3 perlre - Perl regular expressions
7 This page describes the syntax of regular expressions in Perl. For a
8 description of how to I<use> regular expressions in matching
9 operations, plus various examples of the same, see discussions
10 of C<m//>, C<s///>, C<qr//> and C<??> in L<perlop/"Regexp Quote-Like Operators">.
12 Matching operations can have various modifiers. Modifiers
13 that relate to the interpretation of the regular expression inside
14 are listed below. Modifiers that alter the way a regular expression
15 is used by Perl are detailed in L<perlop/"Regexp Quote-Like Operators"> and
16 L<perlop/"Gory details of parsing quoted constructs">.
22 Do case-insensitive pattern matching.
24 If C<use locale> is in effect, the case map is taken from the current
25 locale. See L<perllocale>.
29 Treat string as multiple lines. That is, change "^" and "$" from matching
30 the start or end of the string to matching the start or end of any
31 line anywhere within the string.
35 Treat string as single line. That is, change "." to match any character
36 whatsoever, even a newline, which normally it would not match.
38 The C</s> and C</m> modifiers both override the C<$*> setting. That
39 is, no matter what C<$*> contains, C</s> without C</m> will force
40 "^" to match only at the beginning of the string and "$" to match
41 only at the end (or just before a newline at the end) of the string.
42 Together, as /ms, they let the "." match any character whatsoever,
43 while yet allowing "^" and "$" to match, respectively, just after
44 and just before newlines within the string.
48 Extend your pattern's legibility by permitting whitespace and comments.
52 These are usually written as "the C</x> modifier", even though the delimiter
53 in question might not really be a slash. Any of these
54 modifiers may also be embedded within the regular expression itself using
55 the C<(?...)> construct. See below.
57 The C</x> modifier itself needs a little more explanation. It tells
58 the regular expression parser to ignore whitespace that is neither
59 backslashed nor within a character class. You can use this to break up
60 your regular expression into (slightly) more readable parts. The C<#>
61 character is also treated as a metacharacter introducing a comment,
62 just as in ordinary Perl code. This also means that if you want real
63 whitespace or C<#> characters in the pattern (outside a character
64 class, where they are unaffected by C</x>), that you'll either have to
65 escape them or encode them using octal or hex escapes. Taken together,
66 these features go a long way towards making Perl's regular expressions
67 more readable. Note that you have to be careful not to include the
68 pattern delimiter in the comment--perl has no way of knowing you did
69 not intend to close the pattern early. See the C-comment deletion code
72 =head2 Regular Expressions
74 The patterns used in Perl pattern matching derive from supplied in
75 the Version 8 regex routines. (The routines are derived
76 (distantly) from Henry Spencer's freely redistributable reimplementation
77 of the V8 routines.) See L<Version 8 Regular Expressions> for
80 In particular the following metacharacters have their standard I<egrep>-ish
83 \ Quote the next metacharacter
84 ^ Match the beginning of the line
85 . Match any character (except newline)
86 $ Match the end of the line (or before newline at the end)
91 By default, the "^" character is guaranteed to match only the
92 beginning of the string, the "$" character only the end (or before the
93 newline at the end), and Perl does certain optimizations with the
94 assumption that the string contains only one line. Embedded newlines
95 will not be matched by "^" or "$". You may, however, wish to treat a
96 string as a multi-line buffer, such that the "^" will match after any
97 newline within the string, and "$" will match before any newline. At the
98 cost of a little more overhead, you can do this by using the /m modifier
99 on the pattern match operator. (Older programs did this by setting C<$*>,
100 but this practice is now deprecated.)
102 To simplify multi-line substitutions, the "." character never matches a
103 newline unless you use the C</s> modifier, which in effect tells Perl to pretend
104 the string is a single line--even if it isn't. The C</s> modifier also
105 overrides the setting of C<$*>, in case you have some (badly behaved) older
106 code that sets it in another module.
108 The following standard quantifiers are recognized:
110 * Match 0 or more times
111 + Match 1 or more times
113 {n} Match exactly n times
114 {n,} Match at least n times
115 {n,m} Match at least n but not more than m times
117 (If a curly bracket occurs in any other context, it is treated
118 as a regular character.) The "*" modifier is equivalent to C<{0,}>, the "+"
119 modifier to C<{1,}>, and the "?" modifier to C<{0,1}>. n and m are limited
120 to integral values less than a preset limit defined when perl is built.
121 This is usually 32766 on the most common platforms. The actual limit can
122 be seen in the error message generated by code such as this:
124 $_ **= $_ , / {$_} / for 2 .. 42;
126 By default, a quantified subpattern is "greedy", that is, it will match as
127 many times as possible (given a particular starting location) while still
128 allowing the rest of the pattern to match. If you want it to match the
129 minimum number of times possible, follow the quantifier with a "?". Note
130 that the meanings don't change, just the "greediness":
132 *? Match 0 or more times
133 +? Match 1 or more times
135 {n}? Match exactly n times
136 {n,}? Match at least n times
137 {n,m}? Match at least n but not more than m times
139 Because patterns are processed as double quoted strings, the following
146 \a alarm (bell) (BEL)
147 \e escape (think troff) (ESC)
148 \033 octal char (think of a PDP-11)
150 \x{263a} wide hex char (Unicode SMILEY)
153 \l lowercase next char (think vi)
154 \u uppercase next char (think vi)
155 \L lowercase till \E (think vi)
156 \U uppercase till \E (think vi)
157 \E end case modification (think vi)
158 \Q quote (disable) pattern metacharacters till \E
160 If C<use locale> is in effect, the case map used by C<\l>, C<\L>, C<\u>
161 and C<\U> is taken from the current locale. See L<perllocale>. For
162 documentation of C<\N{name}>, see L<charnames>.
164 You cannot include a literal C<$> or C<@> within a C<\Q> sequence.
165 An unescaped C<$> or C<@> interpolates the corresponding variable,
166 while escaping will cause the literal string C<\$> to be matched.
167 You'll need to write something like C<m/\Quser\E\@\Qhost/>.
169 In addition, Perl defines the following:
171 \w Match a "word" character (alphanumeric plus "_")
172 \W Match a non-word character
173 \s Match a whitespace character
174 \S Match a non-whitespace character
175 \d Match a digit character
176 \D Match a non-digit character
177 \pP Match P, named property. Use \p{Prop} for longer names.
179 \X Match eXtended Unicode "combining character sequence",
180 equivalent to C<(?:\PM\pM*)>
181 \C Match a single C char (octet) even under utf8.
183 A C<\w> matches a single alphanumeric character, not a whole word.
184 Use C<\w+> to match a string of Perl-identifier characters (which isn't
185 the same as matching an English word). If C<use locale> is in effect, the
186 list of alphabetic characters generated by C<\w> is taken from the
187 current locale. See L<perllocale>. You may use C<\w>, C<\W>, C<\s>, C<\S>,
188 C<\d>, and C<\D> within character classes, but if you try to use them
189 as endpoints of a range, that's not a range, the "-" is understood literally.
190 See L<utf8> for details about C<\pP>, C<\PP>, and C<\X>.
192 The POSIX character class syntax
196 is also available. The available classes and their backslash
197 equivalents (if available) are as follows:
213 For example use C<[:upper:]> to match all the uppercase characters.
214 Note that the C<[]> are part of the C<[::]> construct, not part of the whole
215 character class. For example:
219 matches one, zero, any alphabetic character, and the percentage sign.
221 If the C<utf8> pragma is used, the following equivalences to Unicode
222 \p{} constructs hold:
238 For example C<[:lower:]> and C<\p{IsLower}> are equivalent.
240 If the C<utf8> pragma is not used but the C<locale> pragma is, the
241 classes correlate with the isalpha(3) interface (except for `word',
242 which is a Perl extension, mirroring C<\w>).
244 The assumedly non-obviously named classes are:
250 Any control character. Usually characters that don't produce
251 output as such but instead control the terminal somehow:
252 for example newline and backspace are control characters.
253 All characters with ord() less than 32 are most often control
254 classified as characters.
258 Any alphanumeric or punctuation character.
262 Any alphanumeric or punctuation character or space.
266 Any punctuation character.
270 Any hexadecimal digit. Though this may feel silly
271 (/0-9a-f/i would work just fine) it is included
278 You can negate the [::] character classes by prefixing the class name
279 with a '^'. This is a Perl extension. For example:
281 POSIX trad. Perl utf8 Perl
283 [:^digit:] \D \P{IsDigit}
284 [:^space:] \S \P{IsSpace}
285 [:^word:] \W \P{IsWord}
287 The POSIX character classes [.cc.] and [=cc=] are recognized but
288 B<not> supported and trying to use them will cause an error.
290 Perl defines the following zero-width assertions:
292 \b Match a word boundary
293 \B Match a non-(word boundary)
294 \A Match only at beginning of string
295 \Z Match only at end of string, or before newline at the end
296 \z Match only at end of string
297 \G Match only at pos() (e.g. at the end-of-match position
300 A word boundary (C<\b>) is a spot between two characters
301 that has a C<\w> on one side of it and a C<\W> on the other side
302 of it (in either order), counting the imaginary characters off the
303 beginning and end of the string as matching a C<\W>. (Within
304 character classes C<\b> represents backspace rather than a word
305 boundary, just as it normally does in any double-quoted string.)
306 The C<\A> and C<\Z> are just like "^" and "$", except that they
307 won't match multiple times when the C</m> modifier is used, while
308 "^" and "$" will match at every internal line boundary. To match
309 the actual end of the string and not ignore an optional trailing
312 The C<\G> assertion can be used to chain global matches (using
313 C<m//g>), as described in L<perlop/"Regexp Quote-Like Operators">.
314 It is also useful when writing C<lex>-like scanners, when you have
315 several patterns that you want to match against consequent substrings
316 of your string, see the previous reference. The actual location
317 where C<\G> will match can also be influenced by using C<pos()> as
318 an lvalue. See L<perlfunc/pos>.
320 The bracketing construct C<( ... )> creates capture buffers. To
321 refer to the digit'th buffer use \E<lt>digitE<gt> within the
322 match. Outside the match use "$" instead of "\". (The
323 \E<lt>digitE<gt> notation works in certain circumstances outside
324 the match. See the warning below about \1 vs $1 for details.)
325 Referring back to another part of the match is called a
328 There is no limit to the number of captured substrings that you may
329 use. However Perl also uses \10, \11, etc. as aliases for \010,
330 \011, etc. (Recall that 0 means octal, so \011 is the 9'th ASCII
331 character, a tab.) Perl resolves this ambiguity by interpreting
332 \10 as a backreference only if at least 10 left parentheses have
333 opened before it. Likewise \11 is a backreference only if at least
334 11 left parentheses have opened before it. And so on. \1 through
335 \9 are always interpreted as backreferences."
339 s/^([^ ]*) *([^ ]*)/$2 $1/; # swap first two words
341 if (/(.)\1/) { # find first doubled char
342 print "'$1' is the first doubled character\n";
345 if (/Time: (..):(..):(..)/) { # parse out values
351 Several special variables also refer back to portions of the previous
352 match. C<$+> returns whatever the last bracket match matched.
353 C<$&> returns the entire matched string. (At one point C<$0> did
354 also, but now it returns the name of the program.) C<$`> returns
355 everything before the matched string. And C<$'> returns everything
356 after the matched string.
358 The numbered variables ($1, $2, $3, etc.) and the related punctuation
359 set (C<<$+>, C<$&>, C<$`>, and C<$'>) are all dynamically scoped
360 until the end of the enclosing block or until the next successful
361 match, whichever comes first. (See L<perlsyn/"Compound Statements">.)
363 B<WARNING>: Once Perl sees that you need one of C<$&>, C<$`>, or
364 C<$'> anywhere in the program, it has to provide them for every
365 pattern match. This may substantially slow your program. Perl
366 uses the same mechanism to produce $1, $2, etc, so you also pay a
367 price for each pattern that contains capturing parentheses. (To
368 avoid this cost while retaining the grouping behaviour, use the
369 extended regular expression C<(?: ... )> instead.) But if you never
370 use C<$&>, C<$`> or C<$'>, then patterns I<without> capturing
371 parentheses will not be penalized. So avoid C<$&>, C<$'>, and C<$`>
372 if you can, but if you can't (and some algorithms really appreciate
373 them), once you've used them once, use them at will, because you've
374 already paid the price. As of 5.005, C<$&> is not so costly as the
377 Backslashed metacharacters in Perl are alphanumeric, such as C<\b>,
378 C<\w>, C<\n>. Unlike some other regular expression languages, there
379 are no backslashed symbols that aren't alphanumeric. So anything
380 that looks like \\, \(, \), \E<lt>, \E<gt>, \{, or \} is always
381 interpreted as a literal character, not a metacharacter. This was
382 once used in a common idiom to disable or quote the special meanings
383 of regular expression metacharacters in a string that you want to
384 use for a pattern. Simply quote all non-alphanumeric characters:
386 $pattern =~ s/(\W)/\\$1/g;
388 Today it is more common to use the quotemeta() function or the C<\Q>
389 metaquoting escape sequence to disable all metacharacters' special
392 /$unquoted\Q$quoted\E$unquoted/
394 Beware that if you put literal backslashes (those not inside
395 interpolated variables) between C<\Q> and C<\E>, double-quotish
396 backslash interpolation may lead to confusing results. If you
397 I<need> to use literal backslashes within C<\Q...\E>,
398 consult L<perlop/"Gory details of parsing quoted constructs">.
400 =head2 Extended Patterns
402 Perl also defines a consistent extension syntax for features not
403 found in standard tools like B<awk> and B<lex>. The syntax is a
404 pair of parentheses with a question mark as the first thing within
405 the parentheses. The character after the question mark indicates
408 The stability of these extensions varies widely. Some have been
409 part of the core language for many years. Others are experimental
410 and may change without warning or be completely removed. Check
411 the documentation on an individual feature to verify its current
414 A question mark was chosen for this and for the minimal-matching
415 construct because 1) question marks are rare in older regular
416 expressions, and 2) whenever you see one, you should stop and
417 "question" exactly what is going on. That's psychology...
423 A comment. The text is ignored. If the C</x> modifier enables
424 whitespace formatting, a simple C<#> will suffice. Note that Perl closes
425 the comment as soon as it sees a C<)>, so there is no way to put a literal
428 =item C<(?imsx-imsx)>
430 One or more embedded pattern-match modifiers. This is particularly
431 useful for dynamic patterns, such as those read in from a configuration
432 file, read in as an argument, are specified in a table somewhere,
433 etc. Consider the case that some of which want to be case sensitive
434 and some do not. The case insensitive ones need to include merely
435 C<(?i)> at the front of the pattern. For example:
438 if ( /$pattern/i ) { }
442 $pattern = "(?i)foobar";
443 if ( /$pattern/ ) { }
445 Letters after a C<-> turn those modifiers off. These modifiers are
446 localized inside an enclosing group (if any). For example,
450 will match a repeated (I<including the case>!) word C<blah> in any
451 case, assuming C<x> modifier, and no C<i> modifier outside this
456 =item C<(?imsx-imsx:pattern)>
458 This is for clustering, not capturing; it groups subexpressions like
459 "()", but doesn't make backreferences as "()" does. So
461 @fields = split(/\b(?:a|b|c)\b/)
465 @fields = split(/\b(a|b|c)\b/)
467 but doesn't spit out extra fields. It's also cheaper not to capture
468 characters if you don't need to.
470 Any letters between C<?> and C<:> act as flags modifiers as with
471 C<(?imsx-imsx)>. For example,
473 /(?s-i:more.*than).*million/i
475 is equivalent to the more verbose
477 /(?:(?s-i)more.*than).*million/i
481 A zero-width positive look-ahead assertion. For example, C</\w+(?=\t)/>
482 matches a word followed by a tab, without including the tab in C<$&>.
486 A zero-width negative look-ahead assertion. For example C</foo(?!bar)/>
487 matches any occurrence of "foo" that isn't followed by "bar". Note
488 however that look-ahead and look-behind are NOT the same thing. You cannot
489 use this for look-behind.
491 If you are looking for a "bar" that isn't preceded by a "foo", C</(?!foo)bar/>
492 will not do what you want. That's because the C<(?!foo)> is just saying that
493 the next thing cannot be "foo"--and it's not, it's a "bar", so "foobar" will
494 match. You would have to do something like C</(?!foo)...bar/> for that. We
495 say "like" because there's the case of your "bar" not having three characters
496 before it. You could cover that this way: C</(?:(?!foo)...|^.{0,2})bar/>.
497 Sometimes it's still easier just to say:
499 if (/bar/ && $` !~ /foo$/)
501 For look-behind see below.
503 =item C<(?E<lt>=pattern)>
505 A zero-width positive look-behind assertion. For example, C</(?E<lt>=\t)\w+/>
506 matches a word that follows a tab, without including the tab in C<$&>.
507 Works only for fixed-width look-behind.
509 =item C<(?<!pattern)>
511 A zero-width negative look-behind assertion. For example C</(?<!bar)foo/>
512 matches any occurrence of "foo" that does not follow "bar". Works
513 only for fixed-width look-behind.
517 B<WARNING>: This extended regular expression feature is considered
518 highly experimental, and may be changed or deleted without notice.
520 This zero-width assertion evaluate any embedded Perl code. It
521 always succeeds, and its C<code> is not interpolated. Currently,
522 the rules to determine where the C<code> ends are somewhat convoluted.
524 The C<code> is properly scoped in the following sense: If the assertion
525 is backtracked (compare L<"Backtracking">), all changes introduced after
526 C<local>ization are undone, so that
530 (?{ $cnt = 0 }) # Initialize $cnt.
534 local $cnt = $cnt + 1; # Update $cnt, backtracking-safe.
538 (?{ $res = $cnt }) # On success copy to non-localized
542 will set C<$res = 4>. Note that after the match, $cnt returns to the globally
543 introduced value, because the scopes that restrict C<local> operators
546 This assertion may be used as a C<(?(condition)yes-pattern|no-pattern)>
547 switch. If I<not> used in this way, the result of evaluation of
548 C<code> is put into the special variable C<$^R>. This happens
549 immediately, so C<$^R> can be used from other C<(?{ code })> assertions
550 inside the same regular expression.
552 The assignment to C<$^R> above is properly localized, so the old
553 value of C<$^R> is restored if the assertion is backtracked; compare
556 For reasons of security, this construct is forbidden if the regular
557 expression involves run-time interpolation of variables, unless the
558 perilous C<use re 'eval'> pragma has been used (see L<re>), or the
559 variables contain results of C<qr//> operator (see
560 L<perlop/"qr/STRING/imosx">).
562 This restriction is because of the wide-spread and remarkably convenient
563 custom of using run-time determined strings as patterns. For example:
569 Before Perl knew how to execute interpolated code within a pattern,
570 this operation was completely safe from a security point of view,
571 although it could raise an exception from an illegal pattern. If
572 you turn on the C<use re 'eval'>, though, it is no longer secure,
573 so you should only do so if you are also using taint checking.
574 Better yet, use the carefully constrained evaluation within a Safe
575 module. See L<perlsec> for details about both these mechanisms.
577 =item C<(?p{ code })>
579 B<WARNING>: This extended regular expression feature is considered
580 highly experimental, and may be changed or deleted without notice.
581 A simplified version of the syntax may be introduced for commonly
584 This is a "postponed" regular subexpression. The C<code> is evaluated
585 at run time, at the moment this subexpression may match. The result
586 of evaluation is considered as a regular expression and matched as
587 if it were inserted instead of this construct.
589 The C<code> is not interpolated. As before, the rules to determine
590 where the C<code> ends are currently somewhat convoluted.
592 The following pattern matches a parenthesized group:
597 (?> [^()]+ ) # Non-parens without backtracking
599 (?p{ $re }) # Group with matching parens
604 =item C<(?E<gt>pattern)>
606 B<WARNING>: This extended regular expression feature is considered
607 highly experimental, and may be changed or deleted without notice.
609 An "independent" subexpression, one which matches the substring
610 that a I<standalone> C<pattern> would match if anchored at the given
611 position, and it matches I<nothing other than this substring>. This
612 construct is useful for optimizations of what would otherwise be
613 "eternal" matches, because it will not backtrack (see L<"Backtracking">).
614 It may also be useful in places where the "grab all you can, and do not
615 give anything back" semantic is desirable.
617 For example: C<^(?E<gt>a*)ab> will never match, since C<(?E<gt>a*)>
618 (anchored at the beginning of string, as above) will match I<all>
619 characters C<a> at the beginning of string, leaving no C<a> for
620 C<ab> to match. In contrast, C<a*ab> will match the same as C<a+b>,
621 since the match of the subgroup C<a*> is influenced by the following
622 group C<ab> (see L<"Backtracking">). In particular, C<a*> inside
623 C<a*ab> will match fewer characters than a standalone C<a*>, since
624 this makes the tail match.
626 An effect similar to C<(?E<gt>pattern)> may be achieved by writing
627 C<(?=(pattern))\1>. This matches the same substring as a standalone
628 C<a+>, and the following C<\1> eats the matched string; it therefore
629 makes a zero-length assertion into an analogue of C<(?E<gt>...)>.
630 (The difference between these two constructs is that the second one
631 uses a capturing group, thus shifting ordinals of backreferences
632 in the rest of a regular expression.)
634 Consider this pattern:
645 That will efficiently match a nonempty group with matching parentheses
646 two levels deep or less. However, if there is no such group, it
647 will take virtually forever on a long string. That's because there
648 are so many different ways to split a long string into several
649 substrings. This is what C<(.+)+> is doing, and C<(.+)+> is similar
650 to a subpattern of the above pattern. Consider how the pattern
651 above detects no-match on C<((()aaaaaaaaaaaaaaaaaa> in several
652 seconds, but that each extra letter doubles this time. This
653 exponential performance will make it appear that your program has
654 hung. However, a tiny change to this pattern
658 (?> [^()]+ ) # change x+ above to (?> x+ )
665 which uses C<(?E<gt>...)> matches exactly when the one above does (verifying
666 this yourself would be a productive exercise), but finishes in a fourth
667 the time when used on a similar string with 1000000 C<a>s. Be aware,
668 however, that this pattern currently triggers a warning message under
669 B<-w> saying it C<"matches the null string many times">):
671 On simple groups, such as the pattern C<(?E<gt> [^()]+ )>, a comparable
672 effect may be achieved by negative look-ahead, as in C<[^()]+ (?! [^()] )>.
673 This was only 4 times slower on a string with 1000000 C<a>s.
675 The "grab all you can, and do not give anything back" semantic is desirable
676 in many situations where on the first sight a simple C<()*> looks like
677 the correct solution. Suppose we parse text with comments being delimited
678 by C<#> followed by some optional (horizontal) whitespace. Contrary to
679 its appearence, C<#[ \t]*> I<is not> the correct subexpression to match
680 the comment delimiter, because it may "give up" some whitespace if
681 the remainder of the pattern can be made to match that way. The correct
682 answer is either one of these:
687 For example, to grab non-empty comments into $1, one should use either
690 / (?> \# [ \t]* ) ( .+ ) /x;
691 / \# [ \t]* ( [^ \t] .* ) /x;
693 Which one you pick depends on which of these expressions better reflects
694 the above specification of comments.
696 =item C<(?(condition)yes-pattern|no-pattern)>
698 =item C<(?(condition)yes-pattern)>
700 B<WARNING>: This extended regular expression feature is considered
701 highly experimental, and may be changed or deleted without notice.
703 Conditional expression. C<(condition)> should be either an integer in
704 parentheses (which is valid if the corresponding pair of parentheses
705 matched), or look-ahead/look-behind/evaluate zero-width assertion.
714 matches a chunk of non-parentheses, possibly included in parentheses
721 NOTE: This section presents an abstract approximation of regular
722 expression behavior. For a more rigorous (and complicated) view of
723 the rules involved in selecting a match among possible alternatives,
724 see L<Combining pieces together>.
726 A fundamental feature of regular expression matching involves the
727 notion called I<backtracking>, which is currently used (when needed)
728 by all regular expression quantifiers, namely C<*>, C<*?>, C<+>,
729 C<+?>, C<{n,m}>, and C<{n,m}?>. Backtracking is often optimized
730 internally, but the general principle outlined here is valid.
732 For a regular expression to match, the I<entire> regular expression must
733 match, not just part of it. So if the beginning of a pattern containing a
734 quantifier succeeds in a way that causes later parts in the pattern to
735 fail, the matching engine backs up and recalculates the beginning
736 part--that's why it's called backtracking.
738 Here is an example of backtracking: Let's say you want to find the
739 word following "foo" in the string "Food is on the foo table.":
741 $_ = "Food is on the foo table.";
742 if ( /\b(foo)\s+(\w+)/i ) {
743 print "$2 follows $1.\n";
746 When the match runs, the first part of the regular expression (C<\b(foo)>)
747 finds a possible match right at the beginning of the string, and loads up
748 $1 with "Foo". However, as soon as the matching engine sees that there's
749 no whitespace following the "Foo" that it had saved in $1, it realizes its
750 mistake and starts over again one character after where it had the
751 tentative match. This time it goes all the way until the next occurrence
752 of "foo". The complete regular expression matches this time, and you get
753 the expected output of "table follows foo."
755 Sometimes minimal matching can help a lot. Imagine you'd like to match
756 everything between "foo" and "bar". Initially, you write something
759 $_ = "The food is under the bar in the barn.";
760 if ( /foo(.*)bar/ ) {
764 Which perhaps unexpectedly yields:
766 got <d is under the bar in the >
768 That's because C<.*> was greedy, so you get everything between the
769 I<first> "foo" and the I<last> "bar". Here it's more effective
770 to use minimal matching to make sure you get the text between a "foo"
771 and the first "bar" thereafter.
773 if ( /foo(.*?)bar/ ) { print "got <$1>\n" }
774 got <d is under the >
776 Here's another example: let's say you'd like to match a number at the end
777 of a string, and you also want to keep the preceding part the match.
780 $_ = "I have 2 numbers: 53147";
781 if ( /(.*)(\d*)/ ) { # Wrong!
782 print "Beginning is <$1>, number is <$2>.\n";
785 That won't work at all, because C<.*> was greedy and gobbled up the
786 whole string. As C<\d*> can match on an empty string the complete
787 regular expression matched successfully.
789 Beginning is <I have 2 numbers: 53147>, number is <>.
791 Here are some variants, most of which don't work:
793 $_ = "I have 2 numbers: 53147";
806 printf "%-12s ", $pat;
816 (.*)(\d*) <I have 2 numbers: 53147> <>
817 (.*)(\d+) <I have 2 numbers: 5314> <7>
819 (.*?)(\d+) <I have > <2>
820 (.*)(\d+)$ <I have 2 numbers: 5314> <7>
821 (.*?)(\d+)$ <I have 2 numbers: > <53147>
822 (.*)\b(\d+)$ <I have 2 numbers: > <53147>
823 (.*\D)(\d+)$ <I have 2 numbers: > <53147>
825 As you see, this can be a bit tricky. It's important to realize that a
826 regular expression is merely a set of assertions that gives a definition
827 of success. There may be 0, 1, or several different ways that the
828 definition might succeed against a particular string. And if there are
829 multiple ways it might succeed, you need to understand backtracking to
830 know which variety of success you will achieve.
832 When using look-ahead assertions and negations, this can all get even
833 tricker. Imagine you'd like to find a sequence of non-digits not
834 followed by "123". You might try to write that as
837 if ( /^\D*(?!123)/ ) { # Wrong!
838 print "Yup, no 123 in $_\n";
841 But that isn't going to match; at least, not the way you're hoping. It
842 claims that there is no 123 in the string. Here's a clearer picture of
843 why it that pattern matches, contrary to popular expectations:
848 print "1: got $1\n" if $x =~ /^(ABC)(?!123)/ ;
849 print "2: got $1\n" if $y =~ /^(ABC)(?!123)/ ;
851 print "3: got $1\n" if $x =~ /^(\D*)(?!123)/ ;
852 print "4: got $1\n" if $y =~ /^(\D*)(?!123)/ ;
860 You might have expected test 3 to fail because it seems to a more
861 general purpose version of test 1. The important difference between
862 them is that test 3 contains a quantifier (C<\D*>) and so can use
863 backtracking, whereas test 1 will not. What's happening is
864 that you've asked "Is it true that at the start of $x, following 0 or more
865 non-digits, you have something that's not 123?" If the pattern matcher had
866 let C<\D*> expand to "ABC", this would have caused the whole pattern to
869 The search engine will initially match C<\D*> with "ABC". Then it will
870 try to match C<(?!123> with "123", which fails. But because
871 a quantifier (C<\D*>) has been used in the regular expression, the
872 search engine can backtrack and retry the match differently
873 in the hope of matching the complete regular expression.
875 The pattern really, I<really> wants to succeed, so it uses the
876 standard pattern back-off-and-retry and lets C<\D*> expand to just "AB" this
877 time. Now there's indeed something following "AB" that is not
878 "123". It's "C123", which suffices.
880 We can deal with this by using both an assertion and a negation.
881 We'll say that the first part in $1 must be followed both by a digit
882 and by something that's not "123". Remember that the look-aheads
883 are zero-width expressions--they only look, but don't consume any
884 of the string in their match. So rewriting this way produces what
885 you'd expect; that is, case 5 will fail, but case 6 succeeds:
887 print "5: got $1\n" if $x =~ /^(\D*)(?=\d)(?!123)/ ;
888 print "6: got $1\n" if $y =~ /^(\D*)(?=\d)(?!123)/ ;
892 In other words, the two zero-width assertions next to each other work as though
893 they're ANDed together, just as you'd use any built-in assertions: C</^$/>
894 matches only if you're at the beginning of the line AND the end of the
895 line simultaneously. The deeper underlying truth is that juxtaposition in
896 regular expressions always means AND, except when you write an explicit OR
897 using the vertical bar. C</ab/> means match "a" AND (then) match "b",
898 although the attempted matches are made at different positions because "a"
899 is not a zero-width assertion, but a one-width assertion.
901 B<WARNING>: particularly complicated regular expressions can take
902 exponential time to solve because of the immense number of possible
903 ways they can use backtracking to try match. For example, without
904 internal optimizations done by the regular expression engine, this will
905 take a painfully long time to run:
907 'aaaaaaaaaaaa' =~ /((a{0,5}){0,5}){0,5}[c]/
909 And if you used C<*>'s instead of limiting it to 0 through 5 matches,
910 then it would take forever--or until you ran out of stack space.
912 A powerful tool for optimizing such beasts is what is known as an
914 which does not backtrack (see L<C<(?E<gt>pattern)>>). Note also that
915 zero-length look-ahead/look-behind assertions will not backtrack to make
916 the tail match, since they are in "logical" context: only
917 whether they match is considered relevant. For an example
918 where side-effects of look-ahead I<might> have influenced the
919 following match, see L<C<(?E<gt>pattern)>>.
921 =head2 Version 8 Regular Expressions
923 In case you're not familiar with the "regular" Version 8 regex
924 routines, here are the pattern-matching rules not described above.
926 Any single character matches itself, unless it is a I<metacharacter>
927 with a special meaning described here or above. You can cause
928 characters that normally function as metacharacters to be interpreted
929 literally by prefixing them with a "\" (e.g., "\." matches a ".", not any
930 character; "\\" matches a "\"). A series of characters matches that
931 series of characters in the target string, so the pattern C<blurfl>
932 would match "blurfl" in the target string.
934 You can specify a character class, by enclosing a list of characters
935 in C<[]>, which will match any one character from the list. If the
936 first character after the "[" is "^", the class matches any character not
937 in the list. Within a list, the "-" character specifies a
938 range, so that C<a-z> represents all characters between "a" and "z",
939 inclusive. If you want either "-" or "]" itself to be a member of a
940 class, put it at the start of the list (possibly after a "^"), or
941 escape it with a backslash. "-" is also taken literally when it is
942 at the end of the list, just before the closing "]". (The
943 following all specify the same class of three characters: C<[-az]>,
944 C<[az-]>, and C<[a\-z]>. All are different from C<[a-z]>, which
945 specifies a class containing twenty-six characters.)
946 Also, if you try to use the character classes C<\w>, C<\W>, C<\s>,
947 C<\S>, C<\d>, or C<\D> as endpoints of a range, that's not a range,
948 the "-" is understood literally.
950 Note also that the whole range idea is rather unportable between
951 character sets--and even within character sets they may cause results
952 you probably didn't expect. A sound principle is to use only ranges
953 that begin from and end at either alphabets of equal case ([a-e],
954 [A-E]), or digits ([0-9]). Anything else is unsafe. If in doubt,
955 spell out the character sets in full.
957 Characters may be specified using a metacharacter syntax much like that
958 used in C: "\n" matches a newline, "\t" a tab, "\r" a carriage return,
959 "\f" a form feed, etc. More generally, \I<nnn>, where I<nnn> is a string
960 of octal digits, matches the character whose ASCII value is I<nnn>.
961 Similarly, \xI<nn>, where I<nn> are hexadecimal digits, matches the
962 character whose ASCII value is I<nn>. The expression \cI<x> matches the
963 ASCII character control-I<x>. Finally, the "." metacharacter matches any
964 character except "\n" (unless you use C</s>).
966 You can specify a series of alternatives for a pattern using "|" to
967 separate them, so that C<fee|fie|foe> will match any of "fee", "fie",
968 or "foe" in the target string (as would C<f(e|i|o)e>). The
969 first alternative includes everything from the last pattern delimiter
970 ("(", "[", or the beginning of the pattern) up to the first "|", and
971 the last alternative contains everything from the last "|" to the next
972 pattern delimiter. That's why it's common practice to include
973 alternatives in parentheses: to minimize confusion about where they
976 Alternatives are tried from left to right, so the first
977 alternative found for which the entire expression matches, is the one that
978 is chosen. This means that alternatives are not necessarily greedy. For
979 example: when matching C<foo|foot> against "barefoot", only the "foo"
980 part will match, as that is the first alternative tried, and it successfully
981 matches the target string. (This might not seem important, but it is
982 important when you are capturing matched text using parentheses.)
984 Also remember that "|" is interpreted as a literal within square brackets,
985 so if you write C<[fee|fie|foe]> you're really only matching C<[feio|]>.
987 Within a pattern, you may designate subpatterns for later reference
988 by enclosing them in parentheses, and you may refer back to the
989 I<n>th subpattern later in the pattern using the metacharacter
990 \I<n>. Subpatterns are numbered based on the left to right order
991 of their opening parenthesis. A backreference matches whatever
992 actually matched the subpattern in the string being examined, not
993 the rules for that subpattern. Therefore, C<(0|0x)\d*\s\1\d*> will
994 match "0x1234 0x4321", but not "0x1234 01234", because subpattern
995 1 matched "0x", even though the rule C<0|0x> could potentially match
996 the leading 0 in the second number.
998 =head2 Warning on \1 vs $1
1000 Some people get too used to writing things like:
1002 $pattern =~ s/(\W)/\\\1/g;
1004 This is grandfathered for the RHS of a substitute to avoid shocking the
1005 B<sed> addicts, but it's a dirty habit to get into. That's because in
1006 PerlThink, the righthand side of a C<s///> is a double-quoted string. C<\1> in
1007 the usual double-quoted string means a control-A. The customary Unix
1008 meaning of C<\1> is kludged in for C<s///>. However, if you get into the habit
1009 of doing that, you get yourself into trouble if you then add an C</e>
1012 s/(\d+)/ \1 + 1 /eg; # causes warning under -w
1018 You can't disambiguate that by saying C<\{1}000>, whereas you can fix it with
1019 C<${1}000>. The operation of interpolation should not be confused
1020 with the operation of matching a backreference. Certainly they mean two
1021 different things on the I<left> side of the C<s///>.
1023 =head2 Repeated patterns matching zero-length substring
1025 B<WARNING>: Difficult material (and prose) ahead. This section needs a rewrite.
1027 Regular expressions provide a terse and powerful programming language. As
1028 with most other power tools, power comes together with the ability
1031 A common abuse of this power stems from the ability to make infinite
1032 loops using regular expressions, with something as innocuous as:
1034 'foo' =~ m{ ( o? )* }x;
1036 The C<o?> can match at the beginning of C<'foo'>, and since the position
1037 in the string is not moved by the match, C<o?> would match again and again
1038 because of the C<*> modifier. Another common way to create a similar cycle
1039 is with the looping modifier C<//g>:
1041 @matches = ( 'foo' =~ m{ o? }xg );
1045 print "match: <$&>\n" while 'foo' =~ m{ o? }xg;
1047 or the loop implied by split().
1049 However, long experience has shown that many programming tasks may
1050 be significantly simplified by using repeated subexpressions that
1051 may match zero-length substrings. Here's a simple example being:
1053 @chars = split //, $string; # // is not magic in split
1054 ($whitewashed = $string) =~ s/()/ /g; # parens avoid magic s// /
1056 Thus Perl allows such constructs, by I<forcefully breaking
1057 the infinite loop>. The rules for this are different for lower-level
1058 loops given by the greedy modifiers C<*+{}>, and for higher-level
1059 ones like the C</g> modifier or split() operator.
1061 The lower-level loops are I<interrupted> (that is, the loop is
1062 broken) when Perl detects that a repeated expression matched a
1063 zero-length substring. Thus
1065 m{ (?: NON_ZERO_LENGTH | ZERO_LENGTH )* }x;
1067 is made equivalent to
1069 m{ (?: NON_ZERO_LENGTH )*
1074 The higher level-loops preserve an additional state between iterations:
1075 whether the last match was zero-length. To break the loop, the following
1076 match after a zero-length match is prohibited to have a length of zero.
1077 This prohibition interacts with backtracking (see L<"Backtracking">),
1078 and so the I<second best> match is chosen if the I<best> match is of
1086 results in C<"<><b><><a><><r><>">. At each position of the string the best
1087 match given by non-greedy C<??> is the zero-length match, and the I<second
1088 best> match is what is matched by C<\w>. Thus zero-length matches
1089 alternate with one-character-long matches.
1091 Similarly, for repeated C<m/()/g> the second-best match is the match at the
1092 position one notch further in the string.
1094 The additional state of being I<matched with zero-length> is associated with
1095 the matched string, and is reset by each assignment to pos().
1096 Zero-length matches at the end of the previous match are ignored
1099 =head2 Combining pieces together
1101 Each of the elementary pieces of regular expressions which were described
1102 before (such as C<ab> or C<\Z>) could match at most one substring
1103 at the given position of the input string. However, in a typical regular
1104 expression these elementary pieces are combined into more complicated
1105 patterns using combining operators C<ST>, C<S|T>, C<S*> etc
1106 (in these examples C<S> and C<T> are regular subexpressions).
1108 Such combinations can include alternatives, leading to a problem of choice:
1109 if we match a regular expression C<a|ab> against C<"abc">, will it match
1110 substring C<"a"> or C<"ab">? One way to describe which substring is
1111 actually matched is the concept of backtracking (see L<"Backtracking">).
1112 However, this description is too low-level and makes you think
1113 in terms of a particular implementation.
1115 Another description starts with notions of "better"/"worse". All the
1116 substrings which may be matched by the given regular expression can be
1117 sorted from the "best" match to the "worst" match, and it is the "best"
1118 match which is chosen. This substitutes the question of "what is chosen?"
1119 by the question of "which matches are better, and which are worse?".
1121 Again, for elementary pieces there is no such question, since at most
1122 one match at a given position is possible. This section describes the
1123 notion of better/worse for combining operators. In the description
1124 below C<S> and C<T> are regular subexpressions.
1130 Consider two possible matches, C<AB> and C<A'B'>, C<A> and C<A'> are
1131 substrings which can be matched by C<S>, C<B> and C<B'> are substrings
1132 which can be matched by C<T>.
1134 If C<A> is better match for C<S> than C<A'>, C<AB> is a better
1137 If C<A> and C<A'> coincide: C<AB> is a better match than C<AB'> if
1138 C<B> is better match for C<T> than C<B'>.
1142 When C<S> can match, it is a better match than when only C<T> can match.
1144 Ordering of two matches for C<S> is the same as for C<S>. Similar for
1145 two matches for C<T>.
1147 =item C<S{REPEAT_COUNT}>
1149 Matches as C<SSS...S> (repeated as many times as necessary).
1153 Matches as C<S{max}|S{max-1}|...|S{min+1}|S{min}>.
1155 =item C<S{min,max}?>
1157 Matches as C<S{min}|S{min+1}|...|S{max-1}|S{max}>.
1159 =item C<S?>, C<S*>, C<S+>
1161 Same as C<S{0,1}>, C<S{0,BIG_NUMBER}>, C<S{1,BIG_NUMBER}> respectively.
1163 =item C<S??>, C<S*?>, C<S+?>
1165 Same as C<S{0,1}?>, C<S{0,BIG_NUMBER}?>, C<S{1,BIG_NUMBER}?> respectively.
1169 Matches the best match for C<S> and only that.
1171 =item C<(?=S)>, C<(?<=S)>
1173 Only the best match for C<S> is considered. (This is important only if
1174 C<S> has capturing parentheses, and backreferences are used somewhere
1175 else in the whole regular expression.)
1177 =item C<(?!S)>, C<(?<!S)>
1179 For this grouping operator there is no need to describe the ordering, since
1180 only whether or not C<S> can match is important.
1182 =item C<(?p{ EXPR })>
1184 The ordering is the same as for the regular expression which is
1187 =item C<(?(condition)yes-pattern|no-pattern)>
1189 Recall that which of C<yes-pattern> or C<no-pattern> actually matches is
1190 already determined. The ordering of the matches is the same as for the
1191 chosen subexpression.
1195 The above recipes describe the ordering of matches I<at a given position>.
1196 One more rule is needed to understand how a match is determined for the
1197 whole regular expression: a match at an earlier position is always better
1198 than a match at a later position.
1200 =head2 Creating custom RE engines
1202 Overloaded constants (see L<overload>) provide a simple way to extend
1203 the functionality of the RE engine.
1205 Suppose that we want to enable a new RE escape-sequence C<\Y|> which
1206 matches at boundary between white-space characters and non-whitespace
1207 characters. Note that C<(?=\S)(?<!\S)|(?!\S)(?<=\S)> matches exactly
1208 at these positions, so we want to have each C<\Y|> in the place of the
1209 more complicated version. We can create a module C<customre> to do
1217 die "No argument to customre::import allowed" if @_;
1218 overload::constant 'qr' => \&convert;
1221 sub invalid { die "/$_[0]/: invalid escape '\\$_[1]'"}
1223 my %rules = ( '\\' => '\\',
1224 'Y|' => qr/(?=\S)(?<!\S)|(?!\S)(?<=\S)/ );
1230 { $rules{$1} or invalid($re,$1) }sgex;
1234 Now C<use customre> enables the new escape in constant regular
1235 expressions, i.e., those without any runtime variable interpolations.
1236 As documented in L<overload>, this conversion will work only over
1237 literal parts of regular expressions. For C<\Y|$re\Y|> the variable
1238 part of this regular expression needs to be converted explicitly
1239 (but only if the special meaning of C<\Y|> should be enabled inside $re):
1244 $re = customre::convert $re;
1249 This document varies from difficult to understand to completely
1250 and utterly opaque. The wandering prose riddled with jargon is
1251 hard to fathom in several places.
1253 This document needs a rewrite that separates the tutorial content
1254 from the reference content.
1258 L<perlop/"Regexp Quote-Like Operators">.
1260 L<perlop/"Gory details of parsing quoted constructs">.
1268 I<Mastering Regular Expressions> by Jeffrey Friedl, published
1269 by O'Reilly and Associates.