:
817 $_ **= $_ , / {$_} / for 2 .. 42;
819 By default, a quantified subpattern is "greedy", that is, it will match
821 as many times as possible (given a particular starting location) while
822 still allowing the rest of the pattern to match. If you want it to
823 match the minimum number of times possible, follow the quantifier with
824 a "?". Note that the meanings don't change, just the "greediness":
825 *? Match 0 or more times, not greedily
827 +? Match 1 or more times, not greedily
828 ?? Match 0 or 1 time, not greedily
829 {n}? Match exactly n times, not greedily (redundant)
830 {n,}? Match at least n times, not greedily
831 {,n}? Match at most n times, not greedily
832 {n,m}? Match at least n but not more than m times, not greedily
833 Normally when a quantified subpattern does not allow the rest of the
835 overall pattern to match, Perl will backtrack. However, this behaviour
836 is sometimes undesirable. Thus Perl provides the "possessive"
837 quantifier form as well.
838 *+ Match 0 or more times and give nothing back
840 ++ Match 1 or more times and give nothing back
841 ?+ Match 0 or 1 time and give nothing back
842 {n}+ Match exactly n times and give nothing back (redundant)
843 {n,}+ Match at least n times and give nothing back
844 {,n}+ Match at most n times and give nothing back
845 {n,m}+ Match at least n but not more than m times and give nothing back
846 For instance,
848 'aaaa' =~ /a++a/
850 will never match, as the "a++" will gobble up all the "a"'s in the
852 string and won't leave any for the remaining part of the pattern. This
853 feature can be extremely useful to give perl hints about where it
854 shouldn't backtrack. For instance, the typical "match a double-quoted
855 string" problem can be most efficiently performed when written as:
856 /"(?:[^"\\]++|\\.)*+"/
858 as we know that if the final quote does not match, backtracking will
860 not help. See the independent subexpression "(?>pattern)" for more
861 details; possessive quantifiers are just syntactic sugar for that
862 construct. For instance the above example could also be written as
863 follows:
864 /"(?>(?:(?>[^"\\]+)|\\.)*)"/
866 Note that the possessive quantifier modifier can not be combined with
868 the non-greedy modifier. This is because it would make no sense.
869 Consider the follow equivalency table:
870 Illegal Legal
872 ------------ ------
873 X??+ X{0}
874 X+?+ X{1}
875 X{min,max}?+ X{min}
876 Escape sequences
878 Because patterns are processed as double-quoted strings, the following
880 also work:
881 \t tab (HT, TAB)
883 \n newline (LF, NL)
884 \r return (CR)
885 \f form feed (FF)
886 \a alarm (bell) (BEL)
887 \e escape (think troff) (ESC)
888 \cK control char (example: VT)
889 \x{}, \x00 character whose ordinal is the given hexadecimal number
890 \N{name} named Unicode character or character sequence
891 \N{U+263D} Unicode character (example: FIRST QUARTER MOON)
892 \o{}, \000 character whose ordinal is the given octal number
893 \l lowercase next char (think vi)
894 \u uppercase next char (think vi)
895 \L lowercase until \E (think vi)
896 \U uppercase until \E (think vi)
897 \Q quote (disable) pattern metacharacters until \E
898 \E end either case modification or quoted section, think vi
899 Details are in "Quote and Quote-like Operators" in perlop.
901 Character Classes and other Special Escapes
903 In addition, Perl defines the following:
905 Sequence Note Description
907 [...] [1] Match a character according to the rules of the
908 bracketed character class defined by the "...".
909 Example: [a-z] matches "a" or "b" or "c" ... or "z"
910 [[:...:]] [2] Match a character according to the rules of the POSIX
911 character class "..." within the outer bracketed
912 character class. Example: [[:upper:]] matches any
913 uppercase character.
914 (?[...]) [8] Extended bracketed character class
915 \w [3] Match a "word" character (alphanumeric plus "_", plus
916 other connector punctuation chars plus Unicode
917 marks)
918 \W [3] Match a non-"word" character
919 \s [3] Match a whitespace character
920 \S [3] Match a non-whitespace character
921 \d [3] Match a decimal digit character
922 \D [3] Match a non-digit character
923 \pP [3] Match P, named property. Use \p{Prop} for longer names
924 \PP [3] Match non-P
925 \X [4] Match Unicode "eXtended grapheme cluster"
926 \1 [5] Backreference to a specific capture group or buffer.
927 '1' may actually be any positive integer.
928 \g1 [5] Backreference to a specific or previous group,
929 \g{-1} [5] The number may be negative indicating a relative
930 previous group and may optionally be wrapped in
931 curly brackets for safer parsing.
932 \g{name} [5] Named backreference
933 \k<name> [5] Named backreference
934 \k'name' [5] Named backreference
935 \k{name} [5] Named backreference
936 \K [6] Keep the stuff left of the \K, don't include it in $&
937 \N [7] Any character but \n. Not affected by /s modifier
938 \v [3] Vertical whitespace
939 \V [3] Not vertical whitespace
940 \h [3] Horizontal whitespace
941 \H [3] Not horizontal whitespace
942 \R [4] Linebreak
943 [1] See "Bracketed Character Classes" in perlrecharclass for details.
945 [2] See "POSIX Character Classes" in perlrecharclass for details.
947 [3] See "Unicode Character Properties" in perlunicode for details
949 [4] See "Misc" in perlrebackslash for details.
951 [5] See "Capture groups" below for details.
953 [6] See "Extended Patterns" below for details.
955 [7] Note that "\N" has two meanings. When of the form "\N{NAME}", it
957 matches the character or character sequence whose name is NAME; and
958 similarly when of the form "\N{U+hex}", it matches the character
959 whose Unicode code point is hex. Otherwise it matches any
960 character but "\n".
961 [8] See "Extended Bracketed Character Classes" in perlrecharclass for
963 details.
964 Assertions
966 Besides "^" and "$", Perl defines the following zero-width assertions:
968 \b{} Match at Unicode boundary of specified type
970 \B{} Match where corresponding \b{} doesn't match
971 \b Match a \w\W or \W\w boundary
972 \B Match except at a \w\W or \W\w boundary
973 \A Match only at beginning of string
974 \Z Match only at end of string, or before newline at the end
975 \z Match only at end of string
976 \G Match only at pos() (e.g. at the end-of-match position
977 of prior m//g)
978 A Unicode boundary ("\b{}"), available starting in v5.22, is a spot
980 between two characters, or before the first character in the string, or
981 after the final character in the string where certain criteria defined
982 by Unicode are met. See "\b{}, \b, \B{}, \B" in perlrebackslash for
983 details.
984 A word boundary ("\b") is a spot between two characters that has a "\w"
986 on one side of it and a "\W" on the other side of it (in either order),
987 counting the imaginary characters off the beginning and end of the
988 string as matching a "\W". (Within character classes "\b" represents
989 backspace rather than a word boundary, just as it normally does in any
990 double-quoted string.) The "\A" and "\Z" are just like "^" and "$",
991 except that they won't match multiple times when the "/m" modifier is
992 used, while "^" and "$" will match at every internal line boundary. To
993 match the actual end of the string and not ignore an optional trailing
994 newline, use "\z".
995 The "\G" assertion can be used to chain global matches (using "m//g"),
997 as described in "Regexp Quote-Like Operators" in perlop. It is also
998 useful when writing "lex"-like scanners, when you have several patterns
999 that you want to match against consequent substrings of your string;
1000 see the previous reference. The actual location where "\G" will match
1001 can also be influenced by using "pos()" as an lvalue: see "pos" in
1002 perlfunc. Note that the rule for zero-length matches (see "Repeated
1003 Patterns Matching a Zero-length Substring") is modified somewhat, in
1004 that contents to the left of "\G" are not counted when determining the
1005 length of the match. Thus the following will not match forever:
1006 my $string = 'ABC';
1008 pos($string) = 1;
1009 while ($string =~ /(.\G)/g) {
1010 print $1;
1011 }
1012 It will print 'A' and then terminate, as it considers the match to be
1014 zero-width, and thus will not match at the same position twice in a
1015 row.
1016 It is worth noting that "\G" improperly used can result in an infinite
1018 loop. Take care when using patterns that include "\G" in an
1019 alternation.
1020 Note also that "s///" will refuse to overwrite part of a substitution
1022 that has already been replaced; so for example this will stop after the
1023 first iteration, rather than iterating its way backwards through the
1024 string:
1025 $_ = "123456789";
1027 pos = 6;
1028 s/.(?=.\G)/X/g;
1029 print; # prints 1234X6789, not XXXXX6789
1030 Capture groups
1032 The grouping construct "( ... )" creates capture groups (also referred
1034 to as capture buffers). To refer to the current contents of a group
1035 later on, within the same pattern, use "\g1" (or "\g{1}") for the
1036 first, "\g2" (or "\g{2}") for the second, and so on. This is called a
1037 backreference.
1038 There is no limit to the number of captured substrings that you may
1047 use. Groups are numbered with the leftmost open parenthesis being
1048 number 1, etc. If a group did not match, the associated backreference
1049 won't match either. (This can happen if the group is optional, or in a
1050 different branch of an alternation.) You can omit the "g", and write
1051 "\1", etc, but there are some issues with this form, described below.
1052 You can also refer to capture groups relatively, by using a negative
1054 number, so that "\g-1" and "\g{-1}" both refer to the immediately
1055 preceding capture group, and "\g-2" and "\g{-2}" both refer to the
1056 group before it. For example:
1057 /
1059 (Y) # group 1
1060 ( # group 2
1061 (X) # group 3
1062 \g{-1} # backref to group 3
1063 \g{-3} # backref to group 1
1064 )
1065 /x
1066 would match the same as "/(Y) ( (X) \g3 \g1 )/x". This allows you to
1068 interpolate regexes into larger regexes and not have to worry about the
1069 capture groups being renumbered.
1070 You can dispense with numbers altogether and create named capture
1072 groups. The notation is "(?<name>...)" to declare and "\g{name}" to
1073 reference. (To be compatible with .Net regular expressions, "\g{name}"
1074 may also be written as "\k{name}", "\k<name>" or "\k'name'".) name
1075 must not begin with a number, nor contain hyphens. When different
1076 groups within the same pattern have the same name, any reference to
1077 that name assumes the leftmost defined group. Named groups count in
1078 absolute and relative numbering, and so can also be referred to by
1079 those numbers. (It's possible to do things with named capture groups
1080 that would otherwise require "(??{})".)
1081 Capture group contents are dynamically scoped and available to you
1083 outside the pattern until the end of the enclosing block or until the
1084 next successful match, whichever comes first. (See "Compound
1085 Statements" in perlsyn.) You can refer to them by absolute number
1086 (using "$1" instead of "\g1", etc); or by name via the "%+" hash, using
1087 "$+{name}".
1088 Braces are required in referring to named capture groups, but are
1090 optional for absolute or relative numbered ones. Braces are safer when
1091 creating a regex by concatenating smaller strings. For example if you
1092 have "qr/$a$b/", and $a contained "\g1", and $b contained "37", you
1093 would get "/\g137/" which is probably not what you intended.
1094 If you use braces, you may also optionally add any number of blank
1096 (space or tab) characters within but adjacent to the braces, like
1097 "\g{ -1 }", or "\k{ name }".
1098 The "\g" and "\k" notations were introduced in Perl 5.10.0. Prior to
1100 that there were no named nor relative numbered capture groups.
1101 Absolute numbered groups were referred to using "\1", "\2", etc., and
1102 this notation is still accepted (and likely always will be). But it
1103 leads to some ambiguities if there are more than 9 capture groups, as
1104 "\10" could mean either the tenth capture group, or the character whose
1105 ordinal in octal is 010 (a backspace in ASCII). Perl resolves this
1106 ambiguity by interpreting "\10" as a backreference only if at least 10
1107 left parentheses have opened before it. Likewise "\11" is a
1108 backreference only if at least 11 left parentheses have opened before
1109 it. And so on. "\1" through "\9" are always interpreted as
1110 backreferences. There are several examples below that illustrate these
1111 perils. You can avoid the ambiguity by always using "\g{}" or "\g" if
1112 you mean capturing groups; and for octal constants always using "\o{}",
1113 or for "\077" and below, using 3 digits padded with leading zeros,
1114 since a leading zero implies an octal constant.
1115 The "\digit" notation also works in certain circumstances outside the
1117 pattern. See "Warning on \1 Instead of $1" below for details.
1118 Examples:
1120 s/^([^ ]*) *([^ ]*)/$2 $1/; # swap first two words
1122 /(.)\g1/ # find first doubled char
1124 and print "'$1' is the first doubled character\n";
1125 /(?<char>.)\k<char>/ # ... a different way
1127 and print "'$+{char}' is the first doubled character\n";
1128 /(?'char'.)\g1/ # ... mix and match
1130 and print "'$1' is the first doubled character\n";
1131 if (/Time: (..):(..):(..)/) { # parse out values
1133 $hours = $1;
1134 $minutes = $2;
1135 $seconds = $3;
1136 }
1137 /(.)(.)(.)(.)(.)(.)(.)(.)(.)\g10/ # \g10 is a backreference
1139 /(.)(.)(.)(.)(.)(.)(.)(.)(.)\10/ # \10 is octal
1140 /((.)(.)(.)(.)(.)(.)(.)(.)(.))\10/ # \10 is a backreference
1141 /((.)(.)(.)(.)(.)(.)(.)(.)(.))\010/ # \010 is octal
1142 $a = '(.)\1'; # Creates problems when concatenated.
1144 $b = '(.)\g{1}'; # Avoids the problems.
1145 "aa" =~ /${a}/; # True
1146 "aa" =~ /${b}/; # True
1147 "aa0" =~ /${a}0/; # False!
1148 "aa0" =~ /${b}0/; # True
1149 "aa\x08" =~ /${a}0/; # True!
1150 "aa\x08" =~ /${b}0/; # False
1151 Several special variables also refer back to portions of the previous
1153 match. $+ returns whatever the last bracket match matched. $& returns
1154 the entire matched string. (At one point $0 did also, but now it
1155 returns the name of the program.) "$`" returns everything before the
1156 matched string. "$'" returns everything after the matched string. And
1157 $^N contains whatever was matched by the most-recently closed group
1158 (submatch). $^N can be used in extended patterns (see below), for
1159 example to assign a submatch to a variable.
1160 These special variables, like the "%+" hash and the numbered match
1162 variables ($1, $2, $3, etc.) are dynamically scoped until the end of
1163 the enclosing block or until the next successful match, whichever comes
1164 first. (See "Compound Statements" in perlsyn.)
1165 NOTE: Failed matches in Perl do not reset the match variables, which
1167 makes it easier to write code that tests for a series of more specific
1168 cases and remembers the best match.
1169 WARNING: If your code is to run on Perl 5.16 or earlier, beware that
1171 once Perl sees that you need one of $&, "$`", or "$'" anywhere in the
1172 program, it has to provide them for every pattern match. This may
1173 substantially slow your program.
1174 Perl uses the same mechanism to produce $1, $2, etc, so you also pay a
1176 price for each pattern that contains capturing parentheses. (To avoid
1177 this cost while retaining the grouping behaviour, use the extended
1178 regular expression "(?: ... )" instead.) But if you never use $&, "$`"
1179 or "$'", then patterns without capturing parentheses will not be
1180 penalized. So avoid $&, "$'", and "$`" if you can, but if you can't
1181 (and some algorithms really appreciate them), once you've used them
1182 once, use them at will, because you've already paid the price.
1183 Perl 5.16 introduced a slightly more efficient mechanism that notes
1185 separately whether each of "$`", $&, and "$'" have been seen, and thus
1186 may only need to copy part of the string. Perl 5.20 introduced a much
1187 more efficient copy-on-write mechanism which eliminates any slowdown.
1188 As another workaround for this problem, Perl 5.10.0 introduced
1190 "${^PREMATCH}", "${^MATCH}" and "${^POSTMATCH}", which are equivalent
1191 to "$`", $& and "$'", except that they are only guaranteed to be
1192 defined after a successful match that was executed with the "/p"
1193 (preserve) modifier. The use of these variables incurs no global
1194 performance penalty, unlike their punctuation character equivalents,
1195 however at the trade-off that you have to tell perl when you want to
1196 use them. As of Perl 5.20, these three variables are equivalent to
1197 "$`", $& and "$'", and "/p" is ignored.
1198 Quoting metacharacters
1200 Backslashed metacharacters in Perl are alphanumeric, such as "\b",
1201 "\w", "\n". Unlike some other regular expression languages, there are
1202 no backslashed symbols that aren't alphanumeric. So anything that
1203 looks like "\\", "\(", "\)", "\[", "\]", "\{", or "\}" is always
1204 interpreted as a literal character, not a metacharacter. This was once
1205 used in a common idiom to disable or quote the special meanings of
1206 regular expression metacharacters in a string that you want to use for
1207 a pattern. Simply quote all non-"word" characters:
1208 $pattern =~ s/(\W)/\\$1/g;
1210 (If "use locale" is set, then this depends on the current locale.)
1212 Today it is more common to use the "quotemeta()" function or the "\Q"
1213 metaquoting escape sequence to disable all metacharacters' special
1214 meanings like this:
1215 /$unquoted\Q$quoted\E$unquoted/
1217 Beware that if you put literal backslashes (those not inside
1219 interpolated variables) between "\Q" and "\E", double-quotish backslash
1220 interpolation may lead to confusing results. If you need to use
1221 literal backslashes within "\Q...\E", consult "Gory details of parsing
1222 quoted constructs" in perlop.
1223 "quotemeta()" and "\Q" are fully described in "quotemeta" in perlfunc.
1225 Extended Patterns
1227 Perl also defines a consistent extension syntax for features not found
1228 in standard tools like awk and lex. The syntax for most of these is a
1229 pair of parentheses with a question mark as the first thing within the
1230 parentheses. The character after the question mark indicates the
1231 extension.
1232 A question mark was chosen for this and for the minimal-matching
1234 construct because 1) question marks are rare in older regular
1235 expressions, and 2) whenever you see one, you should stop and
1236 "question" exactly what is going on. That's psychology....
1237 "(?#text)"
1239 A comment. The text is ignored. Note that Perl closes the comment
1240 as soon as it sees a ")", so there is no way to put a literal ")"
1241 in the comment. The pattern's closing delimiter must be escaped by
1242 a backslash if it appears in the comment.
1243 See "/x" for another way to have comments in patterns.
1245 Note that a comment can go just about anywhere, except in the
1247 middle of an escape sequence. Examples:
1248 qr/foo(?#comment)bar/' # Matches 'foobar'
1250 # The pattern below matches 'abcd', 'abccd', or 'abcccd'
1252 qr/abc(?#comment between literal and its quantifier){1,3}d/
1253 # The pattern below generates a syntax error, because the '\p' must
1255 # be followed immediately by a '{'.
1256 qr/\p(?#comment between \p and its property name){Any}/
1257 # The pattern below generates a syntax error, because the initial
1259 # '\(' is a literal opening parenthesis, and so there is nothing
1260 # for the closing ')' to match
1261 qr/\(?#the backslash means this isn't a comment)p{Any}/
1262 # Comments can be used to fold long patterns into multiple lines
1264 qr/First part of a long regex(?#
1265 )remaining part/
1266 "(?adlupimnsx-imnsx)"
1268 "(?^alupimnsx)"
1269 Zero or more embedded pattern-match modifiers, to be turned on (or
1270 turned off if preceded by "-") for the remainder of the pattern or
1271 the remainder of the enclosing pattern group (if any).
1272 This is particularly useful for dynamically-generated patterns,
1274 such as those read in from a configuration file, taken from an
1275 argument, or specified in a table somewhere. Consider the case
1276 where some patterns want to be case-sensitive and some do not: The
1277 case-insensitive ones merely need to include "(?i)" at the front of
1278 the pattern. For example:
1279 $pattern = "foobar";
1281 if ( /$pattern/i ) { }
1282 # more flexible:
1284 $pattern = "(?i)foobar";
1286 if ( /$pattern/ ) { }
1287 These modifiers are restored at the end of the enclosing group. For
1289 example,
1290 ( (?i) blah ) \s+ \g1
1292 will match "blah" in any case, some spaces, and an exact (including
1294 the case!) repetition of the previous word, assuming the "/x"
1295 modifier, and no "/i" modifier outside this group.
1296 These modifiers do not carry over into named subpatterns called in
1298 the enclosing group. In other words, a pattern such as
1299 "((?i)(?&NAME))" does not change the case-sensitivity of the NAME
1300 pattern.
1301 A modifier is overridden by later occurrences of this construct in
1303 the same scope containing the same modifier, so that
1304 /((?im)foo(?-m)bar)/
1306 matches all of "foobar" case insensitively, but uses "/m" rules for
1308 only the "foo" portion. The "a" flag overrides "aa" as well;
1309 likewise "aa" overrides "a". The same goes for "x" and "xx".
1310 Hence, in
1311 /(?-x)foo/xx
1313 both "/x" and "/xx" are turned off during matching "foo". And in
1315 /(?x)foo/x
1317 "/x" but NOT "/xx" is turned on for matching "foo". (One might
1319 mistakenly think that since the inner "(?x)" is already in the
1320 scope of "/x", that the result would effectively be the sum of
1321 them, yielding "/xx". It doesn't work that way.) Similarly, doing
1322 something like "(?xx-x)foo" turns off all "x" behavior for matching
1323 "foo", it is not that you subtract 1 "x" from 2 to get 1 "x"
1324 remaining.
1325 Any of these modifiers can be set to apply globally to all regular
1327 expressions compiled within the scope of a "use re". See "'/flags'
1328 mode" in re.
1329 Starting in Perl 5.14, a "^" (caret or circumflex accent)
1331 immediately after the "?" is a shorthand equivalent to "d-imnsx".
1332 Flags (except "d") may follow the caret to override it. But a
1333 minus sign is not legal with it.
1334 Note that the "a", "d", "l", "p", and "u" modifiers are special in
1336 that they can only be enabled, not disabled, and the "a", "d", "l",
1337 and "u" modifiers are mutually exclusive: specifying one de-
1338 specifies the others, and a maximum of one (or two "a"'s) may
1339 appear in the construct. Thus, for example, "(?-p)" will warn when
1340 compiled under "use warnings"; "(?-d:...)" and "(?dl:...)" are
1341 fatal errors.
1342 Note also that the "p" modifier is special in that its presence
1344 anywhere in a pattern has a global effect.
1345 Having zero modifiers makes this a no-op (so why did you specify
1347 it, unless it's generated code), and starting in v5.30, warns under
1348 "use re 'strict'".
1349 "(?:pattern)"
1351 "(?adluimnsx-imnsx:pattern)"
1352 "(?^aluimnsx:pattern)"
1353 This is for clustering, not capturing; it groups subexpressions
1354 like "()", but doesn't make backreferences as "()" does. So
1355 @fields = split(/\b(?:a|b|c)\b/)
1357 matches the same field delimiters as
1359 @fields = split(/\b(a|b|c)\b/)
1361 but doesn't spit out the delimiters themselves as extra fields
1363 (even though that's the behaviour of "split" in perlfunc when its
1364 pattern contains capturing groups). It's also cheaper not to
1365 capture characters if you don't need to.
1366 Any letters between "?" and ":" act as flags modifiers as with
1368 "(?adluimnsx-imnsx)". For example,
1369 /(?s-i:more.*than).*million/i
1371 is equivalent to the more verbose
1373 /(?:(?s-i)more.*than).*million/i
1375 Note that any "()" constructs enclosed within this one will still
1377 capture unless the "/n" modifier is in effect.
1378 Like the "(?adlupimnsx-imnsx)" construct, "aa" and "a" override
1380 each other, as do "xx" and "x". They are not additive. So, doing
1381 something like "(?xx-x:foo)" turns off all "x" behavior for
1382 matching "foo".
1383 Starting in Perl 5.14, a "^" (caret or circumflex accent)
1385 immediately after the "?" is a shorthand equivalent to "d-imnsx".
1386 Any positive flags (except "d") may follow the caret, so
1387 (?^x:foo)
1389 is equivalent to
1391 (?x-imns:foo)
1393 The caret tells Perl that this cluster doesn't inherit the flags of
1395 any surrounding pattern, but uses the system defaults ("d-imnsx"),
1396 modified by any flags specified.
1397 The caret allows for simpler stringification of compiled regular
1399 expressions. These look like
1400 (?^:pattern)
1402 with any non-default flags appearing between the caret and the
1404 colon. A test that looks at such stringification thus doesn't need
1405 to have the system default flags hard-coded in it, just the caret.
1406 If new flags are added to Perl, the meaning of the caret's
1407 expansion will change to include the default for those flags, so
1408 the test will still work, unchanged.
1409 Specifying a negative flag after the caret is an error, as the flag
1411 is redundant.
1412 Mnemonic for "(?^...)": A fresh beginning since the usual use of a
1414 caret is to match at the beginning.
1415 "(?|pattern)"
1417 This is the "branch reset" pattern, which has the special property
1418 that the capture groups are numbered from the same starting point
1419 in each alternation branch. It is available starting from perl
1420 5.10.0.
1421 Capture groups are numbered from left to right, but inside this
1423 construct the numbering is restarted for each branch.
1424 The numbering within each branch will be as normal, and any groups
1426 following this construct will be numbered as though the construct
1427 contained only one branch, that being the one with the most capture
1428 groups in it.
1429 This construct is useful when you want to capture one of a number
1431 of alternative matches.
1432 Consider the following pattern. The numbers underneath show in
1434 which group the captured content will be stored.
1435 # before ---------------branch-reset----------- after
1437 / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
1438 # 1 2 2 3 2 3 4
1439 Be careful when using the branch reset pattern in combination with
1441 named captures. Named captures are implemented as being aliases to
1442 numbered groups holding the captures, and that interferes with the
1443 implementation of the branch reset pattern. If you are using named
1444 captures in a branch reset pattern, it's best to use the same
1445 names, in the same order, in each of the alternations:
1446 /(?| (?<a> x ) (?<b> y )
1448 | (?<a> z ) (?<b> w )) /x
1449 Not doing so may lead to surprises:
1451 "12" =~ /(?| (?<a> \d+ ) | (?<b> \D+))/x;
1453 say $+{a}; # Prints '12'
1454 say $+{b}; # *Also* prints '12'.
1455 The problem here is that both the group named "a" and the group
1457 named "b" are aliases for the group belonging to $1.
1458 Lookaround Assertions
1460 Lookaround assertions are zero-width patterns which match a
1461 specific pattern without including it in $&. Positive assertions
1462 match when their subpattern matches, negative assertions match when
1463 their subpattern fails. Lookbehind matches text up to the current
1464 match position, lookahead matches text following the current match
1465 position.
1466 "(?=pattern)"
1468 "(*pla:pattern)"
1469 "(*positive_lookahead:pattern)"
1470 A zero-width positive lookahead assertion. For example,
1471 "/\w+(?=\t)/" matches a word followed by a tab, without
1472 including the tab in $&.
1473 "(?!pattern)"
1475 "(*nla:pattern)"
1476 "(*negative_lookahead:pattern)"
1477 A zero-width negative lookahead assertion. For example
1478 "/foo(?!bar)/" matches any occurrence of "foo" that isn't
1479 followed by "bar". Note however that lookahead and lookbehind
1480 are NOT the same thing. You cannot use this for lookbehind.
1481 If you are looking for a "bar" that isn't preceded by a "foo",
1483 "/(?!foo)bar/" will not do what you want. That's because the
1484 "(?!foo)" is just saying that the next thing cannot be
1485 "foo"--and it's not, it's a "bar", so "foobar" will match. Use
1486 lookbehind instead (see below).
1487 "(?<=pattern)"
1489 "\K"
1490 "(*plb:pattern)"
1491 "(*positive_lookbehind:pattern)"
1492 A zero-width positive lookbehind assertion. For example,
1493 "/(?<=\t)\w+/" matches a word that follows a tab, without
1494 including the tab in $&.
1495 Prior to Perl 5.30, it worked only for fixed-width lookbehind,
1497 but starting in that release, it can handle variable lengths
1498 from 1 to 255 characters as an experimental feature. The
1499 feature is enabled automatically if you use a variable length
1500 lookbehind assertion, but will raise a warning at pattern
1501 compilation time, unless turned off, in the "experimental::vlb"
1502 category. This is to warn you that the exact behavior is
1503 subject to change should feedback from actual use in the field
1504 indicate to do so; or even complete removal if the problems
1505 found are not practically surmountable. You can achieve close
1506 to pre-5.30 behavior by fatalizing warnings in this category.
1507 There is a special form of this construct, called "\K"
1509 (available since Perl 5.10.0), which causes the regex engine to
1510 "keep" everything it had matched prior to the "\K" and not
1511 include it in $&. This effectively provides non-experimental
1512 variable-length lookbehind of any length.
1513 And, there is a technique that can be used to handle variable
1515 length lookbehinds on earlier releases, and longer than 255
1516 characters. It is described in
1517 <http://www.drregex.com/2019/02/variable-length-lookbehinds-actually.html>.
1518 Note that under "/i", a few single characters match two or
1520 three other characters. This makes them variable length, and
1521 the 255 length applies to the maximum number of characters in
1522 the match. For example "qr/\N{LATIN SMALL LETTER SHARP S}/i"
1523 matches the sequence "ss". Your lookbehind assertion could
1524 contain 127 Sharp S characters under "/i", but adding a 128th
1525 would generate a compilation error, as that could match 256 "s"
1526 characters in a row.
1527 The use of "\K" inside of another lookaround assertion is
1529 allowed, but the behaviour is currently not well defined.
1530 For various reasons "\K" may be significantly more efficient
1532 than the equivalent "(?<=...)" construct, and it is especially
1533 useful in situations where you want to efficiently remove
1534 something following something else in a string. For instance
1535 s/(foo)bar/$1/g;
1537 can be rewritten as the much more efficient
1539 s/foo\Kbar//g;
1541 Use of the non-greedy modifier "?" may not give you the
1543 expected results if it is within a capturing group within the
1544 construct.
1545 "(?<!pattern)"
1547 "(*nlb:pattern)"
1548 "(*negative_lookbehind:pattern)"
1549 A zero-width negative lookbehind assertion. For example
1550 "/(?<!bar)foo/" matches any occurrence of "foo" that does not
1551 follow "bar".
1552 Prior to Perl 5.30, it worked only for fixed-width lookbehind,
1554 but starting in that release, it can handle variable lengths
1555 from 1 to 255 characters as an experimental feature. The
1556 feature is enabled automatically if you use a variable length
1557 lookbehind assertion, but will raise a warning at pattern
1558 compilation time, unless turned off, in the "experimental::vlb"
1559 category. This is to warn you that the exact behavior is
1560 subject to change should feedback from actual use in the field
1561 indicate to do so; or even complete removal if the problems
1562 found are not practically surmountable. You can achieve close
1563 to pre-5.30 behavior by fatalizing warnings in this category.
1564 There is a technique that can be used to handle variable length
1566 lookbehinds on earlier releases, and longer than 255
1567 characters. It is described in
1568 <http://www.drregex.com/2019/02/variable-length-lookbehinds-actually.html>.
1569 Note that under "/i", a few single characters match two or
1571 three other characters. This makes them variable length, and
1572 the 255 length applies to the maximum number of characters in
1573 the match. For example "qr/\N{LATIN SMALL LETTER SHARP S}/i"
1574 matches the sequence "ss". Your lookbehind assertion could
1575 contain 127 Sharp S characters under "/i", but adding a 128th
1576 would generate a compilation error, as that could match 256 "s"
1577 characters in a row.
1578 Use of the non-greedy modifier "?" may not give you the
1580 expected results if it is within a capturing group within the
1581 construct.
1582 "(?<NAME>pattern)"
1584 "(?'NAME'pattern)"
1585 A named capture group. Identical in every respect to normal
1586 capturing parentheses "()" but for the additional fact that the
1587 group can be referred to by name in various regular expression
1588 constructs (like "\g{NAME}") and can be accessed by name after a
1589 successful match via "%+" or "%-". See perlvar for more details on
1590 the "%+" and "%-" hashes.
1591 If multiple distinct capture groups have the same name, then
1593 $+{NAME} will refer to the leftmost defined group in the match.
1594 The forms "(?'NAME'pattern)" and "(?<NAME>pattern)" are equivalent.
1596 NOTE: While the notation of this construct is the same as the
1598 similar function in .NET regexes, the behavior is not. In Perl the
1599 groups are numbered sequentially regardless of being named or not.
1600 Thus in the pattern
1601 /(x)(?<foo>y)(z)/
1603 $+{foo} will be the same as $2, and $3 will contain 'z' instead of
1605 the opposite which is what a .NET regex hacker might expect.
1606 Currently NAME is restricted to simple identifiers only. In other
1608 words, it must match "/^[_A-Za-z][_A-Za-z0-9]*\z/" or its Unicode
1609 extension (see utf8), though it isn't extended by the locale (see
1610 perllocale).
1611 NOTE: In order to make things easier for programmers with
1613 experience with the Python or PCRE regex engines, the pattern
1614 "(?P<NAME>pattern)" may be used instead of "(?<NAME>pattern)";
1615 however this form does not support the use of single quotes as a
1616 delimiter for the name.
1617 "\k<NAME>"
1619 "\k'NAME'"
1620 "\k{NAME}"
1621 Named backreference. Similar to numeric backreferences, except that
1622 the group is designated by name and not number. If multiple groups
1623 have the same name then it refers to the leftmost defined group in
1624 the current match.
1625 It is an error to refer to a name not defined by a "(?<NAME>)"
1627 earlier in the pattern.
1628 All three forms are equivalent, although with "\k{ NAME }", you may
1630 optionally have blanks within but adjacent to the braces, as shown.
1631 NOTE: In order to make things easier for programmers with
1633 experience with the Python or PCRE regex engines, the pattern
1634 "(?P=NAME)" may be used instead of "\k<NAME>".
1635 "(?{ code })"
1637 WARNING: Using this feature safely requires that you understand its
1638 limitations. Code executed that has side effects may not perform
1639 identically from version to version due to the effect of future
1640 optimisations in the regex engine. For more information on this,
1641 see "Embedded Code Execution Frequency".
1642 This zero-width assertion executes any embedded Perl code. It
1644 always succeeds, and its return value is set as $^R.
1645 In literal patterns, the code is parsed at the same time as the
1647 surrounding code. While within the pattern, control is passed
1648 temporarily back to the perl parser, until the logically-balancing
1649 closing brace is encountered. This is similar to the way that an
1650 array index expression in a literal string is handled, for example
1651 "abc$array[ 1 + f('[') + g()]def"
1653 In particular, braces do not need to be balanced:
1655 s/abc(?{ f('{'); })/def/
1657 Even in a pattern that is interpolated and compiled at run-time,
1659 literal code blocks will be compiled once, at perl compile time;
1660 the following prints "ABCD":
1661 print "D";
1663 my $qr = qr/(?{ BEGIN { print "A" } })/;
1664 my $foo = "foo";
1665 /$foo$qr(?{ BEGIN { print "B" } })/;
1666 BEGIN { print "C" }
1667 In patterns where the text of the code is derived from run-time
1669 information rather than appearing literally in a source code
1670 /pattern/, the code is compiled at the same time that the pattern
1671 is compiled, and for reasons of security, "use re 'eval'" must be
1672 in scope. This is to stop user-supplied patterns containing code
1673 snippets from being executable.
1674 In situations where you need to enable this with "use re 'eval'",
1676 you should also have taint checking enabled. Better yet, use the
1677 carefully constrained evaluation within a Safe compartment. See
1678 perlsec for details about both these mechanisms.
1679 From the viewpoint of parsing, lexical variable scope and closures,
1681 /AAA(?{ BBB })CCC/
1683 behaves approximately like
1685 /AAA/ && do { BBB } && /CCC/
1687 Similarly,
1689 qr/AAA(?{ BBB })CCC/
1691 behaves approximately like
1693 sub { /AAA/ && do { BBB } && /CCC/ }
1695 In particular:
1697 { my $i = 1; $r = qr/(?{ print $i })/ }
1699 my $i = 2;
1700 /$r/; # prints "1"
1701 Inside a "(?{...})" block, $_ refers to the string the regular
1703 expression is matching against. You can also use "pos()" to know
1704 what is the current position of matching within this string.
1705 The code block introduces a new scope from the perspective of
1707 lexical variable declarations, but not from the perspective of
1708 "local" and similar localizing behaviours. So later code blocks
1709 within the same pattern will still see the values which were
1710 localized in earlier blocks. These accumulated localizations are
1711 undone either at the end of a successful match, or if the assertion
1712 is backtracked (compare "Backtracking"). For example,
1713 $_ = 'a' x 8;
1715 m<
1716 (?{ $cnt = 0 }) # Initialize $cnt.
1717 (
1718 a
1719 (?{
1720 local $cnt = $cnt + 1; # Update $cnt,
1721 # backtracking-safe.
1722 })
1723 )*
1724 aaaa
1725 (?{ $res = $cnt }) # On success copy to
1726 # non-localized location.
1727 >x;
1728 will initially increment $cnt up to 8; then during backtracking,
1730 its value will be unwound back to 4, which is the value assigned to
1731 $res. At the end of the regex execution, $cnt will be wound back
1732 to its initial value of 0.
1733 This assertion may be used as the condition in a
1735 (?(condition)yes-pattern|no-pattern)
1737 switch. If not used in this way, the result of evaluation of code
1739 is put into the special variable $^R. This happens immediately, so
1740 $^R can be used from other "(?{ code })" assertions inside the same
1741 regular expression.
1742 The assignment to $^R above is properly localized, so the old value
1744 of $^R is restored if the assertion is backtracked; compare
1745 "Backtracking".
1746 Note that the special variable $^N is particularly useful with
1748 code blocks to capture the results of submatches in variables
1749 without having to keep track of the number of nested parentheses.
1750 For example:
1751 $_ = "The brown fox jumps over the lazy dog";
1753 /the (\S+)(?{ $color = $^N }) (\S+)(?{ $animal = $^N })/i;
1754 print "color = $color, animal = $animal\n";
1755 "(??{ code })"
1757 WARNING: Using this feature safely requires that you understand its
1758 limitations. Code executed that has side effects may not perform
1759 identically from version to version due to the effect of future
1760 optimisations in the regex engine. For more information on this,
1761 see "Embedded Code Execution Frequency".
1762 This is a "postponed" regular subexpression. It behaves in exactly
1764 the same way as a "(?{ code })" code block as described above,
1765 except that its return value, rather than being assigned to $^R, is
1766 treated as a pattern, compiled if it's a string (or used as-is if
1767 its a qr// object), then matched as if it were inserted instead of
1768 this construct.
1769 During the matching of this sub-pattern, it has its own set of
1771 captures which are valid during the sub-match, but are discarded
1772 once control returns to the main pattern. For example, the
1773 following matches, with the inner pattern capturing "B" and
1774 matching "BB", while the outer pattern captures "A";
1775 my $inner = '(.)\1';
1777 "ABBA" =~ /^(.)(??{ $inner })\1/;
1778 print $1; # prints "A";
1779 Note that this means that there is no way for the inner pattern to
1781 refer to a capture group defined outside. (The code block itself
1782 can use $1, etc., to refer to the enclosing pattern's capture
1783 groups.) Thus, although
1784 ('a' x 100)=~/(??{'(.)' x 100})/
1786 will match, it will not set $1 on exit.
1788 The following pattern matches a parenthesized group:
1790 $re = qr{
1792 \(
1793 (?:
1794 (?> [^()]+ ) # Non-parens without backtracking
1795 |
1796 (??{ $re }) # Group with matching parens
1797 )*
1798 \)
1799 }x;
1800 See also "(?PARNO)" for a different, more efficient way to
1802 accomplish the same task.
1803 Executing a postponed regular expression too many times without
1805 consuming any input string will also result in a fatal error. The
1806 depth at which that happens is compiled into perl, so it can be
1807 changed with a custom build.
1808 "(?PARNO)" "(?-PARNO)" "(?+PARNO)" "(?R)" "(?0)"
1810 Recursive subpattern. Treat the contents of a given capture buffer
1811 in the current pattern as an independent subpattern and attempt to
1812 match it at the current position in the string. Information about
1813 capture state from the caller for things like backreferences is
1814 available to the subpattern, but capture buffers set by the
1815 subpattern are not visible to the caller.
1816 Similar to "(??{ code })" except that it does not involve executing
1818 any code or potentially compiling a returned pattern string;
1819 instead it treats the part of the current pattern contained within
1820 a specified capture group as an independent pattern that must match
1821 at the current position. Also different is the treatment of capture
1822 buffers, unlike "(??{ code })" recursive patterns have access to
1823 their caller's match state, so one can use backreferences safely.
1824 PARNO is a sequence of digits (not starting with 0) whose value
1826 reflects the paren-number of the capture group to recurse to.
1827 "(?R)" recurses to the beginning of the whole pattern. "(?0)" is an
1828 alternate syntax for "(?R)". If PARNO is preceded by a plus or
1829 minus sign then it is assumed to be relative, with negative numbers
1830 indicating preceding capture groups and positive ones following.
1831 Thus "(?-1)" refers to the most recently declared group, and
1832 "(?+1)" indicates the next group to be declared. Note that the
1833 counting for relative recursion differs from that of relative
1834 backreferences, in that with recursion unclosed groups are
1835 included.
1836 The following pattern matches a function "foo()" which may contain
1838 balanced parentheses as the argument.
1839 $re = qr{ ( # paren group 1 (full function)
1841 foo
1842 ( # paren group 2 (parens)
1843 \(
1844 ( # paren group 3 (contents of parens)
1845 (?:
1846 (?> [^()]+ ) # Non-parens without backtracking
1847 |
1848 (?2) # Recurse to start of paren group 2
1849 )*
1850 )
1851 \)
1852 )
1853 )
1854 }x;
1855 If the pattern was used as follows
1857 'foo(bar(baz)+baz(bop))'=~/$re/
1859 and print "\$1 = $1\n",
1860 "\$2 = $2\n",
1861 "\$3 = $3\n";
1862 the output produced should be the following:
1864 $1 = foo(bar(baz)+baz(bop))
1866 $2 = (bar(baz)+baz(bop))
1867 $3 = bar(baz)+baz(bop)
1868 If there is no corresponding capture group defined, then it is a
1870 fatal error. Recursing deeply without consuming any input string
1871 will also result in a fatal error. The depth at which that happens
1872 is compiled into perl, so it can be changed with a custom build.
1873 The following shows how using negative indexing can make it easier
1875 to embed recursive patterns inside of a "qr//" construct for later
1876 use:
1877 my $parens = qr/(\((?:[^()]++|(?-1))*+\))/;
1879 if (/foo $parens \s+ \+ \s+ bar $parens/x) {
1880 # do something here...
1881 }
1882 Note that this pattern does not behave the same way as the
1884 equivalent PCRE or Python construct of the same form. In Perl you
1885 can backtrack into a recursed group, in PCRE and Python the
1886 recursed into group is treated as atomic. Also, modifiers are
1887 resolved at compile time, so constructs like "(?i:(?1))" or
1888 "(?:(?i)(?1))" do not affect how the sub-pattern will be processed.
1889 "(?&NAME)"
1891 Recurse to a named subpattern. Identical to "(?PARNO)" except that
1892 the parenthesis to recurse to is determined by name. If multiple
1893 parentheses have the same name, then it recurses to the leftmost.
1894 It is an error to refer to a name that is not declared somewhere in
1896 the pattern.
1897 NOTE: In order to make things easier for programmers with
1899 experience with the Python or PCRE regex engines the pattern
1900 "(?P>NAME)" may be used instead of "(?&NAME)".
1901 "(?(condition)yes-pattern|no-pattern)"
1903 "(?(condition)yes-pattern)"
1904 Conditional expression. Matches yes-pattern if condition yields a
1905 true value, matches no-pattern otherwise. A missing pattern always
1906 matches.
1907 "(condition)" should be one of:
1909 an integer in parentheses
1911 (which is valid if the corresponding pair of parentheses
1912 matched);
1913 a lookahead/lookbehind/evaluate zero-width assertion;
1915 a name in angle brackets or single quotes
1916 (which is valid if a group with the given name matched);
1917 the special symbol "(R)"
1919 (true when evaluated inside of recursion or eval).
1920 Additionally the "R" may be followed by a number, (which will
1921 be true when evaluated when recursing inside of the appropriate
1922 group), or by "&NAME", in which case it will be true only when
1923 evaluated during recursion in the named group.
1924 Here's a summary of the possible predicates:
1926 "(1)" "(2)" ...
1928 Checks if the numbered capturing group has matched something.
1929 Full syntax: "(?(1)then|else)"
1930 "(<NAME>)" "('NAME')"
1932 Checks if a group with the given name has matched something.
1933 Full syntax: "(?(<name>)then|else)"
1934 "(?=...)" "(?!...)" "(?<=...)" "(?<!...)"
1936 Checks whether the pattern matches (or does not match, for the
1937 "!" variants). Full syntax: "(?(?=lookahead)then|else)"
1938 "(?{ CODE })"
1940 Treats the return value of the code block as the condition.
1941 Full syntax: "(?(?{ code })then|else)"
1942 "(R)"
1944 Checks if the expression has been evaluated inside of
1945 recursion. Full syntax: "(?(R)then|else)"
1946 "(R1)" "(R2)" ...
1948 Checks if the expression has been evaluated while executing
1949 directly inside of the n-th capture group. This check is the
1950 regex equivalent of
1951 if ((caller(0))[3] eq 'subname') { ... }
1953 In other words, it does not check the full recursion stack.
1955 Full syntax: "(?(R1)then|else)"
1957 "(R&NAME)"
1959 Similar to "(R1)", this predicate checks to see if we're
1960 executing directly inside of the leftmost group with a given
1961 name (this is the same logic used by "(?&NAME)" to
1962 disambiguate). It does not check the full stack, but only the
1963 name of the innermost active recursion. Full syntax:
1964 "(?(R&name)then|else)"
1965 "(DEFINE)"
1967 In this case, the yes-pattern is never directly executed, and
1968 no no-pattern is allowed. Similar in spirit to "(?{0})" but
1969 more efficient. See below for details. Full syntax:
1970 "(?(DEFINE)definitions...)"
1971 For example:
1973 m{ ( \( )?
1975 [^()]+
1976 (?(1) \) )
1977 }x
1978 matches a chunk of non-parentheses, possibly included in
1980 parentheses themselves.
1981 A special form is the "(DEFINE)" predicate, which never executes
1983 its yes-pattern directly, and does not allow a no-pattern. This
1984 allows one to define subpatterns which will be executed only by the
1985 recursion mechanism. This way, you can define a set of regular
1986 expression rules that can be bundled into any pattern you choose.
1987 It is recommended that for this usage you put the DEFINE block at
1989 the end of the pattern, and that you name any subpatterns defined
1990 within it.
1991 Also, it's worth noting that patterns defined this way probably
1993 will not be as efficient, as the optimizer is not very clever about
1994 handling them.
1995 An example of how this might be used is as follows:
1997 /(?<NAME>(?&NAME_PAT))(?<ADDR>(?&ADDRESS_PAT))
1999 (?(DEFINE)
2000 (?<NAME_PAT>....)
2001 (?<ADDRESS_PAT>....)
2002 )/x
2003 Note that capture groups matched inside of recursion are not
2005 accessible after the recursion returns, so the extra layer of
2006 capturing groups is necessary. Thus $+{NAME_PAT} would not be
2007 defined even though $+{NAME} would be.
2008 Finally, keep in mind that subpatterns created inside a DEFINE
2010 block count towards the absolute and relative number of captures,
2011 so this:
2012 my @captures = "a" =~ /(.) # First capture
2014 (?(DEFINE)
2015 (?<EXAMPLE> 1 ) # Second capture
2016 )/x;
2017 say scalar @captures;
2018 Will output 2, not 1. This is particularly important if you intend
2020 to compile the definitions with the "qr//" operator, and later
2021 interpolate them in another pattern.
2022 "(?>pattern)"
2024 "(*atomic:pattern)"
2025 An "independent" subexpression, one which matches the substring
2026 that a standalone pattern would match if anchored at the given
2027 position, and it matches nothing other than this substring. This
2028 construct is useful for optimizations of what would otherwise be
2029 "eternal" matches, because it will not backtrack (see
2030 "Backtracking"). It may also be useful in places where the "grab
2031 all you can, and do not give anything back" semantic is desirable.
2032 For example: "^(?>a*)ab" will never match, since "(?>a*)" (anchored
2034 at the beginning of string, as above) will match all characters "a"
2035 at the beginning of string, leaving no "a" for "ab" to match. In
2036 contrast, "a*ab" will match the same as "a+b", since the match of
2037 the subgroup "a*" is influenced by the following group "ab" (see
2038 "Backtracking"). In particular, "a*" inside "a*ab" will match
2039 fewer characters than a standalone "a*", since this makes the tail
2040 match.
2041 "(?>pattern)" does not disable backtracking altogether once it has
2043 matched. It is still possible to backtrack past the construct, but
2044 not into it. So "((?>a*)|(?>b*))ar" will still match "bar".
2045 An effect similar to "(?>pattern)" may be achieved by writing
2047 "(?=(pattern))\g{-1}". This matches the same substring as a
2048 standalone "a+", and the following "\g{-1}" eats the matched
2049 string; it therefore makes a zero-length assertion into an analogue
2050 of "(?>...)". (The difference between these two constructs is that
2051 the second one uses a capturing group, thus shifting ordinals of
2052 backreferences in the rest of a regular expression.)
2053 Consider this pattern:
2055 m{ \(
2057 (
2058 [^()]+ # x+
2059 |
2060 \( [^()]* \)
2061 )+
2062 \)
2063 }x
2064 That will efficiently match a nonempty group with matching
2066 parentheses two levels deep or less. However, if there is no such
2067 group, it will take virtually forever on a long string. That's
2068 because there are so many different ways to split a long string
2069 into several substrings. This is what "(.+)+" is doing, and
2070 "(.+)+" is similar to a subpattern of the above pattern. Consider
2071 how the pattern above detects no-match on "((()aaaaaaaaaaaaaaaaaa"
2072 in several seconds, but that each extra letter doubles this time.
2073 This exponential performance will make it appear that your program
2074 has hung. However, a tiny change to this pattern
2075 m{ \(
2077 (
2078 (?> [^()]+ ) # change x+ above to (?> x+ )
2079 |
2080 \( [^()]* \)
2081 )+
2082 \)
2083 }x
2084 which uses "(?>...)" matches exactly when the one above does
2086 (verifying this yourself would be a productive exercise), but
2087 finishes in a fourth the time when used on a similar string with
2088 1000000 "a"s. Be aware, however, that, when this construct is
2089 followed by a quantifier, it currently triggers a warning message
2090 under the "use warnings" pragma or -w switch saying it "matches
2091 null string many times in regex".
2092 On simple groups, such as the pattern "(?> [^()]+ )", a comparable
2094 effect may be achieved by negative lookahead, as in "[^()]+ (?!
2095 [^()] )". This was only 4 times slower on a string with 1000000
2096 "a"s.
2097 The "grab all you can, and do not give anything back" semantic is
2099 desirable in many situations where on the first sight a simple
2100 "()*" looks like the correct solution. Suppose we parse text with
2101 comments being delimited by "#" followed by some optional
2102 (horizontal) whitespace. Contrary to its appearance, "#[ \t]*" is
2103 not the correct subexpression to match the comment delimiter,
2104 because it may "give up" some whitespace if the remainder of the
2105 pattern can be made to match that way. The correct answer is
2106 either one of these:
2107 (?>#[ \t]*)
2109 #[ \t]*(?![ \t])
2110 For example, to grab non-empty comments into $1, one should use
2112 either one of these:
2113 / (?> \# [ \t]* ) ( .+ ) /x;
2115 / \# [ \t]* ( [^ \t] .* ) /x;
2116 Which one you pick depends on which of these expressions better
2118 reflects the above specification of comments.
2119 In some literature this construct is called "atomic matching" or
2121 "possessive matching".
2122 Possessive quantifiers are equivalent to putting the item they are
2124 applied to inside of one of these constructs. The following
2125 equivalences apply:
2126 Quantifier Form Bracketing Form
2128 --------------- ---------------
2129 PAT*+ (?>PAT*)
2130 PAT++ (?>PAT+)
2131 PAT?+ (?>PAT?)
2132 PAT{min,max}+ (?>PAT{min,max})
2133 Nested "(?>...)" constructs are not no-ops, even if at first glance
2135 they might seem to be. This is because the nested "(?>...)" can
2136 restrict internal backtracking that otherwise might occur. For
2137 example,
2138 "abc" =~ /(?>a[bc]*c)/
2140 matches, but
2142 "abc" =~ /(?>a(?>[bc]*)c)/
2144 does not.
2146 "(?[ ])"
2148 See "Extended Bracketed Character Classes" in perlrecharclass.
2149 Note that this feature is currently experimental; using it yields a
2151 warning in the "experimental::regex_sets" category.
2152 Backtracking
2154 NOTE: This section presents an abstract approximation of regular
2155 expression behavior. For a more rigorous (and complicated) view of the
2156 rules involved in selecting a match among possible alternatives, see
2157 "Combining RE Pieces".
2158 A fundamental feature of regular expression matching involves the
2160 notion called backtracking, which is currently used (when needed) by
2161 all regular non-possessive expression quantifiers, namely "*", "*?",
2162 "+", "+?", "{n,m}", and "{n,m}?". Backtracking is often optimized
2163 internally, but the general principle outlined here is valid.
2164 For a regular expression to match, the entire regular expression must
2166 match, not just part of it. So if the beginning of a pattern
2167 containing a quantifier succeeds in a way that causes later parts in
2168 the pattern to fail, the matching engine backs up and recalculates the
2169 beginning part--that's why it's called backtracking.
2170 Here is an example of backtracking: Let's say you want to find the
2172 word following "foo" in the string "Food is on the foo table.":
2173 $_ = "Food is on the foo table.";
2175 if ( /\b(foo)\s+(\w+)/i ) {
2176 print "$2 follows $1.\n";
2177 }
2178 When the match runs, the first part of the regular expression
2180 ("\b(foo)") finds a possible match right at the beginning of the
2181 string, and loads up $1 with "Foo". However, as soon as the matching
2182 engine sees that there's no whitespace following the "Foo" that it had
2183 saved in $1, it realizes its mistake and starts over again one
2184 character after where it had the tentative match. This time it goes
2185 all the way until the next occurrence of "foo". The complete regular
2186 expression matches this time, and you get the expected output of "table
2187 follows foo."
2188 Sometimes minimal matching can help a lot. Imagine you'd like to match
2190 everything between "foo" and "bar". Initially, you write something
2191 like this:
2192 $_ = "The food is under the bar in the barn.";
2194 if ( /foo(.*)bar/ ) {
2195 print "got <$1>\n";
2196 }
2197 Which perhaps unexpectedly yields:
2199 got <d is under the bar in the >
2201 That's because ".*" was greedy, so you get everything between the first
2203 "foo" and the last "bar". Here it's more effective to use minimal
2204 matching to make sure you get the text between a "foo" and the first
2205 "bar" thereafter.
2206 if ( /foo(.*?)bar/ ) { print "got <$1>\n" }
2208 got <d is under the >
2209 Here's another example. Let's say you'd like to match a number at the
2211 end of a string, and you also want to keep the preceding part of the
2212 match. So you write this:
2213 $_ = "I have 2 numbers: 53147";
2215 if ( /(.*)(\d*)/ ) { # Wrong!
2216 print "Beginning is <$1>, number is <$2>.\n";
2217 }
2218 That won't work at all, because ".*" was greedy and gobbled up the
2220 whole string. As "\d*" can match on an empty string the complete
2221 regular expression matched successfully.
2222 Beginning is <I have 2 numbers: 53147>, number is <>.
2224 Here are some variants, most of which don't work:
2226 $_ = "I have 2 numbers: 53147";
2228 @pats = qw{
2229 (.*)(\d*)
2230 (.*)(\d+)
2231 (.*?)(\d*)
2232 (.*?)(\d+)
2233 (.*)(\d+)$
2234 (.*?)(\d+)$
2235 (.*)\b(\d+)$
2236 (.*\D)(\d+)$
2237 };
2238 for $pat (@pats) {
2240 printf "%-12s ", $pat;
2241 if ( /$pat/ ) {
2242 print "<$1> <$2>\n";
2243 } else {
2244 print "FAIL\n";
2245 }
2246 }
2247 That will print out:
2249 (.*)(\d*) <I have 2 numbers: 53147> <>
2251 (.*)(\d+) <I have 2 numbers: 5314> <7>
2252 (.*?)(\d*) <> <>
2253 (.*?)(\d+) <I have > <2>
2254 (.*)(\d+)$ <I have 2 numbers: 5314> <7>
2255 (.*?)(\d+)$ <I have 2 numbers: > <53147>
2256 (.*)\b(\d+)$ <I have 2 numbers: > <53147>
2257 (.*\D)(\d+)$ <I have 2 numbers: > <53147>
2258 As you see, this can be a bit tricky. It's important to realize that a
2260 regular expression is merely a set of assertions that gives a
2261 definition of success. There may be 0, 1, or several different ways
2262 that the definition might succeed against a particular string. And if
2263 there are multiple ways it might succeed, you need to understand
2264 backtracking to know which variety of success you will achieve.
2265 When using lookahead assertions and negations, this can all get even
2267 trickier. Imagine you'd like to find a sequence of non-digits not
2268 followed by "123". You might try to write that as
2269 $_ = "ABC123";
2271 if ( /^\D*(?!123)/ ) { # Wrong!
2272 print "Yup, no 123 in $_\n";
2273 }
2274 But that isn't going to match; at least, not the way you're hoping. It
2276 claims that there is no 123 in the string. Here's a clearer picture of
2277 why that pattern matches, contrary to popular expectations:
2278 $x = 'ABC123';
2280 $y = 'ABC445';
2281 print "1: got $1\n" if $x =~ /^(ABC)(?!123)/;
2283 print "2: got $1\n" if $y =~ /^(ABC)(?!123)/;
2284 print "3: got $1\n" if $x =~ /^(\D*)(?!123)/;
2286 print "4: got $1\n" if $y =~ /^(\D*)(?!123)/;
2287 This prints
2289 2: got ABC
2291 3: got AB
2292 4: got ABC
2293 You might have expected test 3 to fail because it seems to a more
2295 general purpose version of test 1. The important difference between
2296 them is that test 3 contains a quantifier ("\D*") and so can use
2297 backtracking, whereas test 1 will not. What's happening is that you've
2298 asked "Is it true that at the start of $x, following 0 or more non-
2299 digits, you have something that's not 123?" If the pattern matcher had
2300 let "\D*" expand to "ABC", this would have caused the whole pattern to
2301 fail.
2302 The search engine will initially match "\D*" with "ABC". Then it will
2304 try to match "(?!123)" with "123", which fails. But because a
2305 quantifier ("\D*") has been used in the regular expression, the search
2306 engine can backtrack and retry the match differently in the hope of
2307 matching the complete regular expression.
2308 The pattern really, really wants to succeed, so it uses the standard
2310 pattern back-off-and-retry and lets "\D*" expand to just "AB" this
2311 time. Now there's indeed something following "AB" that is not "123".
2312 It's "C123", which suffices.
2313 We can deal with this by using both an assertion and a negation. We'll
2315 say that the first part in $1 must be followed both by a digit and by
2316 something that's not "123". Remember that the lookaheads are zero-
2317 width expressions--they only look, but don't consume any of the string
2318 in their match. So rewriting this way produces what you'd expect; that
2319 is, case 5 will fail, but case 6 succeeds:
2320 print "5: got $1\n" if $x =~ /^(\D*)(?=\d)(?!123)/;
2322 print "6: got $1\n" if $y =~ /^(\D*)(?=\d)(?!123)/;
2323 6: got ABC
2325 In other words, the two zero-width assertions next to each other work
2327 as though they're ANDed together, just as you'd use any built-in
2328 assertions: "/^$/" matches only if you're at the beginning of the line
2329 AND the end of the line simultaneously. The deeper underlying truth is
2330 that juxtaposition in regular expressions always means AND, except when
2331 you write an explicit OR using the vertical bar. "/ab/" means match
2332 "a" AND (then) match "b", although the attempted matches are made at
2333 different positions because "a" is not a zero-width assertion, but a
2334 one-width assertion.
2335 WARNING: Particularly complicated regular expressions can take
2337 exponential time to solve because of the immense number of possible
2338 ways they can use backtracking to try for a match. For example,
2339 without internal optimizations done by the regular expression engine,
2340 this will take a painfully long time to run:
2341 'aaaaaaaaaaaa' =~ /((a{0,5}){0,5})*[c]/
2343 And if you used "*"'s in the internal groups instead of limiting them
2345 to 0 through 5 matches, then it would take forever--or until you ran
2346 out of stack space. Moreover, these internal optimizations are not
2347 always applicable. For example, if you put "{0,5}" instead of "*" on
2348 the external group, no current optimization is applicable, and the
2349 match takes a long time to finish.
2350 A powerful tool for optimizing such beasts is what is known as an
2352 "independent group", which does not backtrack (see "(?>pattern)").
2353 Note also that zero-length lookahead/lookbehind assertions will not
2354 backtrack to make the tail match, since they are in "logical" context:
2355 only whether they match is considered relevant. For an example where
2356 side-effects of lookahead might have influenced the following match,
2357 see "(?>pattern)".
2358 Script Runs
2360 A script run is basically a sequence of characters, all from the same
2361 Unicode script (see "Scripts" in perlunicode), such as Latin or Greek.
2362 In most places a single word would never be written in multiple
2363 scripts, unless it is a spoofing attack. An infamous example, is
2364 paypal.com
2366 Those letters could all be Latin (as in the example just above), or
2368 they could be all Cyrillic (except for the dot), or they could be a
2369 mixture of the two. In the case of an internet address the ".com"
2370 would be in Latin, And any Cyrillic ones would cause it to be a
2371 mixture, not a script run. Someone clicking on such a link would not
2372 be directed to the real Paypal website, but an attacker would craft a
2373 look-alike one to attempt to gather sensitive information from the
2374 person.
2375 Starting in Perl 5.28, it is now easy to detect strings that aren't
2377 script runs. Simply enclose just about any pattern like either of
2378 these:
2379 (*script_run:pattern)
2381 (*sr:pattern)
2382 What happens is that after pattern succeeds in matching, it is
2384 subjected to the additional criterion that every character in it must
2385 be from the same script (see exceptions below). If this isn't true,
2386 backtracking occurs until something all in the same script is found
2387 that matches, or all possibilities are exhausted. This can cause a lot
2388 of backtracking, but generally, only malicious input will result in
2389 this, though the slow down could cause a denial of service attack. If
2390 your needs permit, it is best to make the pattern atomic to cut down on
2391 the amount of backtracking. This is so likely to be what you want,
2392 that instead of writing this:
2393 (*script_run:(?>pattern))
2395 you can write either of these:
2397 (*atomic_script_run:pattern)
2399 (*asr:pattern)
2400 (See "(?>pattern)".)
2402 In Taiwan, Japan, and Korea, it is common for text to have a mixture of
2404 characters from their native scripts and base Chinese. Perl follows
2405 Unicode's UTS 39 (<https://unicode.org/reports/tr39/>) Unicode Security
2406 Mechanisms in allowing such mixtures. For example, the Japanese
2407 scripts Katakana and Hiragana are commonly mixed together in practice,
2408 along with some Chinese characters, and hence are treated as being in a
2409 single script run by Perl.
2410 The rules used for matching decimal digits are slightly stricter. Many
2412 scripts have their own sets of digits equivalent to the Western 0
2413 through 9 ones. A few, such as Arabic, have more than one set. For a
2414 string to be considered a script run, all digits in it must come from
2415 the same set of ten, as determined by the first digit encountered. As
2416 an example,
2417 qr/(*script_run: \d+ \b )/x
2419 guarantees that the digits matched will all be from the same set of 10.
2421 You won't get a look-alike digit from a different script that has a
2422 different value than what it appears to be.
2423 Unicode has three pseudo scripts that are handled specially.
2425 "Unknown" is applied to code points whose meaning has yet to be
2427 determined. Perl currently will match as a script run, any single
2428 character string consisting of one of these code points. But any
2429 string longer than one code point containing one of these will not be
2430 considered a script run.
2431 "Inherited" is applied to characters that modify another, such as an
2433 accent of some type. These are considered to be in the script of the
2434 master character, and so never cause a script run to not match.
2435 The other one is "Common". This consists of mostly punctuation, emoji,
2437 and characters used in mathematics and music, the ASCII digits 0
2438 through 9, and full-width forms of these digits. These characters can
2439 appear intermixed in text in many of the world's scripts. These also
2440 don't cause a script run to not match. But like other scripts, all
2441 digits in a run must come from the same set of 10.
2442 This construct is non-capturing. You can add parentheses to pattern to
2444 capture, if desired. You will have to do this if you plan to use
2445 "(*ACCEPT) (*ACCEPT:arg)" and not have it bypass the script run
2446 checking.
2447 The "Script_Extensions" property as modified by UTS 39
2449 (<https://unicode.org/reports/tr39/>) is used as the basis for this
2450 feature.
2451 To summarize,
2453 • All length 0 or length 1 sequences are script runs.
2455 • A longer sequence is a script run if and only if all of the
2457 following conditions are met:
2458 1. No code point in the sequence has the "Script_Extension"
2462 property of "Unknown".
2463 This currently means that all code points in the sequence have
2465 been assigned by Unicode to be characters that aren't private
2466 use nor surrogate code points.
2467 2. All characters in the sequence come from the Common script
2469 and/or the Inherited script and/or a single other script.
2470 The script of a character is determined by the
2472 "Script_Extensions" property as modified by UTS 39
2473 (<https://unicode.org/reports/tr39/>), as described above.
2474 3. All decimal digits in the sequence come from the same block of
2476 10 consecutive digits.
2477 Special Backtracking Control Verbs
2479 These special patterns are generally of the form "(*VERB:arg)". Unless
2480 otherwise stated the arg argument is optional; in some cases, it is
2481 mandatory.
2482 Any pattern containing a special backtracking verb that allows an
2484 argument has the special behaviour that when executed it sets the
2485 current package's $REGERROR and $REGMARK variables. When doing so the
2486 following rules apply:
2487 On failure, the $REGERROR variable will be set to the arg value of the
2489 verb pattern, if the verb was involved in the failure of the match. If
2490 the arg part of the pattern was omitted, then $REGERROR will be set to
2491 the name of the last "(*MARK:NAME)" pattern executed, or to TRUE if
2492 there was none. Also, the $REGMARK variable will be set to FALSE.
2493 On a successful match, the $REGERROR variable will be set to FALSE, and
2495 the $REGMARK variable will be set to the name of the last
2496 "(*MARK:NAME)" pattern executed. See the explanation for the
2497 "(*MARK:NAME)" verb below for more details.
2498 NOTE: $REGERROR and $REGMARK are not magic variables like $1 and most
2500 other regex-related variables. They are not local to a scope, nor
2501 readonly, but instead are volatile package variables similar to
2502 $AUTOLOAD. They are set in the package containing the code that
2503 executed the regex (rather than the one that compiled it, where those
2504 differ). If necessary, you can use "local" to localize changes to
2505 these variables to a specific scope before executing a regex.
2506 If a pattern does not contain a special backtracking verb that allows
2508 an argument, then $REGERROR and $REGMARK are not touched at all.
2509 Verbs
2511 "(*PRUNE)" "(*PRUNE:NAME)"
2512 This zero-width pattern prunes the backtracking tree at the
2513 current point when backtracked into on failure. Consider the
2514 pattern "/A (*PRUNE) B/", where A and B are complex patterns.
2515 Until the "(*PRUNE)" verb is reached, A may backtrack as
2516 necessary to match. Once it is reached, matching continues in B,
2517 which may also backtrack as necessary; however, should B not
2518 match, then no further backtracking will take place, and the
2519 pattern will fail outright at the current starting position.
2520 The following example counts all the possible matching strings
2522 in a pattern (without actually matching any of them).
2523 'aaab' =~ /a+b?(?{print "$&\n"; $count++})(*FAIL)/;
2525 print "Count=$count\n";
2526 which produces:
2528 aaab
2530 aaa
2531 aa
2532 a
2533 aab
2534 aa
2535 a
2536 ab
2537 a
2538 Count=9
2539 If we add a "(*PRUNE)" before the count like the following
2541 'aaab' =~ /a+b?(*PRUNE)(?{print "$&\n"; $count++})(*FAIL)/;
2543 print "Count=$count\n";
2544 we prevent backtracking and find the count of the longest
2546 matching string at each matching starting point like so:
2547 aaab
2549 aab
2550 ab
2551 Count=3
2552 Any number of "(*PRUNE)" assertions may be used in a pattern.
2554 See also "(?>pattern)" and possessive quantifiers for other ways
2556 to control backtracking. In some cases, the use of "(*PRUNE)"
2557 can be replaced with a "(?>pattern)" with no functional
2558 difference; however, "(*PRUNE)" can be used to handle cases that
2559 cannot be expressed using a "(?>pattern)" alone.
2560 "(*SKIP)" "(*SKIP:NAME)"
2562 This zero-width pattern is similar to "(*PRUNE)", except that on
2563 failure it also signifies that whatever text that was matched
2564 leading up to the "(*SKIP)" pattern being executed cannot be
2565 part of any match of this pattern. This effectively means that
2566 the regex engine "skips" forward to this position on failure and
2567 tries to match again, (assuming that there is sufficient room to
2568 match).
2569 The name of the "(*SKIP:NAME)" pattern has special significance.
2571 If a "(*MARK:NAME)" was encountered while matching, then it is
2572 that position which is used as the "skip point". If no "(*MARK)"
2573 of that name was encountered, then the "(*SKIP)" operator has no
2574 effect. When used without a name the "skip point" is where the
2575 match point was when executing the "(*SKIP)" pattern.
2576 Compare the following to the examples in "(*PRUNE)"; note the
2578 string is twice as long:
2579 'aaabaaab' =~ /a+b?(*SKIP)(?{print "$&\n"; $count++})(*FAIL)/;
2581 print "Count=$count\n";
2582 outputs
2584 aaab
2586 aaab
2587 Count=2
2588 Once the 'aaab' at the start of the string has matched, and the
2590 "(*SKIP)" executed, the next starting point will be where the
2591 cursor was when the "(*SKIP)" was executed.
2592 "(*MARK:NAME)" "(*:NAME)"
2594 This zero-width pattern can be used to mark the point reached in
2595 a string when a certain part of the pattern has been
2596 successfully matched. This mark may be given a name. A later
2597 "(*SKIP)" pattern will then skip forward to that point if
2598 backtracked into on failure. Any number of "(*MARK)" patterns
2599 are allowed, and the NAME portion may be duplicated.
2600 In addition to interacting with the "(*SKIP)" pattern,
2602 "(*MARK:NAME)" can be used to "label" a pattern branch, so that
2603 after matching, the program can determine which branches of the
2604 pattern were involved in the match.
2605 When a match is successful, the $REGMARK variable will be set to
2607 the name of the most recently executed "(*MARK:NAME)" that was
2608 involved in the match.
2609 This can be used to determine which branch of a pattern was
2611 matched without using a separate capture group for each branch,
2612 which in turn can result in a performance improvement, as perl
2613 cannot optimize "/(?:(x)|(y)|(z))/" as efficiently as something
2614 like "/(?:x(*MARK:x)|y(*MARK:y)|z(*MARK:z))/".
2615 When a match has failed, and unless another verb has been
2617 involved in failing the match and has provided its own name to
2618 use, the $REGERROR variable will be set to the name of the most
2619 recently executed "(*MARK:NAME)".
2620 See "(*SKIP)" for more details.
2622 As a shortcut "(*MARK:NAME)" can be written "(*:NAME)".
2624 "(*THEN)" "(*THEN:NAME)"
2626 This is similar to the "cut group" operator "::" from Raku.
2627 Like "(*PRUNE)", this verb always matches, and when backtracked
2628 into on failure, it causes the regex engine to try the next
2629 alternation in the innermost enclosing group (capturing or
2630 otherwise) that has alternations. The two branches of a
2631 "(?(condition)yes-pattern|no-pattern)" do not count as an
2632 alternation, as far as "(*THEN)" is concerned.
2633 Its name comes from the observation that this operation combined
2635 with the alternation operator ("|") can be used to create what
2636 is essentially a pattern-based if/then/else block:
2637 ( COND (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ )
2639 Note that if this operator is used and NOT inside of an
2641 alternation then it acts exactly like the "(*PRUNE)" operator.
2642 / A (*PRUNE) B /
2644 is the same as
2646 / A (*THEN) B /
2648 but
2650 / ( A (*THEN) B | C ) /
2652 is not the same as
2654 / ( A (*PRUNE) B | C ) /
2656 as after matching the A but failing on the B the "(*THEN)" verb
2658 will backtrack and try C; but the "(*PRUNE)" verb will simply
2659 fail.
2660 "(*COMMIT)" "(*COMMIT:arg)"
2662 This is the Raku "commit pattern" "<commit>" or ":::". It's a
2663 zero-width pattern similar to "(*SKIP)", except that when
2664 backtracked into on failure it causes the match to fail
2665 outright. No further attempts to find a valid match by advancing
2666 the start pointer will occur again. For example,
2667 'aaabaaab' =~ /a+b?(*COMMIT)(?{print "$&\n"; $count++})(*FAIL)/;
2669 print "Count=$count\n";
2670 outputs
2672 aaab
2674 Count=1
2675 In other words, once the "(*COMMIT)" has been entered, and if
2677 the pattern does not match, the regex engine will not try any
2678 further matching on the rest of the string.
2679 "(*FAIL)" "(*F)" "(*FAIL:arg)"
2681 This pattern matches nothing and always fails. It can be used to
2682 force the engine to backtrack. It is equivalent to "(?!)", but
2683 easier to read. In fact, "(?!)" gets optimised into "(*FAIL)"
2684 internally. You can provide an argument so that if the match
2685 fails because of this "FAIL" directive the argument can be
2686 obtained from $REGERROR.
2687 It is probably useful only when combined with "(?{})" or
2689 "(??{})".
2690 "(*ACCEPT)" "(*ACCEPT:arg)"
2692 This pattern matches nothing and causes the end of successful
2693 matching at the point at which the "(*ACCEPT)" pattern was
2694 encountered, regardless of whether there is actually more to
2695 match in the string. When inside of a nested pattern, such as
2696 recursion, or in a subpattern dynamically generated via
2697 "(??{})", only the innermost pattern is ended immediately.
2698 If the "(*ACCEPT)" is inside of capturing groups then the groups
2700 are marked as ended at the point at which the "(*ACCEPT)" was
2701 encountered. For instance:
2702 'AB' =~ /(A (A|B(*ACCEPT)|C) D)(E)/x;
2704 will match, and $1 will be "AB" and $2 will be "B", $3 will not
2706 be set. If another branch in the inner parentheses was matched,
2707 such as in the string 'ACDE', then the "D" and "E" would have to
2708 be matched as well.
2709 You can provide an argument, which will be available in the var
2711 $REGMARK after the match completes.
2712 Warning on "\1" Instead of $1
2714 Some people get too used to writing things like:
2715 $pattern =~ s/(\W)/\\\1/g;
2717 This is grandfathered (for \1 to \9) for the RHS of a substitute to
2719 avoid shocking the sed addicts, but it's a dirty habit to get into.
2720 That's because in PerlThink, the righthand side of an "s///" is a
2721 double-quoted string. "\1" in the usual double-quoted string means a
2722 control-A. The customary Unix meaning of "\1" is kludged in for
2723 "s///". However, if you get into the habit of doing that, you get
2724 yourself into trouble if you then add an "/e" modifier.
2725 s/(\d+)/ \1 + 1 /eg; # causes warning under -w
2727 Or if you try to do
2729 s/(\d+)/\1000/;
2731 You can't disambiguate that by saying "\{1}000", whereas you can fix it
2733 with "${1}000". The operation of interpolation should not be confused
2734 with the operation of matching a backreference. Certainly they mean
2735 two different things on the left side of the "s///".
2736 Repeated Patterns Matching a Zero-length Substring
2738 WARNING: Difficult material (and prose) ahead. This section needs a
2739 rewrite.
2740 Regular expressions provide a terse and powerful programming language.
2742 As with most other power tools, power comes together with the ability
2743 to wreak havoc.
2744 A common abuse of this power stems from the ability to make infinite
2746 loops using regular expressions, with something as innocuous as:
2747 'foo' =~ m{ ( o? )* }x;
2749 The "o?" matches at the beginning of ""foo"", and since the position in
2751 the string is not moved by the match, "o?" would match again and again
2752 because of the "*" quantifier. Another common way to create a similar
2753 cycle is with the looping modifier "/g":
2754 @matches = ( 'foo' =~ m{ o? }xg );
2756 or
2758 print "match: <$&>\n" while 'foo' =~ m{ o? }xg;
2760 or the loop implied by "split()".
2762 However, long experience has shown that many programming tasks may be
2764 significantly simplified by using repeated subexpressions that may
2765 match zero-length substrings. Here's a simple example being:
2766 @chars = split //, $string; # // is not magic in split
2768 ($whitewashed = $string) =~ s/()/ /g; # parens avoid magic s// /
2769 Thus Perl allows such constructs, by forcefully breaking the infinite
2771 loop. The rules for this are different for lower-level loops given by
2772 the greedy quantifiers "*+{}", and for higher-level ones like the "/g"
2773 modifier or "split()" operator.
2774 The lower-level loops are interrupted (that is, the loop is broken)
2776 when Perl detects that a repeated expression matched a zero-length
2777 substring. Thus
2778 m{ (?: NON_ZERO_LENGTH | ZERO_LENGTH )* }x;
2780 is made equivalent to
2782 m{ (?: NON_ZERO_LENGTH )* (?: ZERO_LENGTH )? }x;
2784 For example, this program
2786 #!perl -l
2788 "aaaaab" =~ /
2789 (?:
2790 a # non-zero
2791 | # or
2792 (?{print "hello"}) # print hello whenever this
2793 # branch is tried
2794 (?=(b)) # zero-width assertion
2795 )* # any number of times
2796 /x;
2797 print $&;
2798 print $1;
2799 prints
2801 hello
2803 aaaaa
2804 b
2805 Notice that "hello" is only printed once, as when Perl sees that the
2807 sixth iteration of the outermost "(?:)*" matches a zero-length string,
2808 it stops the "*".
2809 The higher-level loops preserve an additional state between iterations:
2811 whether the last match was zero-length. To break the loop, the
2812 following match after a zero-length match is prohibited to have a
2813 length of zero. This prohibition interacts with backtracking (see
2814 "Backtracking"), and so the second best match is chosen if the best
2815 match is of zero length.
2816 For example:
2818 $_ = 'bar';
2820 s/\w??/<$&>/g;
2821 results in "<><b><><a><><r><>". At each position of the string the
2823 best match given by non-greedy "??" is the zero-length match, and the
2824 second best match is what is matched by "\w". Thus zero-length matches
2825 alternate with one-character-long matches.
2826 Similarly, for repeated "m/()/g" the second-best match is the match at
2828 the position one notch further in the string.
2829 The additional state of being matched with zero-length is associated
2831 with the matched string, and is reset by each assignment to "pos()".
2832 Zero-length matches at the end of the previous match are ignored during
2833 "split".
2834 Combining RE Pieces
2836 Each of the elementary pieces of regular expressions which were
2837 described before (such as "ab" or "\Z") could match at most one
2838 substring at the given position of the input string. However, in a
2839 typical regular expression these elementary pieces are combined into
2840 more complicated patterns using combining operators "ST", "S|T", "S*"
2841 etc. (in these examples "S" and "T" are regular subexpressions).
2842 Such combinations can include alternatives, leading to a problem of
2844 choice: if we match a regular expression "a|ab" against "abc", will it
2845 match substring "a" or "ab"? One way to describe which substring is
2846 actually matched is the concept of backtracking (see "Backtracking").
2847 However, this description is too low-level and makes you think in terms
2848 of a particular implementation.
2849 Another description starts with notions of "better"/"worse". All the
2851 substrings which may be matched by the given regular expression can be
2852 sorted from the "best" match to the "worst" match, and it is the "best"
2853 match which is chosen. This substitutes the question of "what is
2854 chosen?" by the question of "which matches are better, and which are
2855 worse?".
2856 Again, for elementary pieces there is no such question, since at most
2858 one match at a given position is possible. This section describes the
2859 notion of better/worse for combining operators. In the description
2860 below "S" and "T" are regular subexpressions.
2861 "ST"
2863 Consider two possible matches, "AB" and "A'B'", "A" and "A'" are
2864 substrings which can be matched by "S", "B" and "B'" are substrings
2865 which can be matched by "T".
2866 If "A" is a better match for "S" than "A'", "AB" is a better match
2868 than "A'B'".
2869 If "A" and "A'" coincide: "AB" is a better match than "AB'" if "B"
2871 is a better match for "T" than "B'".
2872 "S|T"
2874 When "S" can match, it is a better match than when only "T" can
2875 match.
2876 Ordering of two matches for "S" is the same as for "S". Similar
2878 for two matches for "T".
2879 "S{REPEAT_COUNT}"
2881 Matches as "SSS...S" (repeated as many times as necessary).
2882 "S{min,max}"
2884 Matches as "S{max}|S{max-1}|...|S{min+1}|S{min}".
2885 "S{min,max}?"
2887 Matches as "S{min}|S{min+1}|...|S{max-1}|S{max}".
2888 "S?", "S*", "S+"
2890 Same as "S{0,1}", "S{0,BIG_NUMBER}", "S{1,BIG_NUMBER}"
2891 respectively.
2892 "S??", "S*?", "S+?"
2894 Same as "S{0,1}?", "S{0,BIG_NUMBER}?", "S{1,BIG_NUMBER}?"
2895 respectively.
2896 "(?>S)"
2898 Matches the best match for "S" and only that.
2899 "(?=S)", "(?<=S)"
2901 Only the best match for "S" is considered. (This is important only
2902 if "S" has capturing parentheses, and backreferences are used
2903 somewhere else in the whole regular expression.)
2904 "(?!S)", "(?<!S)"
2906 For this grouping operator there is no need to describe the
2907 ordering, since only whether or not "S" can match is important.
2908 "(??{ EXPR })", "(?PARNO)"
2910 The ordering is the same as for the regular expression which is the
2911 result of EXPR, or the pattern contained by capture group PARNO.
2912 "(?(condition)yes-pattern|no-pattern)"
2914 Recall that which of yes-pattern or no-pattern actually matches is
2915 already determined. The ordering of the matches is the same as for
2916 the chosen subexpression.
2917 The above recipes describe the ordering of matches at a given position.
2919 One more rule is needed to understand how a match is determined for the
2920 whole regular expression: a match at an earlier position is always
2921 better than a match at a later position.
2922 Creating Custom RE Engines
2924 As of Perl 5.10.0, one can create custom regular expression engines.
2925 This is not for the faint of heart, as they have to plug in at the C
2926 level. See perlreapi for more details.
2927 As an alternative, overloaded constants (see overload) provide a simple
2929 way to extend the functionality of the RE engine, by substituting one
2930 pattern for another.
2931 Suppose that we want to enable a new RE escape-sequence "\Y|" which
2933 matches at a boundary between whitespace characters and non-whitespace
2934 characters. Note that "(?=\S)(?<!\S)|(?!\S)(?<=\S)" matches exactly at
2935 these positions, so we want to have each "\Y|" in the place of the more
2936 complicated version. We can create a module "customre" to do this:
2937 package customre;
2939 use overload;
2940 sub import {
2942 shift;
2943 die "No argument to customre::import allowed" if @_;
2944 overload::constant 'qr' => \&convert;
2945 }
2946 sub invalid { die "/$_[0]/: invalid escape '\\$_[1]'"}
2948 # We must also take care of not escaping the legitimate \\Y|
2950 # sequence, hence the presence of '\\' in the conversion rules.
2951 my %rules = ( '\\' => '\\\\',
2952 'Y|' => qr/(?=\S)(?<!\S)|(?!\S)(?<=\S)/ );
2953 sub convert {
2954 my $re = shift;
2955 $re =~ s{
2956 \\ ( \\ | Y . )
2957 }
2958 { $rules{$1} or invalid($re,$1) }sgex;
2959 return $re;
2960 }
2961 Now "use customre" enables the new escape in constant regular
2963 expressions, i.e., those without any runtime variable interpolations.
2964 As documented in overload, this conversion will work only over literal
2965 parts of regular expressions. For "\Y|$re\Y|" the variable part of
2966 this regular expression needs to be converted explicitly (but only if
2967 the special meaning of "\Y|" should be enabled inside $re):
2968 use customre;
2970 $re = <>;
2971 chomp $re;
2972 $re = customre::convert $re;
2973 /\Y|$re\Y|/;
2974 Embedded Code Execution Frequency
2976 The exact rules for how often "(??{})" and "(?{})" are executed in a
2977 pattern are unspecified. In the case of a successful match you can
2978 assume that they DWIM and will be executed in left to right order the
2979 appropriate number of times in the accepting path of the pattern as
2980 would any other meta-pattern. How non-accepting pathways and match
2981 failures affect the number of times a pattern is executed is
2982 specifically unspecified and may vary depending on what optimizations
2983 can be applied to the pattern and is likely to change from version to
2984 version.
2985 For instance in
2987 "aaabcdeeeee"=~/a(?{print "a"})b(?{print "b"})cde/;
2989 the exact number of times "a" or "b" are printed out is unspecified for
2991 failure, but you may assume they will be printed at least once during a
2992 successful match, additionally you may assume that if "b" is printed,
2993 it will be preceded by at least one "a".
2994 In the case of branching constructs like the following:
2996 /a(b|(?{ print "a" }))c(?{ print "c" })/;
2998 you can assume that the input "ac" will output "ac", and that "abc"
3000 will output only "c".
3001 When embedded code is quantified, successful matches will call the code
3003 once for each matched iteration of the quantifier. For example:
3004 "good" =~ /g(?:o(?{print "o"}))*d/;
3006 will output "o" twice.
3008 PCRE/Python Support
3010 As of Perl 5.10.0, Perl supports several Python/PCRE-specific
3011 extensions to the regex syntax. While Perl programmers are encouraged
3012 to use the Perl-specific syntax, the following are also accepted:
3013 "(?P<NAME>pattern)"
3015 Define a named capture group. Equivalent to "(?<NAME>pattern)".
3016 "(?P=NAME)"
3018 Backreference to a named capture group. Equivalent to "\g{NAME}".
3019 "(?P>NAME)"
3021 Subroutine call to a named capture group. Equivalent to "(?&NAME)".
3022
3024 There are a number of issues with regard to case-insensitive matching
3025 in Unicode rules. See "i" under "Modifiers" above.
3026 This document varies from difficult to understand to completely and
3028 utterly opaque. The wandering prose riddled with jargon is hard to
3029 fathom in several places.
3030 This document needs a rewrite that separates the tutorial content from
3032 the reference content.
3033
3035 The syntax of patterns used in Perl pattern matching evolved from those
3036 supplied in the Bell Labs Research Unix 8th Edition (Version 8) regex
3037 routines. (The code is actually derived (distantly) from Henry
3038 Spencer's freely redistributable reimplementation of those V8
3039 routines.)
3040 perlrequick.
3042 perlretut.
3044 "Regexp Quote-Like Operators" in perlop.
3046 "Gory details of parsing quoted constructs" in perlop.
3048 perlfaq6.
3050 "pos" in perlfunc.
3052 perllocale.
3054 perlebcdic.
3056 Mastering Regular Expressions by Jeffrey Friedl, published by O'Reilly
3058 and Associates.
3059perl v5.34.1 2022-03-15 PERLRE(1)