This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlretut: #109408
[perl5.git] / pod / perlretut.pod
CommitLineData
47f9c88b
GS
1=head1 NAME
2
3perlretut - Perl regular expressions tutorial
4
5=head1 DESCRIPTION
6
7This page provides a basic tutorial on understanding, creating and
8using regular expressions in Perl. It serves as a complement to the
9reference page on regular expressions L<perlre>. Regular expressions
10are an integral part of the C<m//>, C<s///>, C<qr//> and C<split>
11operators and so this tutorial also overlaps with
12L<perlop/"Regexp Quote-Like Operators"> and L<perlfunc/split>.
13
14Perl is widely renowned for excellence in text processing, and regular
15expressions are one of the big factors behind this fame. Perl regular
16expressions display an efficiency and flexibility unknown in most
17other computer languages. Mastering even the basics of regular
18expressions will allow you to manipulate text with surprising ease.
19
20What is a regular expression? A regular expression is simply a string
21that describes a pattern. Patterns are in common use these days;
22examples are the patterns typed into a search engine to find web pages
23and the patterns used to list files in a directory, e.g., C<ls *.txt>
24or C<dir *.*>. In Perl, the patterns described by regular expressions
25are used to search strings, extract desired parts of strings, and to
26do search and replace operations.
27
28Regular expressions have the undeserved reputation of being abstract
29and difficult to understand. Regular expressions are constructed using
30simple concepts like conditionals and loops and are no more difficult
31to understand than the corresponding C<if> conditionals and C<while>
32loops in the Perl language itself. In fact, the main challenge in
33learning regular expressions is just getting used to the terse
34notation used to express these concepts.
35
36This tutorial flattens the learning curve by discussing regular
37expression concepts, along with their notation, one at a time and with
38many examples. The first part of the tutorial will progress from the
39simplest word searches to the basic regular expression concepts. If
40you master the first part, you will have all the tools needed to solve
41about 98% of your needs. The second part of the tutorial is for those
42comfortable with the basics and hungry for more power tools. It
43discusses the more advanced regular expression operators and
8ccb1477 44introduces the latest cutting-edge innovations.
47f9c88b
GS
45
46A note: to save time, 'regular expression' is often abbreviated as
47regexp or regex. Regexp is a more natural abbreviation than regex, but
48is harder to pronounce. The Perl pod documentation is evenly split on
49regexp vs regex; in Perl, there is more than one way to abbreviate it.
50We'll use regexp in this tutorial.
51
52=head1 Part 1: The basics
53
54=head2 Simple word matching
55
56The simplest regexp is simply a word, or more generally, a string of
57characters. A regexp consisting of a word matches any string that
58contains that word:
59
60 "Hello World" =~ /World/; # matches
61
7638d2dc 62What is this Perl statement all about? C<"Hello World"> is a simple
8ccb1477 63double-quoted string. C<World> is the regular expression and the
7638d2dc 64C<//> enclosing C</World/> tells Perl to search a string for a match.
47f9c88b
GS
65The operator C<=~> associates the string with the regexp match and
66produces a true value if the regexp matched, or false if the regexp
67did not match. In our case, C<World> matches the second word in
68C<"Hello World">, so the expression is true. Expressions like this
69are useful in conditionals:
70
71 if ("Hello World" =~ /World/) {
72 print "It matches\n";
73 }
74 else {
75 print "It doesn't match\n";
76 }
77
78There are useful variations on this theme. The sense of the match can
7638d2dc 79be reversed by using the C<!~> operator:
47f9c88b
GS
80
81 if ("Hello World" !~ /World/) {
82 print "It doesn't match\n";
83 }
84 else {
85 print "It matches\n";
86 }
87
88The literal string in the regexp can be replaced by a variable:
89
90 $greeting = "World";
91 if ("Hello World" =~ /$greeting/) {
92 print "It matches\n";
93 }
94 else {
95 print "It doesn't match\n";
96 }
97
98If you're matching against the special default variable C<$_>, the
99C<$_ =~> part can be omitted:
100
101 $_ = "Hello World";
102 if (/World/) {
103 print "It matches\n";
104 }
105 else {
106 print "It doesn't match\n";
107 }
108
109And finally, the C<//> default delimiters for a match can be changed
110to arbitrary delimiters by putting an C<'m'> out front:
111
112 "Hello World" =~ m!World!; # matches, delimited by '!'
113 "Hello World" =~ m{World}; # matches, note the matching '{}'
a6b2f353
GS
114 "/usr/bin/perl" =~ m"/perl"; # matches after '/usr/bin',
115 # '/' becomes an ordinary char
47f9c88b
GS
116
117C</World/>, C<m!World!>, and C<m{World}> all represent the
7638d2dc
WL
118same thing. When, e.g., the quote (C<">) is used as a delimiter, the forward
119slash C<'/'> becomes an ordinary character and can be used in this regexp
47f9c88b
GS
120without trouble.
121
122Let's consider how different regexps would match C<"Hello World">:
123
124 "Hello World" =~ /world/; # doesn't match
125 "Hello World" =~ /o W/; # matches
126 "Hello World" =~ /oW/; # doesn't match
127 "Hello World" =~ /World /; # doesn't match
128
129The first regexp C<world> doesn't match because regexps are
130case-sensitive. The second regexp matches because the substring
7638d2dc 131S<C<'o W'>> occurs in the string S<C<"Hello World">>. The space
47f9c88b
GS
132character ' ' is treated like any other character in a regexp and is
133needed to match in this case. The lack of a space character is the
134reason the third regexp C<'oW'> doesn't match. The fourth regexp
135C<'World '> doesn't match because there is a space at the end of the
136regexp, but not at the end of the string. The lesson here is that
137regexps must match a part of the string I<exactly> in order for the
138statement to be true.
139
7638d2dc 140If a regexp matches in more than one place in the string, Perl will
47f9c88b
GS
141always match at the earliest possible point in the string:
142
143 "Hello World" =~ /o/; # matches 'o' in 'Hello'
144 "That hat is red" =~ /hat/; # matches 'hat' in 'That'
145
146With respect to character matching, there are a few more points you
147need to know about. First of all, not all characters can be used 'as
7638d2dc 148is' in a match. Some characters, called I<metacharacters>, are reserved
47f9c88b
GS
149for use in regexp notation. The metacharacters are
150
151 {}[]()^$.|*+?\
152
153The significance of each of these will be explained
154in the rest of the tutorial, but for now, it is important only to know
155that a metacharacter can be matched by putting a backslash before it:
156
157 "2+2=4" =~ /2+2/; # doesn't match, + is a metacharacter
158 "2+2=4" =~ /2\+2/; # matches, \+ is treated like an ordinary +
159 "The interval is [0,1)." =~ /[0,1)./ # is a syntax error!
160 "The interval is [0,1)." =~ /\[0,1\)\./ # matches
7638d2dc 161 "#!/usr/bin/perl" =~ /#!\/usr\/bin\/perl/; # matches
47f9c88b
GS
162
163In the last regexp, the forward slash C<'/'> is also backslashed,
164because it is used to delimit the regexp. This can lead to LTS
165(leaning toothpick syndrome), however, and it is often more readable
166to change delimiters.
167
7638d2dc 168 "#!/usr/bin/perl" =~ m!#\!/usr/bin/perl!; # easier to read
47f9c88b
GS
169
170The backslash character C<'\'> is a metacharacter itself and needs to
171be backslashed:
172
173 'C:\WIN32' =~ /C:\\WIN/; # matches
174
175In addition to the metacharacters, there are some ASCII characters
176which don't have printable character equivalents and are instead
7638d2dc 177represented by I<escape sequences>. Common examples are C<\t> for a
47f9c88b 178tab, C<\n> for a newline, C<\r> for a carriage return and C<\a> for a
43e59f7b 179bell (or alert). If your string is better thought of as a sequence of arbitrary
47f9c88b
GS
180bytes, the octal escape sequence, e.g., C<\033>, or hexadecimal escape
181sequence, e.g., C<\x1B> may be a more natural representation for your
182bytes. Here are some examples of escapes:
183
184 "1000\t2000" =~ m(0\t2) # matches
185 "1000\n2000" =~ /0\n20/ # matches
186 "1000\t2000" =~ /\000\t2/ # doesn't match, "0" ne "\000"
f0a2b745
KW
187 "cat" =~ /\o{143}\x61\x74/ # matches in ASCII, but a weird way
188 # to spell cat
47f9c88b
GS
189
190If you've been around Perl a while, all this talk of escape sequences
191may seem familiar. Similar escape sequences are used in double-quoted
192strings and in fact the regexps in Perl are mostly treated as
193double-quoted strings. This means that variables can be used in
194regexps as well. Just like double-quoted strings, the values of the
195variables in the regexp will be substituted in before the regexp is
196evaluated for matching purposes. So we have:
197
198 $foo = 'house';
199 'housecat' =~ /$foo/; # matches
200 'cathouse' =~ /cat$foo/; # matches
47f9c88b
GS
201 'housecat' =~ /${foo}cat/; # matches
202
203So far, so good. With the knowledge above you can already perform
204searches with just about any literal string regexp you can dream up.
205Here is a I<very simple> emulation of the Unix grep program:
206
207 % cat > simple_grep
208 #!/usr/bin/perl
209 $regexp = shift;
210 while (<>) {
211 print if /$regexp/;
212 }
213 ^D
214
215 % chmod +x simple_grep
216
217 % simple_grep abba /usr/dict/words
218 Babbage
219 cabbage
220 cabbages
221 sabbath
222 Sabbathize
223 Sabbathizes
224 sabbatical
225 scabbard
226 scabbards
227
228This program is easy to understand. C<#!/usr/bin/perl> is the standard
229way to invoke a perl program from the shell.
7638d2dc 230S<C<$regexp = shift;>> saves the first command line argument as the
47f9c88b 231regexp to be used, leaving the rest of the command line arguments to
7638d2dc
WL
232be treated as files. S<C<< while (<>) >>> loops over all the lines in
233all the files. For each line, S<C<print if /$regexp/;>> prints the
47f9c88b
GS
234line if the regexp matches the line. In this line, both C<print> and
235C</$regexp/> use the default variable C<$_> implicitly.
236
237With all of the regexps above, if the regexp matched anywhere in the
238string, it was considered a match. Sometimes, however, we'd like to
239specify I<where> in the string the regexp should try to match. To do
7638d2dc 240this, we would use the I<anchor> metacharacters C<^> and C<$>. The
47f9c88b
GS
241anchor C<^> means match at the beginning of the string and the anchor
242C<$> means match at the end of the string, or before a newline at the
243end of the string. Here is how they are used:
244
245 "housekeeper" =~ /keeper/; # matches
246 "housekeeper" =~ /^keeper/; # doesn't match
247 "housekeeper" =~ /keeper$/; # matches
248 "housekeeper\n" =~ /keeper$/; # matches
249
250The second regexp doesn't match because C<^> constrains C<keeper> to
251match only at the beginning of the string, but C<"housekeeper"> has
252keeper starting in the middle. The third regexp does match, since the
253C<$> constrains C<keeper> to match only at the end of the string.
254
255When both C<^> and C<$> are used at the same time, the regexp has to
256match both the beginning and the end of the string, i.e., the regexp
257matches the whole string. Consider
258
259 "keeper" =~ /^keep$/; # doesn't match
260 "keeper" =~ /^keeper$/; # matches
261 "" =~ /^$/; # ^$ matches an empty string
262
263The first regexp doesn't match because the string has more to it than
264C<keep>. Since the second regexp is exactly the string, it
265matches. Using both C<^> and C<$> in a regexp forces the complete
266string to match, so it gives you complete control over which strings
267match and which don't. Suppose you are looking for a fellow named
268bert, off in a string by himself:
269
270 "dogbert" =~ /bert/; # matches, but not what you want
271
272 "dilbert" =~ /^bert/; # doesn't match, but ..
273 "bertram" =~ /^bert/; # matches, so still not good enough
274
275 "bertram" =~ /^bert$/; # doesn't match, good
276 "dilbert" =~ /^bert$/; # doesn't match, good
277 "bert" =~ /^bert$/; # matches, perfect
278
279Of course, in the case of a literal string, one could just as easily
7638d2dc 280use the string comparison S<C<$string eq 'bert'>> and it would be
47f9c88b
GS
281more efficient. The C<^...$> regexp really becomes useful when we
282add in the more powerful regexp tools below.
283
284=head2 Using character classes
285
286Although one can already do quite a lot with the literal string
287regexps above, we've only scratched the surface of regular expression
288technology. In this and subsequent sections we will introduce regexp
289concepts (and associated metacharacter notations) that will allow a
8ccb1477 290regexp to represent not just a single character sequence, but a I<whole
47f9c88b
GS
291class> of them.
292
7638d2dc 293One such concept is that of a I<character class>. A character class
47f9c88b
GS
294allows a set of possible characters, rather than just a single
295character, to match at a particular point in a regexp. Character
296classes are denoted by brackets C<[...]>, with the set of characters
297to be possibly matched inside. Here are some examples:
298
299 /cat/; # matches 'cat'
300 /[bcr]at/; # matches 'bat, 'cat', or 'rat'
301 /item[0123456789]/; # matches 'item0' or ... or 'item9'
a6b2f353 302 "abc" =~ /[cab]/; # matches 'a'
47f9c88b
GS
303
304In the last statement, even though C<'c'> is the first character in
305the class, C<'a'> matches because the first character position in the
306string is the earliest point at which the regexp can match.
307
308 /[yY][eE][sS]/; # match 'yes' in a case-insensitive way
309 # 'yes', 'Yes', 'YES', etc.
310
da75cd15 311This regexp displays a common task: perform a case-insensitive
28c3722c 312match. Perl provides a way of avoiding all those brackets by simply
47f9c88b
GS
313appending an C<'i'> to the end of the match. Then C</[yY][eE][sS]/;>
314can be rewritten as C</yes/i;>. The C<'i'> stands for
7638d2dc 315case-insensitive and is an example of a I<modifier> of the matching
47f9c88b
GS
316operation. We will meet other modifiers later in the tutorial.
317
318We saw in the section above that there were ordinary characters, which
319represented themselves, and special characters, which needed a
320backslash C<\> to represent themselves. The same is true in a
321character class, but the sets of ordinary and special characters
322inside a character class are different than those outside a character
7638d2dc 323class. The special characters for a character class are C<-]\^$> (and
353c6505 324the pattern delimiter, whatever it is).
7638d2dc 325C<]> is special because it denotes the end of a character class. C<$> is
47f9c88b
GS
326special because it denotes a scalar variable. C<\> is special because
327it is used in escape sequences, just like above. Here is how the
328special characters C<]$\> are handled:
329
330 /[\]c]def/; # matches ']def' or 'cdef'
331 $x = 'bcr';
a6b2f353 332 /[$x]at/; # matches 'bat', 'cat', or 'rat'
47f9c88b
GS
333 /[\$x]at/; # matches '$at' or 'xat'
334 /[\\$x]at/; # matches '\at', 'bat, 'cat', or 'rat'
335
353c6505 336The last two are a little tricky. In C<[\$x]>, the backslash protects
47f9c88b
GS
337the dollar sign, so the character class has two members C<$> and C<x>.
338In C<[\\$x]>, the backslash is protected, so C<$x> is treated as a
339variable and substituted in double quote fashion.
340
341The special character C<'-'> acts as a range operator within character
342classes, so that a contiguous set of characters can be written as a
343range. With ranges, the unwieldy C<[0123456789]> and C<[abc...xyz]>
344become the svelte C<[0-9]> and C<[a-z]>. Some examples are
345
346 /item[0-9]/; # matches 'item0' or ... or 'item9'
347 /[0-9bx-z]aa/; # matches '0aa', ..., '9aa',
348 # 'baa', 'xaa', 'yaa', or 'zaa'
349 /[0-9a-fA-F]/; # matches a hexadecimal digit
36bbe248 350 /[0-9a-zA-Z_]/; # matches a "word" character,
7638d2dc 351 # like those in a Perl variable name
47f9c88b
GS
352
353If C<'-'> is the first or last character in a character class, it is
354treated as an ordinary character; C<[-ab]>, C<[ab-]> and C<[a\-b]> are
355all equivalent.
356
357The special character C<^> in the first position of a character class
7638d2dc 358denotes a I<negated character class>, which matches any character but
a6b2f353 359those in the brackets. Both C<[...]> and C<[^...]> must match a
47f9c88b
GS
360character, or the match fails. Then
361
362 /[^a]at/; # doesn't match 'aat' or 'at', but matches
363 # all other 'bat', 'cat, '0at', '%at', etc.
364 /[^0-9]/; # matches a non-numeric character
365 /[a^]at/; # matches 'aat' or '^at'; here '^' is ordinary
366
28c3722c 367Now, even C<[0-9]> can be a bother to write multiple times, so in the
47f9c88b 368interest of saving keystrokes and making regexps more readable, Perl
7638d2dc 369has several abbreviations for common character classes, as shown below.
0bd5a82d
KW
370Since the introduction of Unicode, unless the C<//a> modifier is in
371effect, these character classes match more than just a few characters in
372the ASCII range.
47f9c88b
GS
373
374=over 4
375
376=item *
551e1d92 377
7638d2dc 378\d matches a digit, not just [0-9] but also digits from non-roman scripts
47f9c88b
GS
379
380=item *
551e1d92 381
7638d2dc 382\s matches a whitespace character, the set [\ \t\r\n\f] and others
47f9c88b
GS
383
384=item *
551e1d92 385
7638d2dc
WL
386\w matches a word character (alphanumeric or _), not just [0-9a-zA-Z_]
387but also digits and characters from non-roman scripts
47f9c88b
GS
388
389=item *
551e1d92 390
7638d2dc 391\D is a negated \d; it represents any other character than a digit, or [^\d]
47f9c88b
GS
392
393=item *
551e1d92 394
47f9c88b
GS
395\S is a negated \s; it represents any non-whitespace character [^\s]
396
397=item *
551e1d92 398
47f9c88b
GS
399\W is a negated \w; it represents any non-word character [^\w]
400
401=item *
551e1d92 402
7638d2dc
WL
403The period '.' matches any character but "\n" (unless the modifier C<//s> is
404in effect, as explained below).
47f9c88b 405
1ca4ba9b
KW
406=item *
407
408\N, like the period, matches any character but "\n", but it does so
409regardless of whether the modifier C<//s> is in effect.
410
47f9c88b
GS
411=back
412
0bd5a82d
KW
413The C<//a> modifier, available starting in Perl 5.14, is used to
414restrict the matches of \d, \s, and \w to just those in the ASCII range.
415It is useful to keep your program from being needlessly exposed to full
416Unicode (and its accompanying security considerations) when all you want
417is to process English-like text. (The "a" may be doubled, C<//aa>, to
418provide even more restrictions, preventing case-insensitive matching of
419ASCII with non-ASCII characters; otherwise a Unicode "Kelvin Sign"
420would caselessly match a "k" or "K".)
421
47f9c88b
GS
422The C<\d\s\w\D\S\W> abbreviations can be used both inside and outside
423of character classes. Here are some in use:
424
425 /\d\d:\d\d:\d\d/; # matches a hh:mm:ss time format
426 /[\d\s]/; # matches any digit or whitespace character
427 /\w\W\w/; # matches a word char, followed by a
428 # non-word char, followed by a word char
429 /..rt/; # matches any two chars, followed by 'rt'
430 /end\./; # matches 'end.'
431 /end[.]/; # same thing, matches 'end.'
432
433Because a period is a metacharacter, it needs to be escaped to match
434as an ordinary period. Because, for example, C<\d> and C<\w> are sets
435of characters, it is incorrect to think of C<[^\d\w]> as C<[\D\W]>; in
436fact C<[^\d\w]> is the same as C<[^\w]>, which is the same as
437C<[\W]>. Think DeMorgan's laws.
438
7638d2dc 439An anchor useful in basic regexps is the I<word anchor>
47f9c88b
GS
440C<\b>. This matches a boundary between a word character and a non-word
441character C<\w\W> or C<\W\w>:
442
443 $x = "Housecat catenates house and cat";
444 $x =~ /cat/; # matches cat in 'housecat'
445 $x =~ /\bcat/; # matches cat in 'catenates'
446 $x =~ /cat\b/; # matches cat in 'housecat'
447 $x =~ /\bcat\b/; # matches 'cat' at end of string
448
449Note in the last example, the end of the string is considered a word
450boundary.
451
452You might wonder why C<'.'> matches everything but C<"\n"> - why not
453every character? The reason is that often one is matching against
454lines and would like to ignore the newline characters. For instance,
455while the string C<"\n"> represents one line, we would like to think
28c3722c 456of it as empty. Then
47f9c88b
GS
457
458 "" =~ /^$/; # matches
7638d2dc 459 "\n" =~ /^$/; # matches, $ anchors before "\n"
47f9c88b
GS
460
461 "" =~ /./; # doesn't match; it needs a char
462 "" =~ /^.$/; # doesn't match; it needs a char
463 "\n" =~ /^.$/; # doesn't match; it needs a char other than "\n"
464 "a" =~ /^.$/; # matches
7638d2dc 465 "a\n" =~ /^.$/; # matches, $ anchors before "\n"
47f9c88b
GS
466
467This behavior is convenient, because we usually want to ignore
468newlines when we count and match characters in a line. Sometimes,
469however, we want to keep track of newlines. We might even want C<^>
470and C<$> to anchor at the beginning and end of lines within the
471string, rather than just the beginning and end of the string. Perl
472allows us to choose between ignoring and paying attention to newlines
473by using the C<//s> and C<//m> modifiers. C<//s> and C<//m> stand for
474single line and multi-line and they determine whether a string is to
475be treated as one continuous string, or as a set of lines. The two
476modifiers affect two aspects of how the regexp is interpreted: 1) how
477the C<'.'> character class is defined, and 2) where the anchors C<^>
478and C<$> are able to match. Here are the four possible combinations:
479
480=over 4
481
482=item *
551e1d92 483
47f9c88b
GS
484no modifiers (//): Default behavior. C<'.'> matches any character
485except C<"\n">. C<^> matches only at the beginning of the string and
486C<$> matches only at the end or before a newline at the end.
487
488=item *
551e1d92 489
47f9c88b
GS
490s modifier (//s): Treat string as a single long line. C<'.'> matches
491any character, even C<"\n">. C<^> matches only at the beginning of
492the string and C<$> matches only at the end or before a newline at the
493end.
494
495=item *
551e1d92 496
47f9c88b
GS
497m modifier (//m): Treat string as a set of multiple lines. C<'.'>
498matches any character except C<"\n">. C<^> and C<$> are able to match
499at the start or end of I<any> line within the string.
500
501=item *
551e1d92 502
47f9c88b
GS
503both s and m modifiers (//sm): Treat string as a single long line, but
504detect multiple lines. C<'.'> matches any character, even
505C<"\n">. C<^> and C<$>, however, are able to match at the start or end
506of I<any> line within the string.
507
508=back
509
510Here are examples of C<//s> and C<//m> in action:
511
512 $x = "There once was a girl\nWho programmed in Perl\n";
513
514 $x =~ /^Who/; # doesn't match, "Who" not at start of string
515 $x =~ /^Who/s; # doesn't match, "Who" not at start of string
516 $x =~ /^Who/m; # matches, "Who" at start of second line
517 $x =~ /^Who/sm; # matches, "Who" at start of second line
518
519 $x =~ /girl.Who/; # doesn't match, "." doesn't match "\n"
520 $x =~ /girl.Who/s; # matches, "." matches "\n"
521 $x =~ /girl.Who/m; # doesn't match, "." doesn't match "\n"
522 $x =~ /girl.Who/sm; # matches, "." matches "\n"
523
3c12f9b9 524Most of the time, the default behavior is what is wanted, but C<//s> and
47f9c88b 525C<//m> are occasionally very useful. If C<//m> is being used, the start
28c3722c 526of the string can still be matched with C<\A> and the end of the string
47f9c88b
GS
527can still be matched with the anchors C<\Z> (matches both the end and
528the newline before, like C<$>), and C<\z> (matches only the end):
529
530 $x =~ /^Who/m; # matches, "Who" at start of second line
531 $x =~ /\AWho/m; # doesn't match, "Who" is not at start of string
532
533 $x =~ /girl$/m; # matches, "girl" at end of first line
534 $x =~ /girl\Z/m; # doesn't match, "girl" is not at end of string
535
536 $x =~ /Perl\Z/m; # matches, "Perl" is at newline before end
537 $x =~ /Perl\z/m; # doesn't match, "Perl" is not at end of string
538
539We now know how to create choices among classes of characters in a
540regexp. What about choices among words or character strings? Such
541choices are described in the next section.
542
543=head2 Matching this or that
544
28c3722c 545Sometimes we would like our regexp to be able to match different
47f9c88b 546possible words or character strings. This is accomplished by using
7638d2dc
WL
547the I<alternation> metacharacter C<|>. To match C<dog> or C<cat>, we
548form the regexp C<dog|cat>. As before, Perl will try to match the
47f9c88b 549regexp at the earliest possible point in the string. At each
7638d2dc
WL
550character position, Perl will first try to match the first
551alternative, C<dog>. If C<dog> doesn't match, Perl will then try the
47f9c88b 552next alternative, C<cat>. If C<cat> doesn't match either, then the
7638d2dc 553match fails and Perl moves to the next position in the string. Some
47f9c88b
GS
554examples:
555
556 "cats and dogs" =~ /cat|dog|bird/; # matches "cat"
557 "cats and dogs" =~ /dog|cat|bird/; # matches "cat"
558
559Even though C<dog> is the first alternative in the second regexp,
560C<cat> is able to match earlier in the string.
561
562 "cats" =~ /c|ca|cat|cats/; # matches "c"
563 "cats" =~ /cats|cat|ca|c/; # matches "cats"
564
565Here, all the alternatives match at the first string position, so the
566first alternative is the one that matches. If some of the
567alternatives are truncations of the others, put the longest ones first
568to give them a chance to match.
569
570 "cab" =~ /a|b|c/ # matches "c"
571 # /a|b|c/ == /[abc]/
572
573The last example points out that character classes are like
574alternations of characters. At a given character position, the first
210b36aa 575alternative that allows the regexp match to succeed will be the one
47f9c88b
GS
576that matches.
577
578=head2 Grouping things and hierarchical matching
579
580Alternation allows a regexp to choose among alternatives, but by
7638d2dc 581itself it is unsatisfying. The reason is that each alternative is a whole
47f9c88b
GS
582regexp, but sometime we want alternatives for just part of a
583regexp. For instance, suppose we want to search for housecats or
584housekeepers. The regexp C<housecat|housekeeper> fits the bill, but is
585inefficient because we had to type C<house> twice. It would be nice to
da75cd15 586have parts of the regexp be constant, like C<house>, and some
47f9c88b
GS
587parts have alternatives, like C<cat|keeper>.
588
7638d2dc 589The I<grouping> metacharacters C<()> solve this problem. Grouping
47f9c88b
GS
590allows parts of a regexp to be treated as a single unit. Parts of a
591regexp are grouped by enclosing them in parentheses. Thus we could solve
592the C<housecat|housekeeper> by forming the regexp as
593C<house(cat|keeper)>. The regexp C<house(cat|keeper)> means match
594C<house> followed by either C<cat> or C<keeper>. Some more examples
595are
596
597 /(a|b)b/; # matches 'ab' or 'bb'
598 /(ac|b)b/; # matches 'acb' or 'bb'
599 /(^a|b)c/; # matches 'ac' at start of string or 'bc' anywhere
600 /(a|[bc])d/; # matches 'ad', 'bd', or 'cd'
601
602 /house(cat|)/; # matches either 'housecat' or 'house'
603 /house(cat(s|)|)/; # matches either 'housecats' or 'housecat' or
604 # 'house'. Note groups can be nested.
605
606 /(19|20|)\d\d/; # match years 19xx, 20xx, or the Y2K problem, xx
607 "20" =~ /(19|20|)\d\d/; # matches the null alternative '()\d\d',
608 # because '20\d\d' can't match
609
610Alternations behave the same way in groups as out of them: at a given
611string position, the leftmost alternative that allows the regexp to
210b36aa 612match is taken. So in the last example at the first string position,
47f9c88b 613C<"20"> matches the second alternative, but there is nothing left over
7638d2dc 614to match the next two digits C<\d\d>. So Perl moves on to the next
47f9c88b
GS
615alternative, which is the null alternative and that works, since
616C<"20"> is two digits.
617
618The process of trying one alternative, seeing if it matches, and
7638d2dc
WL
619moving on to the next alternative, while going back in the string
620from where the previous alternative was tried, if it doesn't, is called
621I<backtracking>. The term 'backtracking' comes from the idea that
47f9c88b
GS
622matching a regexp is like a walk in the woods. Successfully matching
623a regexp is like arriving at a destination. There are many possible
624trailheads, one for each string position, and each one is tried in
625order, left to right. From each trailhead there may be many paths,
626some of which get you there, and some which are dead ends. When you
627walk along a trail and hit a dead end, you have to backtrack along the
628trail to an earlier point to try another trail. If you hit your
629destination, you stop immediately and forget about trying all the
630other trails. You are persistent, and only if you have tried all the
631trails from all the trailheads and not arrived at your destination, do
632you declare failure. To be concrete, here is a step-by-step analysis
7638d2dc 633of what Perl does when it tries to match the regexp
47f9c88b
GS
634
635 "abcde" =~ /(abd|abc)(df|d|de)/;
636
637=over 4
638
551e1d92
RB
639=item 0
640
641Start with the first letter in the string 'a'.
642
643=item 1
47f9c88b 644
551e1d92 645Try the first alternative in the first group 'abd'.
47f9c88b 646
551e1d92 647=item 2
47f9c88b 648
551e1d92
RB
649Match 'a' followed by 'b'. So far so good.
650
651=item 3
652
653'd' in the regexp doesn't match 'c' in the string - a dead
47f9c88b
GS
654end. So backtrack two characters and pick the second alternative in
655the first group 'abc'.
656
551e1d92
RB
657=item 4
658
659Match 'a' followed by 'b' followed by 'c'. We are on a roll
47f9c88b
GS
660and have satisfied the first group. Set $1 to 'abc'.
661
551e1d92
RB
662=item 5
663
664Move on to the second group and pick the first alternative
47f9c88b
GS
665'df'.
666
551e1d92 667=item 6
47f9c88b 668
551e1d92
RB
669Match the 'd'.
670
671=item 7
672
673'f' in the regexp doesn't match 'e' in the string, so a dead
47f9c88b
GS
674end. Backtrack one character and pick the second alternative in the
675second group 'd'.
676
551e1d92
RB
677=item 8
678
679'd' matches. The second grouping is satisfied, so set $2 to
47f9c88b
GS
680'd'.
681
551e1d92
RB
682=item 9
683
684We are at the end of the regexp, so we are done! We have
47f9c88b
GS
685matched 'abcd' out of the string "abcde".
686
687=back
688
689There are a couple of things to note about this analysis. First, the
690third alternative in the second group 'de' also allows a match, but we
691stopped before we got to it - at a given character position, leftmost
692wins. Second, we were able to get a match at the first character
693position of the string 'a'. If there were no matches at the first
7638d2dc 694position, Perl would move to the second character position 'b' and
47f9c88b 695attempt the match all over again. Only when all possible paths at all
7638d2dc
WL
696possible character positions have been exhausted does Perl give
697up and declare S<C<$string =~ /(abd|abc)(df|d|de)/;>> to be false.
47f9c88b
GS
698
699Even with all this work, regexp matching happens remarkably fast. To
353c6505
DL
700speed things up, Perl compiles the regexp into a compact sequence of
701opcodes that can often fit inside a processor cache. When the code is
7638d2dc
WL
702executed, these opcodes can then run at full throttle and search very
703quickly.
47f9c88b
GS
704
705=head2 Extracting matches
706
707The grouping metacharacters C<()> also serve another completely
708different function: they allow the extraction of the parts of a string
709that matched. This is very useful to find out what matched and for
710text processing in general. For each grouping, the part that matched
711inside goes into the special variables C<$1>, C<$2>, etc. They can be
712used just as ordinary variables:
713
714 # extract hours, minutes, seconds
2275acdc
RGS
715 if ($time =~ /(\d\d):(\d\d):(\d\d)/) { # match hh:mm:ss format
716 $hours = $1;
717 $minutes = $2;
718 $seconds = $3;
719 }
47f9c88b
GS
720
721Now, we know that in scalar context,
7638d2dc 722S<C<$time =~ /(\d\d):(\d\d):(\d\d)/>> returns a true or false
47f9c88b
GS
723value. In list context, however, it returns the list of matched values
724C<($1,$2,$3)>. So we could write the code more compactly as
725
726 # extract hours, minutes, seconds
727 ($hours, $minutes, $second) = ($time =~ /(\d\d):(\d\d):(\d\d)/);
728
729If the groupings in a regexp are nested, C<$1> gets the group with the
730leftmost opening parenthesis, C<$2> the next opening parenthesis,
7638d2dc 731etc. Here is a regexp with nested groups:
47f9c88b
GS
732
733 /(ab(cd|ef)((gi)|j))/;
734 1 2 34
735
7638d2dc
WL
736If this regexp matches, C<$1> contains a string starting with
737C<'ab'>, C<$2> is either set to C<'cd'> or C<'ef'>, C<$3> equals either
738C<'gi'> or C<'j'>, and C<$4> is either set to C<'gi'>, just like C<$3>,
739or it remains undefined.
740
741For convenience, Perl sets C<$+> to the string held by the highest numbered
742C<$1>, C<$2>,... that got assigned (and, somewhat related, C<$^N> to the
743value of the C<$1>, C<$2>,... most-recently assigned; i.e. the C<$1>,
744C<$2>,... associated with the rightmost closing parenthesis used in the
a01268b5 745match).
47f9c88b 746
7638d2dc
WL
747
748=head2 Backreferences
749
47f9c88b 750Closely associated with the matching variables C<$1>, C<$2>, ... are
d8b950dc 751the I<backreferences> C<\g1>, C<\g2>,... Backreferences are simply
47f9c88b 752matching variables that can be used I<inside> a regexp. This is a
ac036724 753really nice feature; what matches later in a regexp is made to depend on
47f9c88b 754what matched earlier in the regexp. Suppose we wanted to look
7638d2dc 755for doubled words in a text, like 'the the'. The following regexp finds
47f9c88b
GS
756all 3-letter doubles with a space in between:
757
d8b950dc 758 /\b(\w\w\w)\s\g1\b/;
47f9c88b 759
8ccb1477 760The grouping assigns a value to \g1, so that the same 3-letter sequence
7638d2dc
WL
761is used for both parts.
762
763A similar task is to find words consisting of two identical parts:
47f9c88b 764
d8b950dc 765 % simple_grep '^(\w\w\w\w|\w\w\w|\w\w|\w)\g1$' /usr/dict/words
47f9c88b
GS
766 beriberi
767 booboo
768 coco
769 mama
770 murmur
771 papa
772
773The regexp has a single grouping which considers 4-letter
d8b950dc
KW
774combinations, then 3-letter combinations, etc., and uses C<\g1> to look for
775a repeat. Although C<$1> and C<\g1> represent the same thing, care should be
7638d2dc 776taken to use matched variables C<$1>, C<$2>,... only I<outside> a regexp
d8b950dc 777and backreferences C<\g1>, C<\g2>,... only I<inside> a regexp; not doing
7638d2dc
WL
778so may lead to surprising and unsatisfactory results.
779
780
781=head2 Relative backreferences
782
783Counting the opening parentheses to get the correct number for a
7698aede 784backreference is error-prone as soon as there is more than one
7638d2dc
WL
785capturing group. A more convenient technique became available
786with Perl 5.10: relative backreferences. To refer to the immediately
787preceding capture group one now may write C<\g{-1}>, the next but
788last is available via C<\g{-2}>, and so on.
789
790Another good reason in addition to readability and maintainability
8ccb1477 791for using relative backreferences is illustrated by the following example,
7638d2dc
WL
792where a simple pattern for matching peculiar strings is used:
793
d8b950dc 794 $a99a = '([a-z])(\d)\g2\g1'; # matches a11a, g22g, x33x, etc.
7638d2dc
WL
795
796Now that we have this pattern stored as a handy string, we might feel
797tempted to use it as a part of some other pattern:
798
799 $line = "code=e99e";
800 if ($line =~ /^(\w+)=$a99a$/){ # unexpected behavior!
801 print "$1 is valid\n";
802 } else {
803 print "bad line: '$line'\n";
804 }
805
ac036724 806But this doesn't match, at least not the way one might expect. Only
7638d2dc
WL
807after inserting the interpolated C<$a99a> and looking at the resulting
808full text of the regexp is it obvious that the backreferences have
ac036724 809backfired. The subexpression C<(\w+)> has snatched number 1 and
7638d2dc
WL
810demoted the groups in C<$a99a> by one rank. This can be avoided by
811using relative backreferences:
812
813 $a99a = '([a-z])(\d)\g{-1}\g{-2}'; # safe for being interpolated
814
815
816=head2 Named backreferences
817
c27a5cfe 818Perl 5.10 also introduced named capture groups and named backreferences.
7638d2dc
WL
819To attach a name to a capturing group, you write either
820C<< (?<name>...) >> or C<< (?'name'...) >>. The backreference may
821then be written as C<\g{name}>. It is permissible to attach the
822same name to more than one group, but then only the leftmost one of the
823eponymous set can be referenced. Outside of the pattern a named
c27a5cfe 824capture group is accessible through the C<%+> hash.
7638d2dc 825
353c6505 826Assuming that we have to match calendar dates which may be given in one
7638d2dc 827of the three formats yyyy-mm-dd, mm/dd/yyyy or dd.mm.yyyy, we can write
353c6505 828three suitable patterns where we use 'd', 'm' and 'y' respectively as the
c27a5cfe 829names of the groups capturing the pertaining components of a date. The
7638d2dc
WL
830matching operation combines the three patterns as alternatives:
831
832 $fmt1 = '(?<y>\d\d\d\d)-(?<m>\d\d)-(?<d>\d\d)';
833 $fmt2 = '(?<m>\d\d)/(?<d>\d\d)/(?<y>\d\d\d\d)';
834 $fmt3 = '(?<d>\d\d)\.(?<m>\d\d)\.(?<y>\d\d\d\d)';
835 for my $d qw( 2006-10-21 15.01.2007 10/31/2005 ){
836 if ( $d =~ m{$fmt1|$fmt2|$fmt3} ){
837 print "day=$+{d} month=$+{m} year=$+{y}\n";
838 }
839 }
840
841If any of the alternatives matches, the hash C<%+> is bound to contain the
842three key-value pairs.
843
844
845=head2 Alternative capture group numbering
846
847Yet another capturing group numbering technique (also as from Perl 5.10)
848deals with the problem of referring to groups within a set of alternatives.
849Consider a pattern for matching a time of the day, civil or military style:
47f9c88b 850
7638d2dc
WL
851 if ( $time =~ /(\d\d|\d):(\d\d)|(\d\d)(\d\d)/ ){
852 # process hour and minute
853 }
854
855Processing the results requires an additional if statement to determine
353c6505 856whether C<$1> and C<$2> or C<$3> and C<$4> contain the goodies. It would
c27a5cfe 857be easier if we could use group numbers 1 and 2 in second alternative as
353c6505 858well, and this is exactly what the parenthesized construct C<(?|...)>,
7638d2dc
WL
859set around an alternative achieves. Here is an extended version of the
860previous pattern:
861
862 if ( $time =~ /(?|(\d\d|\d):(\d\d)|(\d\d)(\d\d))\s+([A-Z][A-Z][A-Z])/ ){
863 print "hour=$1 minute=$2 zone=$3\n";
864 }
865
c27a5cfe 866Within the alternative numbering group, group numbers start at the same
7638d2dc 867position for each alternative. After the group, numbering continues
353c6505 868with one higher than the maximum reached across all the alternatives.
7638d2dc
WL
869
870=head2 Position information
871
13e5d9cd 872In addition to what was matched, Perl also provides the
7638d2dc 873positions of what was matched as contents of the C<@-> and C<@+>
47f9c88b
GS
874arrays. C<$-[0]> is the position of the start of the entire match and
875C<$+[0]> is the position of the end. Similarly, C<$-[n]> is the
876position of the start of the C<$n> match and C<$+[n]> is the position
877of the end. If C<$n> is undefined, so are C<$-[n]> and C<$+[n]>. Then
878this code
879
880 $x = "Mmm...donut, thought Homer";
881 $x =~ /^(Mmm|Yech)\.\.\.(donut|peas)/; # matches
882 foreach $expr (1..$#-) {
883 print "Match $expr: '${$expr}' at position ($-[$expr],$+[$expr])\n";
884 }
885
886prints
887
888 Match 1: 'Mmm' at position (0,3)
889 Match 2: 'donut' at position (6,11)
890
891Even if there are no groupings in a regexp, it is still possible to
7638d2dc 892find out what exactly matched in a string. If you use them, Perl
47f9c88b
GS
893will set C<$`> to the part of the string before the match, will set C<$&>
894to the part of the string that matched, and will set C<$'> to the part
895of the string after the match. An example:
896
897 $x = "the cat caught the mouse";
898 $x =~ /cat/; # $` = 'the ', $& = 'cat', $' = ' caught the mouse'
899 $x =~ /the/; # $` = '', $& = 'the', $' = ' cat caught the mouse'
900
7638d2dc
WL
901In the second match, C<$`> equals C<''> because the regexp matched at the
902first character position in the string and stopped; it never saw the
47f9c88b 903second 'the'. It is important to note that using C<$`> and C<$'>
7638d2dc 904slows down regexp matching quite a bit, while C<$&> slows it down to a
47f9c88b 905lesser extent, because if they are used in one regexp in a program,
7638d2dc 906they are generated for I<all> regexps in the program. So if raw
47f9c88b 907performance is a goal of your application, they should be avoided.
7638d2dc
WL
908If you need to extract the corresponding substrings, use C<@-> and
909C<@+> instead:
47f9c88b
GS
910
911 $` is the same as substr( $x, 0, $-[0] )
912 $& is the same as substr( $x, $-[0], $+[0]-$-[0] )
913 $' is the same as substr( $x, $+[0] )
914
78622607
FC
915As of Perl 5.10, the C<${^PREMATCH}>, C<${^MATCH}> and C<${^POSTMATCH}>
916variables may be used. These are only set if the C</p> modifier is present.
917Consequently they do not penalize the rest of the program.
7638d2dc
WL
918
919=head2 Non-capturing groupings
920
353c6505 921A group that is required to bundle a set of alternatives may or may not be
7638d2dc 922useful as a capturing group. If it isn't, it just creates a superfluous
c27a5cfe 923addition to the set of available capture group values, inside as well as
7638d2dc 924outside the regexp. Non-capturing groupings, denoted by C<(?:regexp)>,
353c6505 925still allow the regexp to be treated as a single unit, but don't establish
c27a5cfe 926a capturing group at the same time. Both capturing and non-capturing
7638d2dc
WL
927groupings are allowed to co-exist in the same regexp. Because there is
928no extraction, non-capturing groupings are faster than capturing
929groupings. Non-capturing groupings are also handy for choosing exactly
930which parts of a regexp are to be extracted to matching variables:
931
932 # match a number, $1-$4 are set, but we only want $1
933 /([+-]?\ *(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?)/;
934
935 # match a number faster , only $1 is set
936 /([+-]?\ *(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][+-]?\d+)?)/;
937
938 # match a number, get $1 = whole number, $2 = exponent
939 /([+-]?\ *(?:\d+(?:\.\d*)?|\.\d+)(?:[eE]([+-]?\d+))?)/;
940
941Non-capturing groupings are also useful for removing nuisance
942elements gathered from a split operation where parentheses are
943required for some reason:
944
945 $x = '12aba34ba5';
9b846e30 946 @num = split /(a|b)+/, $x; # @num = ('12','a','34','a','5')
7638d2dc
WL
947 @num = split /(?:a|b)+/, $x; # @num = ('12','34','5')
948
949
47f9c88b
GS
950=head2 Matching repetitions
951
952The examples in the previous section display an annoying weakness. We
7638d2dc
WL
953were only matching 3-letter words, or chunks of words of 4 letters or
954less. We'd like to be able to match words or, more generally, strings
955of any length, without writing out tedious alternatives like
47f9c88b
GS
956C<\w\w\w\w|\w\w\w|\w\w|\w>.
957
7638d2dc
WL
958This is exactly the problem the I<quantifier> metacharacters C<?>,
959C<*>, C<+>, and C<{}> were created for. They allow us to delimit the
960number of repeats for a portion of a regexp we consider to be a
47f9c88b
GS
961match. Quantifiers are put immediately after the character, character
962class, or grouping that we want to specify. They have the following
963meanings:
964
965=over 4
966
551e1d92 967=item *
47f9c88b 968
7638d2dc 969C<a?> means: match 'a' 1 or 0 times
47f9c88b 970
551e1d92
RB
971=item *
972
7638d2dc 973C<a*> means: match 'a' 0 or more times, i.e., any number of times
551e1d92
RB
974
975=item *
47f9c88b 976
7638d2dc 977C<a+> means: match 'a' 1 or more times, i.e., at least once
551e1d92
RB
978
979=item *
980
7638d2dc 981C<a{n,m}> means: match at least C<n> times, but not more than C<m>
47f9c88b
GS
982times.
983
551e1d92
RB
984=item *
985
7638d2dc 986C<a{n,}> means: match at least C<n> or more times
551e1d92
RB
987
988=item *
47f9c88b 989
7638d2dc 990C<a{n}> means: match exactly C<n> times
47f9c88b
GS
991
992=back
993
994Here are some examples:
995
7638d2dc 996 /[a-z]+\s+\d*/; # match a lowercase word, at least one space, and
47f9c88b 997 # any number of digits
d8b950dc 998 /(\w+)\s+\g1/; # match doubled words of arbitrary length
47f9c88b 999 /y(es)?/i; # matches 'y', 'Y', or a case-insensitive 'yes'
c2ac8995
NS
1000 $year =~ /^\d{2,4}$/; # make sure year is at least 2 but not more
1001 # than 4 digits
f5b885cd 1002 $year =~ /^\d{4}$|^\d{2}$/; # better match; throw out 3-digit dates
c2ac8995
NS
1003 $year =~ /^\d{2}(\d{2})?$/; # same thing written differently. However,
1004 # this captures the last two digits in $1
1005 # and the other does not.
47f9c88b 1006
d8b950dc 1007 % simple_grep '^(\w+)\g1$' /usr/dict/words # isn't this easier?
47f9c88b
GS
1008 beriberi
1009 booboo
1010 coco
1011 mama
1012 murmur
1013 papa
1014
7638d2dc 1015For all of these quantifiers, Perl will try to match as much of the
47f9c88b 1016string as possible, while still allowing the regexp to succeed. Thus
7638d2dc
WL
1017with C</a?.../>, Perl will first try to match the regexp with the C<a>
1018present; if that fails, Perl will try to match the regexp without the
47f9c88b
GS
1019C<a> present. For the quantifier C<*>, we get the following:
1020
1021 $x = "the cat in the hat";
1022 $x =~ /^(.*)(cat)(.*)$/; # matches,
1023 # $1 = 'the '
1024 # $2 = 'cat'
1025 # $3 = ' in the hat'
1026
1027Which is what we might expect, the match finds the only C<cat> in the
1028string and locks onto it. Consider, however, this regexp:
1029
1030 $x =~ /^(.*)(at)(.*)$/; # matches,
1031 # $1 = 'the cat in the h'
1032 # $2 = 'at'
7638d2dc 1033 # $3 = '' (0 characters match)
47f9c88b 1034
7638d2dc 1035One might initially guess that Perl would find the C<at> in C<cat> and
47f9c88b
GS
1036stop there, but that wouldn't give the longest possible string to the
1037first quantifier C<.*>. Instead, the first quantifier C<.*> grabs as
1038much of the string as possible while still having the regexp match. In
a6b2f353 1039this example, that means having the C<at> sequence with the final C<at>
f5b885cd 1040in the string. The other important principle illustrated here is that,
47f9c88b 1041when there are two or more elements in a regexp, the I<leftmost>
f5b885cd 1042quantifier, if there is one, gets to grab as much of the string as
47f9c88b
GS
1043possible, leaving the rest of the regexp to fight over scraps. Thus in
1044our example, the first quantifier C<.*> grabs most of the string, while
1045the second quantifier C<.*> gets the empty string. Quantifiers that
7638d2dc
WL
1046grab as much of the string as possible are called I<maximal match> or
1047I<greedy> quantifiers.
47f9c88b
GS
1048
1049When a regexp can match a string in several different ways, we can use
1050the principles above to predict which way the regexp will match:
1051
1052=over 4
1053
1054=item *
551e1d92 1055
47f9c88b
GS
1056Principle 0: Taken as a whole, any regexp will be matched at the
1057earliest possible position in the string.
1058
1059=item *
551e1d92 1060
47f9c88b
GS
1061Principle 1: In an alternation C<a|b|c...>, the leftmost alternative
1062that allows a match for the whole regexp will be the one used.
1063
1064=item *
551e1d92 1065
47f9c88b
GS
1066Principle 2: The maximal matching quantifiers C<?>, C<*>, C<+> and
1067C<{n,m}> will in general match as much of the string as possible while
1068still allowing the whole regexp to match.
1069
1070=item *
551e1d92 1071
47f9c88b
GS
1072Principle 3: If there are two or more elements in a regexp, the
1073leftmost greedy quantifier, if any, will match as much of the string
1074as possible while still allowing the whole regexp to match. The next
1075leftmost greedy quantifier, if any, will try to match as much of the
1076string remaining available to it as possible, while still allowing the
1077whole regexp to match. And so on, until all the regexp elements are
1078satisfied.
1079
1080=back
1081
ac036724 1082As we have seen above, Principle 0 overrides the others. The regexp
47f9c88b
GS
1083will be matched as early as possible, with the other principles
1084determining how the regexp matches at that earliest character
1085position.
1086
1087Here is an example of these principles in action:
1088
1089 $x = "The programming republic of Perl";
1090 $x =~ /^(.+)(e|r)(.*)$/; # matches,
1091 # $1 = 'The programming republic of Pe'
1092 # $2 = 'r'
1093 # $3 = 'l'
1094
1095This regexp matches at the earliest string position, C<'T'>. One
1096might think that C<e>, being leftmost in the alternation, would be
1097matched, but C<r> produces the longest string in the first quantifier.
1098
1099 $x =~ /(m{1,2})(.*)$/; # matches,
1100 # $1 = 'mm'
1101 # $2 = 'ing republic of Perl'
1102
1103Here, The earliest possible match is at the first C<'m'> in
1104C<programming>. C<m{1,2}> is the first quantifier, so it gets to match
1105a maximal C<mm>.
1106
1107 $x =~ /.*(m{1,2})(.*)$/; # matches,
1108 # $1 = 'm'
1109 # $2 = 'ing republic of Perl'
1110
1111Here, the regexp matches at the start of the string. The first
1112quantifier C<.*> grabs as much as possible, leaving just a single
1113C<'m'> for the second quantifier C<m{1,2}>.
1114
1115 $x =~ /(.?)(m{1,2})(.*)$/; # matches,
1116 # $1 = 'a'
1117 # $2 = 'mm'
1118 # $3 = 'ing republic of Perl'
1119
1120Here, C<.?> eats its maximal one character at the earliest possible
1121position in the string, C<'a'> in C<programming>, leaving C<m{1,2}>
1122the opportunity to match both C<m>'s. Finally,
1123
1124 "aXXXb" =~ /(X*)/; # matches with $1 = ''
1125
1126because it can match zero copies of C<'X'> at the beginning of the
1127string. If you definitely want to match at least one C<'X'>, use
1128C<X+>, not C<X*>.
1129
1130Sometimes greed is not good. At times, we would like quantifiers to
1131match a I<minimal> piece of string, rather than a maximal piece. For
7638d2dc
WL
1132this purpose, Larry Wall created the I<minimal match> or
1133I<non-greedy> quantifiers C<??>, C<*?>, C<+?>, and C<{}?>. These are
47f9c88b
GS
1134the usual quantifiers with a C<?> appended to them. They have the
1135following meanings:
1136
1137=over 4
1138
551e1d92
RB
1139=item *
1140
7638d2dc 1141C<a??> means: match 'a' 0 or 1 times. Try 0 first, then 1.
47f9c88b 1142
551e1d92
RB
1143=item *
1144
7638d2dc 1145C<a*?> means: match 'a' 0 or more times, i.e., any number of times,
47f9c88b
GS
1146but as few times as possible
1147
551e1d92
RB
1148=item *
1149
7638d2dc 1150C<a+?> means: match 'a' 1 or more times, i.e., at least once, but
47f9c88b
GS
1151as few times as possible
1152
551e1d92
RB
1153=item *
1154
7638d2dc 1155C<a{n,m}?> means: match at least C<n> times, not more than C<m>
47f9c88b
GS
1156times, as few times as possible
1157
551e1d92
RB
1158=item *
1159
7638d2dc 1160C<a{n,}?> means: match at least C<n> times, but as few times as
47f9c88b
GS
1161possible
1162
551e1d92
RB
1163=item *
1164
7638d2dc 1165C<a{n}?> means: match exactly C<n> times. Because we match exactly
47f9c88b
GS
1166C<n> times, C<a{n}?> is equivalent to C<a{n}> and is just there for
1167notational consistency.
1168
1169=back
1170
1171Let's look at the example above, but with minimal quantifiers:
1172
1173 $x = "The programming republic of Perl";
1174 $x =~ /^(.+?)(e|r)(.*)$/; # matches,
1175 # $1 = 'Th'
1176 # $2 = 'e'
1177 # $3 = ' programming republic of Perl'
1178
1179The minimal string that will allow both the start of the string C<^>
1180and the alternation to match is C<Th>, with the alternation C<e|r>
1181matching C<e>. The second quantifier C<.*> is free to gobble up the
1182rest of the string.
1183
1184 $x =~ /(m{1,2}?)(.*?)$/; # matches,
1185 # $1 = 'm'
1186 # $2 = 'ming republic of Perl'
1187
1188The first string position that this regexp can match is at the first
1189C<'m'> in C<programming>. At this position, the minimal C<m{1,2}?>
1190matches just one C<'m'>. Although the second quantifier C<.*?> would
1191prefer to match no characters, it is constrained by the end-of-string
1192anchor C<$> to match the rest of the string.
1193
1194 $x =~ /(.*?)(m{1,2}?)(.*)$/; # matches,
1195 # $1 = 'The progra'
1196 # $2 = 'm'
1197 # $3 = 'ming republic of Perl'
1198
1199In this regexp, you might expect the first minimal quantifier C<.*?>
1200to match the empty string, because it is not constrained by a C<^>
1201anchor to match the beginning of the word. Principle 0 applies here,
1202however. Because it is possible for the whole regexp to match at the
1203start of the string, it I<will> match at the start of the string. Thus
1204the first quantifier has to match everything up to the first C<m>. The
1205second minimal quantifier matches just one C<m> and the third
1206quantifier matches the rest of the string.
1207
1208 $x =~ /(.??)(m{1,2})(.*)$/; # matches,
1209 # $1 = 'a'
1210 # $2 = 'mm'
1211 # $3 = 'ing republic of Perl'
1212
1213Just as in the previous regexp, the first quantifier C<.??> can match
1214earliest at position C<'a'>, so it does. The second quantifier is
1215greedy, so it matches C<mm>, and the third matches the rest of the
1216string.
1217
1218We can modify principle 3 above to take into account non-greedy
1219quantifiers:
1220
1221=over 4
1222
1223=item *
551e1d92 1224
47f9c88b
GS
1225Principle 3: If there are two or more elements in a regexp, the
1226leftmost greedy (non-greedy) quantifier, if any, will match as much
1227(little) of the string as possible while still allowing the whole
1228regexp to match. The next leftmost greedy (non-greedy) quantifier, if
1229any, will try to match as much (little) of the string remaining
1230available to it as possible, while still allowing the whole regexp to
1231match. And so on, until all the regexp elements are satisfied.
1232
1233=back
1234
1235Just like alternation, quantifiers are also susceptible to
1236backtracking. Here is a step-by-step analysis of the example
1237
1238 $x = "the cat in the hat";
1239 $x =~ /^(.*)(at)(.*)$/; # matches,
1240 # $1 = 'the cat in the h'
1241 # $2 = 'at'
1242 # $3 = '' (0 matches)
1243
1244=over 4
1245
551e1d92
RB
1246=item 0
1247
1248Start with the first letter in the string 't'.
47f9c88b 1249
551e1d92
RB
1250=item 1
1251
1252The first quantifier '.*' starts out by matching the whole
47f9c88b
GS
1253string 'the cat in the hat'.
1254
551e1d92
RB
1255=item 2
1256
1257'a' in the regexp element 'at' doesn't match the end of the
47f9c88b
GS
1258string. Backtrack one character.
1259
551e1d92
RB
1260=item 3
1261
1262'a' in the regexp element 'at' still doesn't match the last
47f9c88b
GS
1263letter of the string 't', so backtrack one more character.
1264
551e1d92
RB
1265=item 4
1266
1267Now we can match the 'a' and the 't'.
47f9c88b 1268
551e1d92
RB
1269=item 5
1270
1271Move on to the third element '.*'. Since we are at the end of
47f9c88b
GS
1272the string and '.*' can match 0 times, assign it the empty string.
1273
551e1d92
RB
1274=item 6
1275
1276We are done!
47f9c88b
GS
1277
1278=back
1279
1280Most of the time, all this moving forward and backtracking happens
7638d2dc 1281quickly and searching is fast. There are some pathological regexps,
47f9c88b
GS
1282however, whose execution time exponentially grows with the size of the
1283string. A typical structure that blows up in your face is of the form
1284
1285 /(a|b+)*/;
1286
1287The problem is the nested indeterminate quantifiers. There are many
1288different ways of partitioning a string of length n between the C<+>
1289and C<*>: one repetition with C<b+> of length n, two repetitions with
1290the first C<b+> length k and the second with length n-k, m repetitions
1291whose bits add up to length n, etc. In fact there are an exponential
7638d2dc 1292number of ways to partition a string as a function of its length. A
47f9c88b 1293regexp may get lucky and match early in the process, but if there is
7638d2dc 1294no match, Perl will try I<every> possibility before giving up. So be
47f9c88b 1295careful with nested C<*>'s, C<{n,m}>'s, and C<+>'s. The book
7638d2dc 1296I<Mastering Regular Expressions> by Jeffrey Friedl gives a wonderful
47f9c88b
GS
1297discussion of this and other efficiency issues.
1298
7638d2dc
WL
1299
1300=head2 Possessive quantifiers
1301
1302Backtracking during the relentless search for a match may be a waste
1303of time, particularly when the match is bound to fail. Consider
1304the simple pattern
1305
1306 /^\w+\s+\w+$/; # a word, spaces, a word
1307
1308Whenever this is applied to a string which doesn't quite meet the
1309pattern's expectations such as S<C<"abc ">> or S<C<"abc def ">>,
353c6505
DL
1310the regex engine will backtrack, approximately once for each character
1311in the string. But we know that there is no way around taking I<all>
1312of the initial word characters to match the first repetition, that I<all>
7638d2dc 1313spaces must be eaten by the middle part, and the same goes for the second
353c6505
DL
1314word.
1315
1316With the introduction of the I<possessive quantifiers> in Perl 5.10, we
1317have a way of instructing the regex engine not to backtrack, with the
1318usual quantifiers with a C<+> appended to them. This makes them greedy as
1319well as stingy; once they succeed they won't give anything back to permit
1320another solution. They have the following meanings:
7638d2dc
WL
1321
1322=over 4
1323
1324=item *
1325
353c6505
DL
1326C<a{n,m}+> means: match at least C<n> times, not more than C<m> times,
1327as many times as possible, and don't give anything up. C<a?+> is short
7638d2dc
WL
1328for C<a{0,1}+>
1329
1330=item *
1331
1332C<a{n,}+> means: match at least C<n> times, but as many times as possible,
353c6505 1333and don't give anything up. C<a*+> is short for C<a{0,}+> and C<a++> is
7638d2dc
WL
1334short for C<a{1,}+>.
1335
1336=item *
1337
1338C<a{n}+> means: match exactly C<n> times. It is just there for
1339notational consistency.
1340
1341=back
1342
353c6505
DL
1343These possessive quantifiers represent a special case of a more general
1344concept, the I<independent subexpression>, see below.
7638d2dc
WL
1345
1346As an example where a possessive quantifier is suitable we consider
1347matching a quoted string, as it appears in several programming languages.
1348The backslash is used as an escape character that indicates that the
1349next character is to be taken literally, as another character for the
1350string. Therefore, after the opening quote, we expect a (possibly
353c6505 1351empty) sequence of alternatives: either some character except an
7638d2dc
WL
1352unescaped quote or backslash or an escaped character.
1353
1354 /"(?:[^"\\]++|\\.)*+"/;
1355
1356
47f9c88b
GS
1357=head2 Building a regexp
1358
1359At this point, we have all the basic regexp concepts covered, so let's
1360give a more involved example of a regular expression. We will build a
1361regexp that matches numbers.
1362
1363The first task in building a regexp is to decide what we want to match
1364and what we want to exclude. In our case, we want to match both
1365integers and floating point numbers and we want to reject any string
1366that isn't a number.
1367
1368The next task is to break the problem down into smaller problems that
1369are easily converted into a regexp.
1370
1371The simplest case is integers. These consist of a sequence of digits,
1372with an optional sign in front. The digits we can represent with
1373C<\d+> and the sign can be matched with C<[+-]>. Thus the integer
1374regexp is
1375
1376 /[+-]?\d+/; # matches integers
1377
1378A floating point number potentially has a sign, an integral part, a
1379decimal point, a fractional part, and an exponent. One or more of these
1380parts is optional, so we need to check out the different
1381possibilities. Floating point numbers which are in proper form include
1382123., 0.345, .34, -1e6, and 25.4E-72. As with integers, the sign out
1383front is completely optional and can be matched by C<[+-]?>. We can
1384see that if there is no exponent, floating point numbers must have a
1385decimal point, otherwise they are integers. We might be tempted to
1386model these with C<\d*\.\d*>, but this would also match just a single
1387decimal point, which is not a number. So the three cases of floating
7638d2dc 1388point number without exponent are
47f9c88b
GS
1389
1390 /[+-]?\d+\./; # 1., 321., etc.
1391 /[+-]?\.\d+/; # .1, .234, etc.
1392 /[+-]?\d+\.\d+/; # 1.0, 30.56, etc.
1393
1394These can be combined into a single regexp with a three-way alternation:
1395
1396 /[+-]?(\d+\.\d+|\d+\.|\.\d+)/; # floating point, no exponent
1397
1398In this alternation, it is important to put C<'\d+\.\d+'> before
1399C<'\d+\.'>. If C<'\d+\.'> were first, the regexp would happily match that
1400and ignore the fractional part of the number.
1401
1402Now consider floating point numbers with exponents. The key
1403observation here is that I<both> integers and numbers with decimal
1404points are allowed in front of an exponent. Then exponents, like the
1405overall sign, are independent of whether we are matching numbers with
1406or without decimal points, and can be 'decoupled' from the
1407mantissa. The overall form of the regexp now becomes clear:
1408
1409 /^(optional sign)(integer | f.p. mantissa)(optional exponent)$/;
1410
1411The exponent is an C<e> or C<E>, followed by an integer. So the
1412exponent regexp is
1413
1414 /[eE][+-]?\d+/; # exponent
1415
1416Putting all the parts together, we get a regexp that matches numbers:
1417
1418 /^[+-]?(\d+\.\d+|\d+\.|\.\d+|\d+)([eE][+-]?\d+)?$/; # Ta da!
1419
1420Long regexps like this may impress your friends, but can be hard to
1421decipher. In complex situations like this, the C<//x> modifier for a
1422match is invaluable. It allows one to put nearly arbitrary whitespace
1423and comments into a regexp without affecting their meaning. Using it,
1424we can rewrite our 'extended' regexp in the more pleasing form
1425
1426 /^
1427 [+-]? # first, match an optional sign
1428 ( # then match integers or f.p. mantissas:
1429 \d+\.\d+ # mantissa of the form a.b
1430 |\d+\. # mantissa of the form a.
1431 |\.\d+ # mantissa of the form .b
1432 |\d+ # integer of the form a
1433 )
1434 ([eE][+-]?\d+)? # finally, optionally match an exponent
1435 $/x;
1436
1437If whitespace is mostly irrelevant, how does one include space
1438characters in an extended regexp? The answer is to backslash it
7638d2dc 1439S<C<'\ '>> or put it in a character class S<C<[ ]>>. The same thing
f5b885cd 1440goes for pound signs: use C<\#> or C<[#]>. For instance, Perl allows
7638d2dc 1441a space between the sign and the mantissa or integer, and we could add
47f9c88b
GS
1442this to our regexp as follows:
1443
1444 /^
1445 [+-]?\ * # first, match an optional sign *and space*
1446 ( # then match integers or f.p. mantissas:
1447 \d+\.\d+ # mantissa of the form a.b
1448 |\d+\. # mantissa of the form a.
1449 |\.\d+ # mantissa of the form .b
1450 |\d+ # integer of the form a
1451 )
1452 ([eE][+-]?\d+)? # finally, optionally match an exponent
1453 $/x;
1454
1455In this form, it is easier to see a way to simplify the
1456alternation. Alternatives 1, 2, and 4 all start with C<\d+>, so it
1457could be factored out:
1458
1459 /^
1460 [+-]?\ * # first, match an optional sign
1461 ( # then match integers or f.p. mantissas:
1462 \d+ # start out with a ...
1463 (
1464 \.\d* # mantissa of the form a.b or a.
1465 )? # ? takes care of integers of the form a
1466 |\.\d+ # mantissa of the form .b
1467 )
1468 ([eE][+-]?\d+)? # finally, optionally match an exponent
1469 $/x;
1470
1471or written in the compact form,
1472
1473 /^[+-]?\ *(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?$/;
1474
1475This is our final regexp. To recap, we built a regexp by
1476
1477=over 4
1478
551e1d92
RB
1479=item *
1480
1481specifying the task in detail,
47f9c88b 1482
551e1d92
RB
1483=item *
1484
1485breaking down the problem into smaller parts,
1486
1487=item *
47f9c88b 1488
551e1d92 1489translating the small parts into regexps,
47f9c88b 1490
551e1d92
RB
1491=item *
1492
1493combining the regexps,
1494
1495=item *
47f9c88b 1496
551e1d92 1497and optimizing the final combined regexp.
47f9c88b
GS
1498
1499=back
1500
1501These are also the typical steps involved in writing a computer
1502program. This makes perfect sense, because regular expressions are
7638d2dc 1503essentially programs written in a little computer language that specifies
47f9c88b
GS
1504patterns.
1505
1506=head2 Using regular expressions in Perl
1507
1508The last topic of Part 1 briefly covers how regexps are used in Perl
1509programs. Where do they fit into Perl syntax?
1510
1511We have already introduced the matching operator in its default
1512C</regexp/> and arbitrary delimiter C<m!regexp!> forms. We have used
1513the binding operator C<=~> and its negation C<!~> to test for string
1514matches. Associated with the matching operator, we have discussed the
1515single line C<//s>, multi-line C<//m>, case-insensitive C<//i> and
353c6505
DL
1516extended C<//x> modifiers. There are a few more things you might
1517want to know about matching operators.
47f9c88b 1518
7638d2dc
WL
1519=head3 Prohibiting substitution
1520
1521If you change C<$pattern> after the first substitution happens, Perl
47f9c88b
GS
1522will ignore it. If you don't want any substitutions at all, use the
1523special delimiter C<m''>:
1524
16e8b840 1525 @pattern = ('Seuss');
47f9c88b 1526 while (<>) {
16e8b840 1527 print if m'@pattern'; # matches literal '@pattern', not 'Seuss'
47f9c88b
GS
1528 }
1529
353c6505 1530Similar to strings, C<m''> acts like apostrophes on a regexp; all other
7638d2dc 1531C<m> delimiters act like quotes. If the regexp evaluates to the empty string,
47f9c88b
GS
1532the regexp in the I<last successful match> is used instead. So we have
1533
1534 "dog" =~ /d/; # 'd' matches
1535 "dogbert =~ //; # this matches the 'd' regexp used before
1536
7638d2dc
WL
1537
1538=head3 Global matching
1539
7698aede 1540The final two modifiers we will discuss here,
5f67e4c9 1541C<//g> and C<//c>, concern multiple matches.
da75cd15 1542The modifier C<//g> stands for global matching and allows the
47f9c88b
GS
1543matching operator to match within a string as many times as possible.
1544In scalar context, successive invocations against a string will have
f5b885cd 1545C<//g> jump from match to match, keeping track of position in the
47f9c88b
GS
1546string as it goes along. You can get or set the position with the
1547C<pos()> function.
1548
1549The use of C<//g> is shown in the following example. Suppose we have
1550a string that consists of words separated by spaces. If we know how
1551many words there are in advance, we could extract the words using
1552groupings:
1553
1554 $x = "cat dog house"; # 3 words
1555 $x =~ /^\s*(\w+)\s+(\w+)\s+(\w+)\s*$/; # matches,
1556 # $1 = 'cat'
1557 # $2 = 'dog'
1558 # $3 = 'house'
1559
1560But what if we had an indeterminate number of words? This is the sort
1561of task C<//g> was made for. To extract all words, form the simple
1562regexp C<(\w+)> and loop over all matches with C</(\w+)/g>:
1563
1564 while ($x =~ /(\w+)/g) {
1565 print "Word is $1, ends at position ", pos $x, "\n";
1566 }
1567
1568prints
1569
1570 Word is cat, ends at position 3
1571 Word is dog, ends at position 7
1572 Word is house, ends at position 13
1573
1574A failed match or changing the target string resets the position. If
1575you don't want the position reset after failure to match, add the
1576C<//c>, as in C</regexp/gc>. The current position in the string is
1577associated with the string, not the regexp. This means that different
1578strings have different positions and their respective positions can be
1579set or read independently.
1580
1581In list context, C<//g> returns a list of matched groupings, or if
1582there are no groupings, a list of matches to the whole regexp. So if
1583we wanted just the words, we could use
1584
1585 @words = ($x =~ /(\w+)/g); # matches,
5a0c7e9d
PJ
1586 # $words[0] = 'cat'
1587 # $words[1] = 'dog'
1588 # $words[2] = 'house'
47f9c88b
GS
1589
1590Closely associated with the C<//g> modifier is the C<\G> anchor. The
1591C<\G> anchor matches at the point where the previous C<//g> match left
1592off. C<\G> allows us to easily do context-sensitive matching:
1593
1594 $metric = 1; # use metric units
1595 ...
1596 $x = <FILE>; # read in measurement
1597 $x =~ /^([+-]?\d+)\s*/g; # get magnitude
1598 $weight = $1;
1599 if ($metric) { # error checking
1600 print "Units error!" unless $x =~ /\Gkg\./g;
1601 }
1602 else {
1603 print "Units error!" unless $x =~ /\Glbs\./g;
1604 }
1605 $x =~ /\G\s+(widget|sprocket)/g; # continue processing
1606
1607The combination of C<//g> and C<\G> allows us to process the string a
1608bit at a time and use arbitrary Perl logic to decide what to do next.
25cf8c22
HS
1609Currently, the C<\G> anchor is only fully supported when used to anchor
1610to the start of the pattern.
47f9c88b 1611
f5b885cd 1612C<\G> is also invaluable in processing fixed-length records with
47f9c88b
GS
1613regexps. Suppose we have a snippet of coding region DNA, encoded as
1614base pair letters C<ATCGTTGAAT...> and we want to find all the stop
1615codons C<TGA>. In a coding region, codons are 3-letter sequences, so
1616we can think of the DNA snippet as a sequence of 3-letter records. The
1617naive regexp
1618
1619 # expanded, this is "ATC GTT GAA TGC AAA TGA CAT GAC"
1620 $dna = "ATCGTTGAATGCAAATGACATGAC";
1621 $dna =~ /TGA/;
1622
d1be9408 1623doesn't work; it may match a C<TGA>, but there is no guarantee that
47f9c88b 1624the match is aligned with codon boundaries, e.g., the substring
7638d2dc 1625S<C<GTT GAA>> gives a match. A better solution is
47f9c88b
GS
1626
1627 while ($dna =~ /(\w\w\w)*?TGA/g) { # note the minimal *?
1628 print "Got a TGA stop codon at position ", pos $dna, "\n";
1629 }
1630
1631which prints
1632
1633 Got a TGA stop codon at position 18
1634 Got a TGA stop codon at position 23
1635
1636Position 18 is good, but position 23 is bogus. What happened?
1637
1638The answer is that our regexp works well until we get past the last
1639real match. Then the regexp will fail to match a synchronized C<TGA>
1640and start stepping ahead one character position at a time, not what we
1641want. The solution is to use C<\G> to anchor the match to the codon
1642alignment:
1643
1644 while ($dna =~ /\G(\w\w\w)*?TGA/g) {
1645 print "Got a TGA stop codon at position ", pos $dna, "\n";
1646 }
1647
1648This prints
1649
1650 Got a TGA stop codon at position 18
1651
1652which is the correct answer. This example illustrates that it is
1653important not only to match what is desired, but to reject what is not
1654desired.
1655
0bd5a82d 1656(There are other regexp modifiers that are available, such as
615d795d 1657C<//o>, but their specialized uses are beyond the
0bd5a82d
KW
1658scope of this introduction. )
1659
7638d2dc 1660=head3 Search and replace
47f9c88b 1661
7638d2dc 1662Regular expressions also play a big role in I<search and replace>
47f9c88b
GS
1663operations in Perl. Search and replace is accomplished with the
1664C<s///> operator. The general form is
1665C<s/regexp/replacement/modifiers>, with everything we know about
1666regexps and modifiers applying in this case as well. The
f5b885cd 1667C<replacement> is a Perl double-quoted string that replaces in the
47f9c88b
GS
1668string whatever is matched with the C<regexp>. The operator C<=~> is
1669also used here to associate a string with C<s///>. If matching
7638d2dc 1670against C<$_>, the S<C<$_ =~>> can be dropped. If there is a match,
f5b885cd 1671C<s///> returns the number of substitutions made; otherwise it returns
47f9c88b
GS
1672false. Here are a few examples:
1673
1674 $x = "Time to feed the cat!";
1675 $x =~ s/cat/hacker/; # $x contains "Time to feed the hacker!"
1676 if ($x =~ s/^(Time.*hacker)!$/$1 now!/) {
1677 $more_insistent = 1;
1678 }
1679 $y = "'quoted words'";
1680 $y =~ s/^'(.*)'$/$1/; # strip single quotes,
1681 # $y contains "quoted words"
1682
1683In the last example, the whole string was matched, but only the part
1684inside the single quotes was grouped. With the C<s///> operator, the
f5b885cd 1685matched variables C<$1>, C<$2>, etc. are immediately available for use
47f9c88b
GS
1686in the replacement expression, so we use C<$1> to replace the quoted
1687string with just what was quoted. With the global modifier, C<s///g>
1688will search and replace all occurrences of the regexp in the string:
1689
1690 $x = "I batted 4 for 4";
1691 $x =~ s/4/four/; # doesn't do it all:
1692 # $x contains "I batted four for 4"
1693 $x = "I batted 4 for 4";
1694 $x =~ s/4/four/g; # does it all:
1695 # $x contains "I batted four for four"
1696
1697If you prefer 'regex' over 'regexp' in this tutorial, you could use
1698the following program to replace it:
1699
1700 % cat > simple_replace
1701 #!/usr/bin/perl
1702 $regexp = shift;
1703 $replacement = shift;
1704 while (<>) {
c2e2285d 1705 s/$regexp/$replacement/g;
47f9c88b
GS
1706 print;
1707 }
1708 ^D
1709
1710 % simple_replace regexp regex perlretut.pod
1711
1712In C<simple_replace> we used the C<s///g> modifier to replace all
c2e2285d
KW
1713occurrences of the regexp on each line. (Even though the regular
1714expression appears in a loop, Perl is smart enough to compile it
1715only once.) As with C<simple_grep>, both the
1716C<print> and the C<s/$regexp/$replacement/g> use C<$_> implicitly.
47f9c88b 1717
4f4d7508
DC
1718If you don't want C<s///> to change your original variable you can use
1719the non-destructive substitute modifier, C<s///r>. This changes the
d6b8a906
KW
1720behavior so that C<s///r> returns the final substituted string
1721(instead of the number of substitutions):
4f4d7508
DC
1722
1723 $x = "I like dogs.";
1724 $y = $x =~ s/dogs/cats/r;
1725 print "$x $y\n";
1726
1727That example will print "I like dogs. I like cats". Notice the original
f5b885cd 1728C<$x> variable has not been affected. The overall
4f4d7508
DC
1729result of the substitution is instead stored in C<$y>. If the
1730substitution doesn't affect anything then the original string is
1731returned:
1732
1733 $x = "I like dogs.";
1734 $y = $x =~ s/elephants/cougars/r;
1735 print "$x $y\n"; # prints "I like dogs. I like dogs."
1736
1737One other interesting thing that the C<s///r> flag allows is chaining
1738substitutions:
1739
1740 $x = "Cats are great.";
1741 print $x =~ s/Cats/Dogs/r =~ s/Dogs/Frogs/r =~ s/Frogs/Hedgehogs/r, "\n";
1742 # prints "Hedgehogs are great."
1743
47f9c88b 1744A modifier available specifically to search and replace is the
f5b885cd
FC
1745C<s///e> evaluation modifier. C<s///e> treats the
1746replacement text as Perl code, rather than a double-quoted
1747string. The value that the code returns is substituted for the
47f9c88b
GS
1748matched substring. C<s///e> is useful if you need to do a bit of
1749computation in the process of replacing text. This example counts
1750character frequencies in a line:
1751
1752 $x = "Bill the cat";
1753 $x =~ s/(.)/$chars{$1}++;$1/eg; # final $1 replaces char with itself
1754 print "frequency of '$_' is $chars{$_}\n"
1755 foreach (sort {$chars{$b} <=> $chars{$a}} keys %chars);
1756
1757This prints
1758
1759 frequency of ' ' is 2
1760 frequency of 't' is 2
1761 frequency of 'l' is 2
1762 frequency of 'B' is 1
1763 frequency of 'c' is 1
1764 frequency of 'e' is 1
1765 frequency of 'h' is 1
1766 frequency of 'i' is 1
1767 frequency of 'a' is 1
1768
1769As with the match C<m//> operator, C<s///> can use other delimiters,
1770such as C<s!!!> and C<s{}{}>, and even C<s{}//>. If single quotes are
f5b885cd
FC
1771used C<s'''>, then the regexp and replacement are
1772treated as single-quoted strings and there are no
1773variable substitutions. C<s///> in list context
47f9c88b
GS
1774returns the same thing as in scalar context, i.e., the number of
1775matches.
1776
7638d2dc 1777=head3 The split function
47f9c88b 1778
7638d2dc 1779The C<split()> function is another place where a regexp is used.
353c6505
DL
1780C<split /regexp/, string, limit> separates the C<string> operand into
1781a list of substrings and returns that list. The regexp must be designed
7638d2dc 1782to match whatever constitutes the separators for the desired substrings.
353c6505 1783The C<limit>, if present, constrains splitting into no more than C<limit>
7638d2dc 1784number of strings. For example, to split a string into words, use
47f9c88b
GS
1785
1786 $x = "Calvin and Hobbes";
1787 @words = split /\s+/, $x; # $word[0] = 'Calvin'
1788 # $word[1] = 'and'
1789 # $word[2] = 'Hobbes'
1790
1791If the empty regexp C<//> is used, the regexp always matches and
1792the string is split into individual characters. If the regexp has
7638d2dc 1793groupings, then the resulting list contains the matched substrings from the
47f9c88b
GS
1794groupings as well. For instance,
1795
1796 $x = "/usr/bin/perl";
1797 @dirs = split m!/!, $x; # $dirs[0] = ''
1798 # $dirs[1] = 'usr'
1799 # $dirs[2] = 'bin'
1800 # $dirs[3] = 'perl'
1801 @parts = split m!(/)!, $x; # $parts[0] = ''
1802 # $parts[1] = '/'
1803 # $parts[2] = 'usr'
1804 # $parts[3] = '/'
1805 # $parts[4] = 'bin'
1806 # $parts[5] = '/'
1807 # $parts[6] = 'perl'
1808
1809Since the first character of $x matched the regexp, C<split> prepended
1810an empty initial element to the list.
1811
1812If you have read this far, congratulations! You now have all the basic
1813tools needed to use regular expressions to solve a wide range of text
1814processing problems. If this is your first time through the tutorial,
f5b885cd 1815why not stop here and play around with regexps a while.... S<Part 2>
47f9c88b
GS
1816concerns the more esoteric aspects of regular expressions and those
1817concepts certainly aren't needed right at the start.
1818
1819=head1 Part 2: Power tools
1820
1821OK, you know the basics of regexps and you want to know more. If
1822matching regular expressions is analogous to a walk in the woods, then
1823the tools discussed in Part 1 are analogous to topo maps and a
1824compass, basic tools we use all the time. Most of the tools in part 2
da75cd15 1825are analogous to flare guns and satellite phones. They aren't used
47f9c88b
GS
1826too often on a hike, but when we are stuck, they can be invaluable.
1827
1828What follows are the more advanced, less used, or sometimes esoteric
7638d2dc 1829capabilities of Perl regexps. In Part 2, we will assume you are
7c579eed 1830comfortable with the basics and concentrate on the advanced features.
47f9c88b
GS
1831
1832=head2 More on characters, strings, and character classes
1833
1834There are a number of escape sequences and character classes that we
1835haven't covered yet.
1836
1837There are several escape sequences that convert characters or strings
7638d2dc 1838between upper and lower case, and they are also available within
353c6505 1839patterns. C<\l> and C<\u> convert the next character to lower or
7638d2dc 1840upper case, respectively:
47f9c88b
GS
1841
1842 $x = "perl";
1843 $string =~ /\u$x/; # matches 'Perl' in $string
1844 $x = "M(rs?|s)\\."; # note the double backslash
1845 $string =~ /\l$x/; # matches 'mr.', 'mrs.', and 'ms.',
1846
7638d2dc
WL
1847A C<\L> or C<\U> indicates a lasting conversion of case, until
1848terminated by C<\E> or thrown over by another C<\U> or C<\L>:
47f9c88b
GS
1849
1850 $x = "This word is in lower case:\L SHOUT\E";
1851 $x =~ /shout/; # matches
1852 $x = "I STILL KEYPUNCH CARDS FOR MY 360"
1853 $x =~ /\Ukeypunch/; # matches punch card string
1854
1855If there is no C<\E>, case is converted until the end of the
1856string. The regexps C<\L\u$word> or C<\u\L$word> convert the first
1857character of C<$word> to uppercase and the rest of the characters to
1858lowercase.
1859
1860Control characters can be escaped with C<\c>, so that a control-Z
1861character would be matched with C<\cZ>. The escape sequence
1862C<\Q>...C<\E> quotes, or protects most non-alphabetic characters. For
1863instance,
1864
1865 $x = "\QThat !^*&%~& cat!";
1866 $x =~ /\Q!^*&%~&\E/; # check for rough language
1867
1868It does not protect C<$> or C<@>, so that variables can still be
1869substituted.
1870
8e71069f
FC
1871C<\Q>, C<\L>, C<\l>, C<\U>, C<\u> and C<\E> are actually part of
1872double-quotish syntax, and not part of regexp syntax proper. They will
7698aede 1873work if they appear in a regular expression embedded directly in a
8e71069f
FC
1874program, but not when contained in a string that is interpolated in a
1875pattern.
7c579eed 1876
13e5d9cd
BF
1877Perl regexps can handle more than just the
1878standard ASCII character set. Perl supports I<Unicode>, a standard
7638d2dc 1879for representing the alphabets from virtually all of the world's written
38a44b82 1880languages, and a host of symbols. Perl's text strings are Unicode strings, so
2575c402 1881they can contain characters with a value (codepoint or character number) higher
7c579eed 1882than 255.
47f9c88b
GS
1883
1884What does this mean for regexps? Well, regexp users don't need to know
7638d2dc 1885much about Perl's internal representation of strings. But they do need
2575c402
JW
1886to know 1) how to represent Unicode characters in a regexp and 2) that
1887a matching operation will treat the string to be searched as a sequence
1888of characters, not bytes. The answer to 1) is that Unicode characters
f0a2b745 1889greater than C<chr(255)> are represented using the C<\x{hex}> notation, because
5f67e4c9
KW
1890\x hex (without curly braces) doesn't go further than 255. (Starting in Perl
18915.14, if you're an octal fan, you can also use C<\o{oct}>.)
47f9c88b 1892
47f9c88b
GS
1893 /\x{263a}/; # match a Unicode smiley face :)
1894
7638d2dc 1895B<NOTE>: In Perl 5.6.0 it used to be that one needed to say C<use
72ff2908
JH
1896utf8> to use any Unicode features. This is no more the case: for
1897almost all Unicode processing, the explicit C<utf8> pragma is not
1898needed. (The only case where it matters is if your Perl script is in
1899Unicode and encoded in UTF-8, then an explicit C<use utf8> is needed.)
47f9c88b
GS
1900
1901Figuring out the hexadecimal sequence of a Unicode character you want
1902or deciphering someone else's hexadecimal Unicode regexp is about as
1903much fun as programming in machine code. So another way to specify
e526e8bb
KW
1904Unicode characters is to use the I<named character> escape
1905sequence C<\N{I<name>}>. I<name> is a name for the Unicode character, as
55eda711
JH
1906specified in the Unicode standard. For instance, if we wanted to
1907represent or match the astrological sign for the planet Mercury, we
1908could use
47f9c88b 1909
47f9c88b
GS
1910 $x = "abc\N{MERCURY}def";
1911 $x =~ /\N{MERCURY}/; # matches
1912
fbb93542 1913One can also use "short" names:
47f9c88b 1914
47f9c88b 1915 print "\N{GREEK SMALL LETTER SIGMA} is called sigma.\n";
47f9c88b
GS
1916 print "\N{greek:Sigma} is an upper-case sigma.\n";
1917
fbb93542
KW
1918You can also restrict names to a certain alphabet by specifying the
1919L<charnames> pragma:
1920
47f9c88b
GS
1921 use charnames qw(greek);
1922 print "\N{sigma} is Greek sigma\n";
1923
0bd42786
KW
1924An index of character names is available on-line from the Unicode
1925Consortium, L<http://www.unicode.org/charts/charindex.html>; explanatory
1926material with links to other resources at
1927L<http://www.unicode.org/standard/where>.
47f9c88b 1928
13e5d9cd
BF
1929The answer to requirement 2) is that a regexp (mostly)
1930uses Unicode characters. The "mostly" is for messy backward
615d795d
KW
1931compatibility reasons, but starting in Perl 5.14, any regex compiled in
1932the scope of a C<use feature 'unicode_strings'> (which is automatically
1933turned on within the scope of a C<use 5.012> or higher) will turn that
1934"mostly" into "always". If you want to handle Unicode properly, you
13e5d9cd 1935should ensure that C<'unicode_strings'> is turned on.
0bd5a82d
KW
1936Internally, this is encoded to bytes using either UTF-8 or a native 8
1937bit encoding, depending on the history of the string, but conceptually
1938it is a sequence of characters, not bytes. See L<perlunitut> for a
1939tutorial about that.
2575c402
JW
1940
1941Let us now discuss Unicode character classes. Just as with Unicode
1942characters, there are named Unicode character classes represented by the
1943C<\p{name}> escape sequence. Closely associated is the C<\P{name}>
1944character class, which is the negation of the C<\p{name}> class. For
1945example, to match lower and uppercase characters,
47f9c88b 1946
47f9c88b
GS
1947 $x = "BOB";
1948 $x =~ /^\p{IsUpper}/; # matches, uppercase char class
1949 $x =~ /^\P{IsUpper}/; # doesn't match, char class sans uppercase
1950 $x =~ /^\p{IsLower}/; # doesn't match, lowercase char class
1951 $x =~ /^\P{IsLower}/; # matches, char class sans lowercase
1952
5f67e4c9
KW
1953(The "Is" is optional.)
1954
86929931
JH
1955Here is the association between some Perl named classes and the
1956traditional Unicode classes:
47f9c88b 1957
86929931 1958 Perl class name Unicode class name or regular expression
47f9c88b 1959
f5868911
JH
1960 IsAlpha /^[LM]/
1961 IsAlnum /^[LMN]/
1962 IsASCII $code <= 127
1963 IsCntrl /^C/
1964 IsBlank $code =~ /^(0020|0009)$/ || /^Z[^lp]/
47f9c88b 1965 IsDigit Nd
f5868911 1966 IsGraph /^([LMNPS]|Co)/
47f9c88b 1967 IsLower Ll
f5868911
JH
1968 IsPrint /^([LMNPS]|Co|Zs)/
1969 IsPunct /^P/
1970 IsSpace /^Z/ || ($code =~ /^(0009|000A|000B|000C|000D)$/
08ce8fc6 1971 IsSpacePerl /^Z/ || ($code =~ /^(0009|000A|000C|000D|0085|2028|2029)$/
f5868911
JH
1972 IsUpper /^L[ut]/
1973 IsWord /^[LMN]/ || $code eq "005F"
47f9c88b
GS
1974 IsXDigit $code =~ /^00(3[0-9]|[46][1-6])$/
1975
7c579eed
FC
1976You can also use the official Unicode class names with C<\p> and
1977C<\P>, like C<\p{L}> for Unicode 'letters', C<\p{Lu}> for uppercase
86929931
JH
1978letters, or C<\P{Nd}> for non-digits. If a C<name> is just one
1979letter, the braces can be dropped. For instance, C<\pM> is the
98f22ffc 1980character class of Unicode 'marks', for example accent marks.
32293815
JH
1981For the full list see L<perlunicode>.
1982
7c579eed 1983Unicode has also been separated into various sets of characters
7638d2dc
WL
1984which you can test with C<\p{...}> (in) and C<\P{...}> (not in).
1985To test whether a character is (or is not) an element of a script
353c6505 1986you would use the script name, for example C<\p{Latin}>, C<\p{Greek}>,
1cd08ccc 1987or C<\P{Katakana}>.
e1b711da
KW
1988
1989What we have described so far is the single form of the C<\p{...}> character
1990classes. There is also a compound form which you may run into. These
1991look like C<\p{name=value}> or C<\p{name:value}> (the equals sign and colon
1992can be used interchangeably). These are more general than the single form,
1993and in fact most of the single forms are just Perl-defined shortcuts for common
1994compound forms. For example, the script examples in the previous paragraph
1995could be written equivalently as C<\p{Script=Latin}>, C<\p{Script:Greek}>, and
1996C<\P{script=katakana}> (case is irrelevant between the C<{}> braces). You may
1997never have to use the compound forms, but sometimes it is necessary, and their
1998use can make your code easier to understand.
47f9c88b 1999
7638d2dc 2000C<\X> is an abbreviation for a character class that comprises
5f67e4c9 2001a Unicode I<extended grapheme cluster>. This represents a "logical character":
e1b711da
KW
2002what appears to be a single character, but may be represented internally by more
2003than one. As an example, using the Unicode full names, e.g., S<C<A + COMBINING
2004RING>> is a grapheme cluster with base character C<A> and combining character
2005S<C<COMBINING RING>>, which translates in Danish to A with the circle atop it,
2006as in the word Angstrom.
47f9c88b 2007
da75cd15 2008For the full and latest information about Unicode see the latest
e1b711da 2009Unicode standard, or the Unicode Consortium's website L<http://www.unicode.org>
5e42d7b4 2010
7c579eed 2011As if all those classes weren't enough, Perl also defines POSIX-style
47f9c88b 2012character classes. These have the form C<[:name:]>, with C<name> the
aaa51d5e
JF
2013name of the POSIX class. The POSIX classes are C<alpha>, C<alnum>,
2014C<ascii>, C<cntrl>, C<digit>, C<graph>, C<lower>, C<print>, C<punct>,
2015C<space>, C<upper>, and C<xdigit>, and two extensions, C<word> (a Perl
0bd5a82d
KW
2016extension to match C<\w>), and C<blank> (a GNU extension). The C<//a>
2017modifier restricts these to matching just in the ASCII range; otherwise
2018they can match the same as their corresponding Perl Unicode classes:
2019C<[:upper:]> is the same as C<\p{IsUpper}>, etc. (There are some
2020exceptions and gotchas with this; see L<perlrecharclass> for a full
2021discussion.) The C<[:digit:]>, C<[:word:]>, and
47f9c88b 2022C<[:space:]> correspond to the familiar C<\d>, C<\w>, and C<\s>
aaa51d5e 2023character classes. To negate a POSIX class, put a C<^> in front of
7c579eed
FC
2024the name, so that, e.g., C<[:^digit:]> corresponds to C<\D> and, under
2025Unicode, C<\P{IsDigit}>. The Unicode and POSIX character classes can
54c18d04
MK
2026be used just like C<\d>, with the exception that POSIX character
2027classes can only be used inside of a character class:
47f9c88b
GS
2028
2029 /\s+[abc[:digit:]xyz]\s*/; # match a,b,c,x,y,z, or a digit
54c18d04 2030 /^=item\s[[:digit:]]/; # match '=item',
47f9c88b 2031 # followed by a space and a digit
47f9c88b
GS
2032 /\s+[abc\p{IsDigit}xyz]\s+/; # match a,b,c,x,y,z, or a digit
2033 /^=item\s\p{IsDigit}/; # match '=item',
2034 # followed by a space and a digit
2035
2036Whew! That is all the rest of the characters and character classes.
2037
2038=head2 Compiling and saving regular expressions
2039
c2e2285d
KW
2040In Part 1 we mentioned that Perl compiles a regexp into a compact
2041sequence of opcodes. Thus, a compiled regexp is a data structure
47f9c88b
GS
2042that can be stored once and used again and again. The regexp quote
2043C<qr//> does exactly that: C<qr/string/> compiles the C<string> as a
2044regexp and transforms the result into a form that can be assigned to a
2045variable:
2046
2047 $reg = qr/foo+bar?/; # reg contains a compiled regexp
2048
2049Then C<$reg> can be used as a regexp:
2050
2051 $x = "fooooba";
2052 $x =~ $reg; # matches, just like /foo+bar?/
2053 $x =~ /$reg/; # same thing, alternate form
2054
2055C<$reg> can also be interpolated into a larger regexp:
2056
2057 $x =~ /(abc)?$reg/; # still matches
2058
2059As with the matching operator, the regexp quote can use different
7638d2dc
WL
2060delimiters, e.g., C<qr!!>, C<qr{}> or C<qr~~>. Apostrophes
2061as delimiters (C<qr''>) inhibit any interpolation.
47f9c88b
GS
2062
2063Pre-compiled regexps are useful for creating dynamic matches that
2064don't need to be recompiled each time they are encountered. Using
7638d2dc
WL
2065pre-compiled regexps, we write a C<grep_step> program which greps
2066for a sequence of patterns, advancing to the next pattern as soon
2067as one has been satisfied.
47f9c88b 2068
7638d2dc 2069 % cat > grep_step
47f9c88b 2070 #!/usr/bin/perl
7638d2dc 2071 # grep_step - match <number> regexps, one after the other
47f9c88b
GS
2072 # usage: multi_grep <number> regexp1 regexp2 ... file1 file2 ...
2073
2074 $number = shift;
2075 $regexp[$_] = shift foreach (0..$number-1);
2076 @compiled = map qr/$_/, @regexp;
2077 while ($line = <>) {
7638d2dc
WL
2078 if ($line =~ /$compiled[0]/) {
2079 print $line;
2080 shift @compiled;
2081 last unless @compiled;
47f9c88b
GS
2082 }
2083 }
2084 ^D
2085
7638d2dc
WL
2086 % grep_step 3 shift print last grep_step
2087 $number = shift;
2088 print $line;
2089 last unless @compiled;
47f9c88b
GS
2090
2091Storing pre-compiled regexps in an array C<@compiled> allows us to
2092simply loop through the regexps without any recompilation, thus gaining
2093flexibility without sacrificing speed.
2094
7638d2dc
WL
2095
2096=head2 Composing regular expressions at runtime
2097
2098Backtracking is more efficient than repeated tries with different regular
2099expressions. If there are several regular expressions and a match with
353c6505 2100any of them is acceptable, then it is possible to combine them into a set
7638d2dc 2101of alternatives. If the individual expressions are input data, this
353c6505
DL
2102can be done by programming a join operation. We'll exploit this idea in
2103an improved version of the C<simple_grep> program: a program that matches
7638d2dc
WL
2104multiple patterns:
2105
2106 % cat > multi_grep
2107 #!/usr/bin/perl
2108 # multi_grep - match any of <number> regexps
2109 # usage: multi_grep <number> regexp1 regexp2 ... file1 file2 ...
2110
2111 $number = shift;
2112 $regexp[$_] = shift foreach (0..$number-1);
2113 $pattern = join '|', @regexp;
2114
2115 while ($line = <>) {
c2e2285d 2116 print $line if $line =~ /$pattern/;
7638d2dc
WL
2117 }
2118 ^D
2119
2120 % multi_grep 2 shift for multi_grep
2121 $number = shift;
2122 $regexp[$_] = shift foreach (0..$number-1);
2123
2124Sometimes it is advantageous to construct a pattern from the I<input>
2125that is to be analyzed and use the permissible values on the left
2126hand side of the matching operations. As an example for this somewhat
353c6505 2127paradoxical situation, let's assume that our input contains a command
7638d2dc 2128verb which should match one out of a set of available command verbs,
353c6505 2129with the additional twist that commands may be abbreviated as long as
7638d2dc
WL
2130the given string is unique. The program below demonstrates the basic
2131algorithm.
2132
2133 % cat > keymatch
2134 #!/usr/bin/perl
2135 $kwds = 'copy compare list print';
2136 while( $command = <> ){
2137 $command =~ s/^\s+|\s+$//g; # trim leading and trailing spaces
2138 if( ( @matches = $kwds =~ /\b$command\w*/g ) == 1 ){
92a24ac3 2139 print "command: '@matches'\n";
7638d2dc
WL
2140 } elsif( @matches == 0 ){
2141 print "no such command: '$command'\n";
2142 } else {
2143 print "not unique: '$command' (could be one of: @matches)\n";
2144 }
2145 }
2146 ^D
2147
2148 % keymatch
2149 li
2150 command: 'list'
2151 co
2152 not unique: 'co' (could be one of: copy compare)
2153 printer
2154 no such command: 'printer'
2155
2156Rather than trying to match the input against the keywords, we match the
2157combined set of keywords against the input. The pattern matching
353c6505
DL
2158operation S<C<$kwds =~ /\b($command\w*)/g>> does several things at the
2159same time. It makes sure that the given command begins where a keyword
2160begins (C<\b>). It tolerates abbreviations due to the added C<\w*>. It
2161tells us the number of matches (C<scalar @matches>) and all the keywords
7638d2dc 2162that were actually matched. You could hardly ask for more.
7638d2dc 2163
47f9c88b
GS
2164=head2 Embedding comments and modifiers in a regular expression
2165
2166Starting with this section, we will be discussing Perl's set of
7638d2dc 2167I<extended patterns>. These are extensions to the traditional regular
47f9c88b
GS
2168expression syntax that provide powerful new tools for pattern
2169matching. We have already seen extensions in the form of the minimal
6b3ddc02
FC
2170matching constructs C<??>, C<*?>, C<+?>, C<{n,m}?>, and C<{n,}?>. Most
2171of the extensions below have the form C<(?char...)>, where the
47f9c88b
GS
2172C<char> is a character that determines the type of extension.
2173
2174The first extension is an embedded comment C<(?#text)>. This embeds a
2175comment into the regular expression without affecting its meaning. The
2176comment should not have any closing parentheses in the text. An
2177example is
2178
2179 /(?# Match an integer:)[+-]?\d+/;
2180
2181This style of commenting has been largely superseded by the raw,
2182freeform commenting that is allowed with the C<//x> modifier.
2183
5f67e4c9 2184Most modifiers, such as C<//i>, C<//m>, C<//s> and C<//x> (or any
24549070 2185combination thereof) can also be embedded in
47f9c88b
GS
2186a regexp using C<(?i)>, C<(?m)>, C<(?s)>, and C<(?x)>. For instance,
2187
2188 /(?i)yes/; # match 'yes' case insensitively
2189 /yes/i; # same thing
2190 /(?x)( # freeform version of an integer regexp
2191 [+-]? # match an optional sign
2192 \d+ # match a sequence of digits
2193 )
2194 /x;
2195
2196Embedded modifiers can have two important advantages over the usual
2197modifiers. Embedded modifiers allow a custom set of modifiers to
2198I<each> regexp pattern. This is great for matching an array of regexps
2199that must have different modifiers:
2200
2201 $pattern[0] = '(?i)doctor';
2202 $pattern[1] = 'Johnson';
2203 ...
2204 while (<>) {
2205 foreach $patt (@pattern) {
2206 print if /$patt/;
2207 }
2208 }
2209
24549070 2210The second advantage is that embedded modifiers (except C<//p>, which
7638d2dc 2211modifies the entire regexp) only affect the regexp
47f9c88b
GS
2212inside the group the embedded modifier is contained in. So grouping
2213can be used to localize the modifier's effects:
2214
2215 /Answer: ((?i)yes)/; # matches 'Answer: yes', 'Answer: YES', etc.
2216
2217Embedded modifiers can also turn off any modifiers already present
2218by using, e.g., C<(?-i)>. Modifiers can also be combined into
2219a single expression, e.g., C<(?s-i)> turns on single line mode and
2220turns off case insensitivity.
2221
7638d2dc 2222Embedded modifiers may also be added to a non-capturing grouping.
47f9c88b
GS
2223C<(?i-m:regexp)> is a non-capturing grouping that matches C<regexp>
2224case insensitively and turns off multi-line mode.
2225
7638d2dc 2226
47f9c88b
GS
2227=head2 Looking ahead and looking behind
2228
2229This section concerns the lookahead and lookbehind assertions. First,
2230a little background.
2231
2232In Perl regular expressions, most regexp elements 'eat up' a certain
2233amount of string when they match. For instance, the regexp element
2234C<[abc}]> eats up one character of the string when it matches, in the
7638d2dc 2235sense that Perl moves to the next character position in the string
47f9c88b
GS
2236after the match. There are some elements, however, that don't eat up
2237characters (advance the character position) if they match. The examples
2238we have seen so far are the anchors. The anchor C<^> matches the
2239beginning of the line, but doesn't eat any characters. Similarly, the
7638d2dc 2240word boundary anchor C<\b> matches wherever a character matching C<\w>
353c6505 2241is next to a character that doesn't, but it doesn't eat up any
6b3ddc02
FC
2242characters itself. Anchors are examples of I<zero-width assertions>:
2243zero-width, because they consume
47f9c88b
GS
2244no characters, and assertions, because they test some property of the
2245string. In the context of our walk in the woods analogy to regexp
2246matching, most regexp elements move us along a trail, but anchors have
2247us stop a moment and check our surroundings. If the local environment
2248checks out, we can proceed forward. But if the local environment
2249doesn't satisfy us, we must backtrack.
2250
2251Checking the environment entails either looking ahead on the trail,
2252looking behind, or both. C<^> looks behind, to see that there are no
2253characters before. C<$> looks ahead, to see that there are no
2254characters after. C<\b> looks both ahead and behind, to see if the
7638d2dc 2255characters on either side differ in their "word-ness".
47f9c88b
GS
2256
2257The lookahead and lookbehind assertions are generalizations of the
2258anchor concept. Lookahead and lookbehind are zero-width assertions
2259that let us specify which characters we want to test for. The
2260lookahead assertion is denoted by C<(?=regexp)> and the lookbehind
a6b2f353 2261assertion is denoted by C<< (?<=fixed-regexp) >>. Some examples are
47f9c88b
GS
2262
2263 $x = "I catch the housecat 'Tom-cat' with catnip";
7638d2dc 2264 $x =~ /cat(?=\s)/; # matches 'cat' in 'housecat'
47f9c88b
GS
2265 @catwords = ($x =~ /(?<=\s)cat\w+/g); # matches,
2266 # $catwords[0] = 'catch'
2267 # $catwords[1] = 'catnip'
2268 $x =~ /\bcat\b/; # matches 'cat' in 'Tom-cat'
2269 $x =~ /(?<=\s)cat(?=\s)/; # doesn't match; no isolated 'cat' in
2270 # middle of $x
2271
a6b2f353 2272Note that the parentheses in C<(?=regexp)> and C<< (?<=regexp) >> are
47f9c88b
GS
2273non-capturing, since these are zero-width assertions. Thus in the
2274second regexp, the substrings captured are those of the whole regexp
a6b2f353
GS
2275itself. Lookahead C<(?=regexp)> can match arbitrary regexps, but
2276lookbehind C<< (?<=fixed-regexp) >> only works for regexps of fixed
2277width, i.e., a fixed number of characters long. Thus
2278C<< (?<=(ab|bc)) >> is fine, but C<< (?<=(ab)*) >> is not. The
2279negated versions of the lookahead and lookbehind assertions are
2280denoted by C<(?!regexp)> and C<< (?<!fixed-regexp) >> respectively.
2281They evaluate true if the regexps do I<not> match:
47f9c88b
GS
2282
2283 $x = "foobar";
2284 $x =~ /foo(?!bar)/; # doesn't match, 'bar' follows 'foo'
2285 $x =~ /foo(?!baz)/; # matches, 'baz' doesn't follow 'foo'
2286 $x =~ /(?<!\s)foo/; # matches, there is no \s before 'foo'
2287
f14c76ed
RGS
2288The C<\C> is unsupported in lookbehind, because the already
2289treacherous definition of C<\C> would become even more so
2290when going backwards.
2291
7638d2dc
WL
2292Here is an example where a string containing blank-separated words,
2293numbers and single dashes is to be split into its components.
2294Using C</\s+/> alone won't work, because spaces are not required between
2295dashes, or a word or a dash. Additional places for a split are established
2296by looking ahead and behind:
47f9c88b 2297
7638d2dc
WL
2298 $str = "one two - --6-8";
2299 @toks = split / \s+ # a run of spaces
2300 | (?<=\S) (?=-) # any non-space followed by '-'
2301 | (?<=-) (?=\S) # a '-' followed by any non-space
2302 /x, $str; # @toks = qw(one two - - - 6 - 8)
47f9c88b 2303
7638d2dc
WL
2304
2305=head2 Using independent subexpressions to prevent backtracking
2306
2307I<Independent subexpressions> are regular expressions, in the
47f9c88b
GS
2308context of a larger regular expression, that function independently of
2309the larger regular expression. That is, they consume as much or as
2310little of the string as they wish without regard for the ability of
2311the larger regexp to match. Independent subexpressions are represented
2312by C<< (?>regexp) >>. We can illustrate their behavior by first
2313considering an ordinary regexp:
2314
2315 $x = "ab";
2316 $x =~ /a*ab/; # matches
2317
2318This obviously matches, but in the process of matching, the
2319subexpression C<a*> first grabbed the C<a>. Doing so, however,
2320wouldn't allow the whole regexp to match, so after backtracking, C<a*>
2321eventually gave back the C<a> and matched the empty string. Here, what
2322C<a*> matched was I<dependent> on what the rest of the regexp matched.
2323
2324Contrast that with an independent subexpression:
2325
2326 $x =~ /(?>a*)ab/; # doesn't match!
2327
2328The independent subexpression C<< (?>a*) >> doesn't care about the rest
2329of the regexp, so it sees an C<a> and grabs it. Then the rest of the
2330regexp C<ab> cannot match. Because C<< (?>a*) >> is independent, there
da75cd15 2331is no backtracking and the independent subexpression does not give
47f9c88b
GS
2332up its C<a>. Thus the match of the regexp as a whole fails. A similar
2333behavior occurs with completely independent regexps:
2334
2335 $x = "ab";
2336 $x =~ /a*/g; # matches, eats an 'a'
2337 $x =~ /\Gab/g; # doesn't match, no 'a' available
2338
2339Here C<//g> and C<\G> create a 'tag team' handoff of the string from
2340one regexp to the other. Regexps with an independent subexpression are
2341much like this, with a handoff of the string to the independent
2342subexpression, and a handoff of the string back to the enclosing
2343regexp.
2344
2345The ability of an independent subexpression to prevent backtracking
2346can be quite useful. Suppose we want to match a non-empty string
2347enclosed in parentheses up to two levels deep. Then the following
2348regexp matches:
2349
2350 $x = "abc(de(fg)h"; # unbalanced parentheses
2351 $x =~ /\( ( [^()]+ | \([^()]*\) )+ \)/x;
2352
2353The regexp matches an open parenthesis, one or more copies of an
2354alternation, and a close parenthesis. The alternation is two-way, with
2355the first alternative C<[^()]+> matching a substring with no
2356parentheses and the second alternative C<\([^()]*\)> matching a
2357substring delimited by parentheses. The problem with this regexp is
2358that it is pathological: it has nested indeterminate quantifiers
07698885 2359of the form C<(a+|b)+>. We discussed in Part 1 how nested quantifiers
47f9c88b
GS
2360like this could take an exponentially long time to execute if there
2361was no match possible. To prevent the exponential blowup, we need to
2362prevent useless backtracking at some point. This can be done by
2363enclosing the inner quantifier as an independent subexpression:
2364
2365 $x =~ /\( ( (?>[^()]+) | \([^()]*\) )+ \)/x;
2366
2367Here, C<< (?>[^()]+) >> breaks the degeneracy of string partitioning
2368by gobbling up as much of the string as possible and keeping it. Then
2369match failures fail much more quickly.
2370
7638d2dc 2371
47f9c88b
GS
2372=head2 Conditional expressions
2373
7638d2dc 2374A I<conditional expression> is a form of if-then-else statement
47f9c88b
GS
2375that allows one to choose which patterns are to be matched, based on
2376some condition. There are two types of conditional expression:
2377C<(?(condition)yes-regexp)> and
2378C<(?(condition)yes-regexp|no-regexp)>. C<(?(condition)yes-regexp)> is
7638d2dc 2379like an S<C<'if () {}'>> statement in Perl. If the C<condition> is true,
47f9c88b 2380the C<yes-regexp> will be matched. If the C<condition> is false, the
7638d2dc
WL
2381C<yes-regexp> will be skipped and Perl will move onto the next regexp
2382element. The second form is like an S<C<'if () {} else {}'>> statement
47f9c88b
GS
2383in Perl. If the C<condition> is true, the C<yes-regexp> will be
2384matched, otherwise the C<no-regexp> will be matched.
2385
7638d2dc 2386The C<condition> can have several forms. The first form is simply an
47f9c88b 2387integer in parentheses C<(integer)>. It is true if the corresponding
7638d2dc 2388backreference C<\integer> matched earlier in the regexp. The same
c27a5cfe 2389thing can be done with a name associated with a capture group, written
7638d2dc 2390as C<< (<name>) >> or C<< ('name') >>. The second form is a bare
6b3ddc02 2391zero-width assertion C<(?...)>, either a lookahead, a lookbehind, or a
7638d2dc
WL
2392code assertion (discussed in the next section). The third set of forms
2393provides tests that return true if the expression is executed within
2394a recursion (C<(R)>) or is being called from some capturing group,
2395referenced either by number (C<(R1)>, C<(R2)>,...) or by name
2396(C<(R&name)>).
2397
2398The integer or name form of the C<condition> allows us to choose,
2399with more flexibility, what to match based on what matched earlier in the
2400regexp. This searches for words of the form C<"$x$x"> or C<"$x$y$y$x">:
47f9c88b 2401
d8b950dc 2402 % simple_grep '^(\w+)(\w+)?(?(2)\g2\g1|\g1)$' /usr/dict/words
47f9c88b
GS
2403 beriberi
2404 coco
2405 couscous
2406 deed
2407 ...
2408 toot
2409 toto
2410 tutu
2411
2412The lookbehind C<condition> allows, along with backreferences,
2413an earlier part of the match to influence a later part of the
2414match. For instance,
2415
2416 /[ATGC]+(?(?<=AA)G|C)$/;
2417
2418matches a DNA sequence such that it either ends in C<AAG>, or some
2419other base pair combination and C<C>. Note that the form is
a6b2f353
GS
2420C<< (?(?<=AA)G|C) >> and not C<< (?((?<=AA))G|C) >>; for the
2421lookahead, lookbehind or code assertions, the parentheses around the
2422conditional are not needed.
47f9c88b 2423
7638d2dc
WL
2424
2425=head2 Defining named patterns
2426
2427Some regular expressions use identical subpatterns in several places.
2428Starting with Perl 5.10, it is possible to define named subpatterns in
2429a section of the pattern so that they can be called up by name
2430anywhere in the pattern. This syntactic pattern for this definition
2431group is C<< (?(DEFINE)(?<name>pattern)...) >>. An insertion
2432of a named pattern is written as C<(?&name)>.
2433
2434The example below illustrates this feature using the pattern for
2435floating point numbers that was presented earlier on. The three
2436subpatterns that are used more than once are the optional sign, the
2437digit sequence for an integer and the decimal fraction. The DEFINE
2438group at the end of the pattern contains their definition. Notice
2439that the decimal fraction pattern is the first place where we can
2440reuse the integer pattern.
2441
353c6505 2442 /^ (?&osg)\ * ( (?&int)(?&dec)? | (?&dec) )
7638d2dc
WL
2443 (?: [eE](?&osg)(?&int) )?
2444 $
2445 (?(DEFINE)
2446 (?<osg>[-+]?) # optional sign
2447 (?<int>\d++) # integer
2448 (?<dec>\.(?&int)) # decimal fraction
2449 )/x
2450
2451
2452=head2 Recursive patterns
2453
2454This feature (introduced in Perl 5.10) significantly extends the
2455power of Perl's pattern matching. By referring to some other
2456capture group anywhere in the pattern with the construct
353c6505 2457C<(?group-ref)>, the I<pattern> within the referenced group is used
7638d2dc
WL
2458as an independent subpattern in place of the group reference itself.
2459Because the group reference may be contained I<within> the group it
2460refers to, it is now possible to apply pattern matching to tasks that
2461hitherto required a recursive parser.
2462
2463To illustrate this feature, we'll design a pattern that matches if
2464a string contains a palindrome. (This is a word or a sentence that,
2465while ignoring spaces, interpunctuation and case, reads the same backwards
2466as forwards. We begin by observing that the empty string or a string
2467containing just one word character is a palindrome. Otherwise it must
2468have a word character up front and the same at its end, with another
2469palindrome in between.
2470
fd2b7f55 2471 /(?: (\w) (?...Here be a palindrome...) \g{-1} | \w? )/x
7638d2dc 2472
e57a4e52 2473Adding C<\W*> at either end to eliminate what is to be ignored, we already
7638d2dc
WL
2474have the full pattern:
2475
2476 my $pp = qr/^(\W* (?: (\w) (?1) \g{-1} | \w? ) \W*)$/ix;
2477 for $s ( "saippuakauppias", "A man, a plan, a canal: Panama!" ){
2478 print "'$s' is a palindrome\n" if $s =~ /$pp/;
2479 }
2480
2481In C<(?...)> both absolute and relative backreferences may be used.
2482The entire pattern can be reinserted with C<(?R)> or C<(?0)>.
c27a5cfe
KW
2483If you prefer to name your groups, you can use C<(?&name)> to
2484recurse into that group.
7638d2dc
WL
2485
2486
47f9c88b
GS
2487=head2 A bit of magic: executing Perl code in a regular expression
2488
2489Normally, regexps are a part of Perl expressions.
7638d2dc 2490I<Code evaluation> expressions turn that around by allowing
da75cd15 2491arbitrary Perl code to be a part of a regexp. A code evaluation
7638d2dc 2492expression is denoted C<(?{code})>, with I<code> a string of Perl
47f9c88b
GS
2493statements.
2494
353c6505 2495Be warned that this feature is considered experimental, and may be
7638d2dc
WL
2496changed without notice.
2497
47f9c88b
GS
2498Code expressions are zero-width assertions, and the value they return
2499depends on their environment. There are two possibilities: either the
2500code expression is used as a conditional in a conditional expression
2501C<(?(condition)...)>, or it is not. If the code expression is a
2502conditional, the code is evaluated and the result (i.e., the result of
2503the last statement) is used to determine truth or falsehood. If the
2504code expression is not used as a conditional, the assertion always
2505evaluates true and the result is put into the special variable
2506C<$^R>. The variable C<$^R> can then be used in code expressions later
2507in the regexp. Here are some silly examples:
2508
2509 $x = "abcdef";
2510 $x =~ /abc(?{print "Hi Mom!";})def/; # matches,
2511 # prints 'Hi Mom!'
2512 $x =~ /aaa(?{print "Hi Mom!";})def/; # doesn't match,
2513 # no 'Hi Mom!'
745e1e41
DC
2514
2515Pay careful attention to the next example:
2516
47f9c88b
GS
2517 $x =~ /abc(?{print "Hi Mom!";})ddd/; # doesn't match,
2518 # no 'Hi Mom!'
745e1e41
DC
2519 # but why not?
2520
2521At first glance, you'd think that it shouldn't print, because obviously
2522the C<ddd> isn't going to match the target string. But look at this
2523example:
2524
87167316
RGS
2525 $x =~ /abc(?{print "Hi Mom!";})[dD]dd/; # doesn't match,
2526 # but _does_ print
745e1e41
DC
2527
2528Hmm. What happened here? If you've been following along, you know that
ac036724 2529the above pattern should be effectively (almost) the same as the last one;
2530enclosing the C<d> in a character class isn't going to change what it
745e1e41
DC
2531matches. So why does the first not print while the second one does?
2532
7638d2dc 2533The answer lies in the optimizations the regex engine makes. In the first
745e1e41
DC
2534case, all the engine sees are plain old characters (aside from the
2535C<?{}> construct). It's smart enough to realize that the string 'ddd'
2536doesn't occur in our target string before actually running the pattern
2537through. But in the second case, we've tricked it into thinking that our
87167316 2538pattern is more complicated. It takes a look, sees our
745e1e41
DC
2539character class, and decides that it will have to actually run the
2540pattern to determine whether or not it matches, and in the process of
2541running it hits the print statement before it discovers that we don't
2542have a match.
2543
2544To take a closer look at how the engine does optimizations, see the
2545section L<"Pragmas and debugging"> below.
2546
2547More fun with C<?{}>:
2548
47f9c88b
GS
2549 $x =~ /(?{print "Hi Mom!";})/; # matches,
2550 # prints 'Hi Mom!'
2551 $x =~ /(?{$c = 1;})(?{print "$c";})/; # matches,
2552 # prints '1'
2553 $x =~ /(?{$c = 1;})(?{print "$^R";})/; # matches,
2554 # prints '1'
2555
2556The bit of magic mentioned in the section title occurs when the regexp
2557backtracks in the process of searching for a match. If the regexp
2558backtracks over a code expression and if the variables used within are
2559localized using C<local>, the changes in the variables produced by the
2560code expression are undone! Thus, if we wanted to count how many times
2561a character got matched inside a group, we could use, e.g.,
2562
2563 $x = "aaaa";
2564 $count = 0; # initialize 'a' count
2565 $c = "bob"; # test if $c gets clobbered
2566 $x =~ /(?{local $c = 0;}) # initialize count
2567 ( a # match 'a'
2568 (?{local $c = $c + 1;}) # increment count
2569 )* # do this any number of times,
2570 aa # but match 'aa' at the end
2571 (?{$count = $c;}) # copy local $c var into $count
2572 /x;
2573 print "'a' count is $count, \$c variable is '$c'\n";
2574
2575This prints
2576
2577 'a' count is 2, $c variable is 'bob'
2578
7638d2dc
WL
2579If we replace the S<C< (?{local $c = $c + 1;})>> with
2580S<C< (?{$c = $c + 1;})>>, the variable changes are I<not> undone
47f9c88b
GS
2581during backtracking, and we get
2582
2583 'a' count is 4, $c variable is 'bob'
2584
2585Note that only localized variable changes are undone. Other side
2586effects of code expression execution are permanent. Thus
2587
2588 $x = "aaaa";
2589 $x =~ /(a(?{print "Yow\n";}))*aa/;
2590
2591produces
2592
2593 Yow
2594 Yow
2595 Yow
2596 Yow
2597
2598The result C<$^R> is automatically localized, so that it will behave
2599properly in the presence of backtracking.
2600
7638d2dc
WL
2601This example uses a code expression in a conditional to match a
2602definite article, either 'the' in English or 'der|die|das' in German:
47f9c88b 2603
47f9c88b
GS
2604 $lang = 'DE'; # use German
2605 ...
2606 $text = "das";
2607 print "matched\n"
2608 if $text =~ /(?(?{
2609 $lang eq 'EN'; # is the language English?
2610 })
2611 the | # if so, then match 'the'
7638d2dc 2612 (der|die|das) # else, match 'der|die|das'
47f9c88b
GS
2613 )
2614 /xi;
2615
2616Note that the syntax here is C<(?(?{...})yes-regexp|no-regexp)>, not
2617C<(?((?{...}))yes-regexp|no-regexp)>. In other words, in the case of a
2618code expression, we don't need the extra parentheses around the
2619conditional.
2620
e128ab2c
DM
2621If you try to use code expressions where the code text is contained within
2622an interpolated variable, rather than appearing literally in the pattern,
2623Perl may surprise you:
a6b2f353
GS
2624
2625 $bar = 5;
2626 $pat = '(?{ 1 })';
2627 /foo(?{ $bar })bar/; # compiles ok, $bar not interpolated
e128ab2c 2628 /foo(?{ 1 })$bar/; # compiles ok, $bar interpolated
a6b2f353
GS
2629 /foo${pat}bar/; # compile error!
2630
2631 $pat = qr/(?{ $foo = 1 })/; # precompile code regexp
2632 /foo${pat}bar/; # compiles ok
2633
e128ab2c
DM
2634If a regexp has a variable that interpolates a code expression, Perl
2635treats the regexp as an error. If the code expression is precompiled into
2636a variable, however, interpolating is ok. The question is, why is this an
2637error?
a6b2f353
GS
2638
2639The reason is that variable interpolation and code expressions
2640together pose a security risk. The combination is dangerous because
2641many programmers who write search engines often take user input and
2642plug it directly into a regexp:
47f9c88b
GS
2643
2644 $regexp = <>; # read user-supplied regexp
2645 $chomp $regexp; # get rid of possible newline
2646 $text =~ /$regexp/; # search $text for the $regexp
2647
a6b2f353
GS
2648If the C<$regexp> variable contains a code expression, the user could
2649then execute arbitrary Perl code. For instance, some joker could
7638d2dc
WL
2650search for S<C<system('rm -rf *');>> to erase your files. In this
2651sense, the combination of interpolation and code expressions I<taints>
47f9c88b 2652your regexp. So by default, using both interpolation and code
a6b2f353
GS
2653expressions in the same regexp is not allowed. If you're not
2654concerned about malicious users, it is possible to bypass this
7638d2dc 2655security check by invoking S<C<use re 'eval'>>:
a6b2f353
GS
2656
2657 use re 'eval'; # throw caution out the door
2658 $bar = 5;
2659 $pat = '(?{ 1 })';
a6b2f353 2660 /foo${pat}bar/; # compiles ok
47f9c88b 2661
7638d2dc 2662Another form of code expression is the I<pattern code expression>.
47f9c88b
GS
2663The pattern code expression is like a regular code expression, except
2664that the result of the code evaluation is treated as a regular
2665expression and matched immediately. A simple example is
2666
2667 $length = 5;
2668 $char = 'a';
2669 $x = 'aaaaabb';
2670 $x =~ /(??{$char x $length})/x; # matches, there are 5 of 'a'
2671
2672
2673This final example contains both ordinary and pattern code
7638d2dc 2674expressions. It detects whether a binary string C<1101010010001...> has a
47f9c88b
GS
2675Fibonacci spacing 0,1,1,2,3,5,... of the C<1>'s:
2676
47f9c88b 2677 $x = "1101010010001000001";
7638d2dc 2678 $z0 = ''; $z1 = '0'; # initial conditions
47f9c88b
GS
2679 print "It is a Fibonacci sequence\n"
2680 if $x =~ /^1 # match an initial '1'
7638d2dc
WL
2681 (?:
2682 ((??{ $z0 })) # match some '0'
2683 1 # and then a '1'
2684 (?{ $z0 = $z1; $z1 .= $^N; })
47f9c88b
GS
2685 )+ # repeat as needed
2686 $ # that is all there is
2687 /x;
7638d2dc 2688 printf "Largest sequence matched was %d\n", length($z1)-length($z0);
47f9c88b 2689
7638d2dc
WL
2690Remember that C<$^N> is set to whatever was matched by the last
2691completed capture group. This prints
47f9c88b
GS
2692
2693 It is a Fibonacci sequence
2694 Largest sequence matched was 5
2695
2696Ha! Try that with your garden variety regexp package...
2697
7638d2dc 2698Note that the variables C<$z0> and C<$z1> are not substituted when the
47f9c88b 2699regexp is compiled, as happens for ordinary variables outside a code
e128ab2c
DM
2700expression. Rather, the whole code block is parsed as perl code at the
2701same time as perl is compiling the code containing the literal regexp
2702pattern.
47f9c88b
GS
2703
2704The regexp without the C<//x> modifier is
2705
7638d2dc
WL
2706 /^1(?:((??{ $z0 }))1(?{ $z0 = $z1; $z1 .= $^N; }))+$/
2707
2708which shows that spaces are still possible in the code parts. Nevertheless,
353c6505 2709when working with code and conditional expressions, the extended form of
7638d2dc
WL
2710regexps is almost necessary in creating and debugging regexps.
2711
2712
2713=head2 Backtracking control verbs
2714
2715Perl 5.10 introduced a number of control verbs intended to provide
2716detailed control over the backtracking process, by directly influencing
2717the regexp engine and by providing monitoring techniques. As all
2718the features in this group are experimental and subject to change or
2719removal in a future version of Perl, the interested reader is
2720referred to L<perlre/"Special Backtracking Control Verbs"> for a
2721detailed description.
2722
2723Below is just one example, illustrating the control verb C<(*FAIL)>,
2724which may be abbreviated as C<(*F)>. If this is inserted in a regexp
6b3ddc02
FC
2725it will cause it to fail, just as it would at some
2726mismatch between the pattern and the string. Processing
2727of the regexp continues as it would after any "normal"
353c6505
DL
2728failure, so that, for instance, the next position in the string or another
2729alternative will be tried. As failing to match doesn't preserve capture
c27a5cfe 2730groups or produce results, it may be necessary to use this in
7638d2dc
WL
2731combination with embedded code.
2732
2733 %count = ();
b539c2c9 2734 "supercalifragilisticexpialidocious" =~
c2e2285d 2735 /([aeiou])(?{ $count{$1}++; })(*FAIL)/i;
7638d2dc
WL
2736 printf "%3d '%s'\n", $count{$_}, $_ for (sort keys %count);
2737
353c6505
DL
2738The pattern begins with a class matching a subset of letters. Whenever
2739this matches, a statement like C<$count{'a'}++;> is executed, incrementing
2740the letter's counter. Then C<(*FAIL)> does what it says, and
6b3ddc02
FC
2741the regexp engine proceeds according to the book: as long as the end of
2742the string hasn't been reached, the position is advanced before looking
7638d2dc 2743for another vowel. Thus, match or no match makes no difference, and the
e1020413 2744regexp engine proceeds until the entire string has been inspected.
7638d2dc
WL
2745(It's remarkable that an alternative solution using something like
2746
b539c2c9 2747 $count{lc($_)}++ for split('', "supercalifragilisticexpialidocious");
7638d2dc
WL
2748 printf "%3d '%s'\n", $count2{$_}, $_ for ( qw{ a e i o u } );
2749
2750is considerably slower.)
47f9c88b 2751
47f9c88b
GS
2752
2753=head2 Pragmas and debugging
2754
2755Speaking of debugging, there are several pragmas available to control
2756and debug regexps in Perl. We have already encountered one pragma in
7638d2dc 2757the previous section, S<C<use re 'eval';>>, that allows variable
a6b2f353
GS
2758interpolation and code expressions to coexist in a regexp. The other
2759pragmas are
47f9c88b
GS
2760
2761 use re 'taint';
2762 $tainted = <>;
2763 @parts = ($tainted =~ /(\w+)\s+(\w+)/; # @parts is now tainted
2764
2765The C<taint> pragma causes any substrings from a match with a tainted
2766variable to be tainted as well. This is not normally the case, as
2767regexps are often used to extract the safe bits from a tainted
2768variable. Use C<taint> when you are not extracting safe bits, but are
2769performing some other processing. Both C<taint> and C<eval> pragmas
a6b2f353 2770are lexically scoped, which means they are in effect only until
47f9c88b
GS
2771the end of the block enclosing the pragmas.
2772
511eb430
FC
2773 use re '/m'; # or any other flags
2774 $multiline_string =~ /^foo/; # /m is implied
2775
9fa86798
FC
2776The C<re '/flags'> pragma (introduced in Perl
27775.14) turns on the given regular expression flags
3fd67154
KW
2778until the end of the lexical scope. See
2779L<re/"'E<sol>flags' mode"> for more
511eb430
FC
2780detail.
2781
47f9c88b
GS
2782 use re 'debug';
2783 /^(.*)$/s; # output debugging info
2784
2785 use re 'debugcolor';
2786 /^(.*)$/s; # output debugging info in living color
2787
2788The global C<debug> and C<debugcolor> pragmas allow one to get
2789detailed debugging info about regexp compilation and
2790execution. C<debugcolor> is the same as debug, except the debugging
2791information is displayed in color on terminals that can display
2792termcap color sequences. Here is example output:
2793
2794 % perl -e 'use re "debug"; "abc" =~ /a*b+c/;'
ccf3535a 2795 Compiling REx 'a*b+c'
47f9c88b
GS
2796 size 9 first at 1
2797 1: STAR(4)
2798 2: EXACT <a>(0)
2799 4: PLUS(7)
2800 5: EXACT <b>(0)
2801 7: EXACT <c>(9)
2802 9: END(0)
ccf3535a
JK
2803 floating 'bc' at 0..2147483647 (checking floating) minlen 2
2804 Guessing start of match, REx 'a*b+c' against 'abc'...
2805 Found floating substr 'bc' at offset 1...
47f9c88b 2806 Guessed: match at offset 0
ccf3535a 2807 Matching REx 'a*b+c' against 'abc'
47f9c88b
GS
2808 Setting an EVAL scope, savestack=3
2809 0 <> <abc> | 1: STAR
2810 EXACT <a> can match 1 times out of 32767...
2811 Setting an EVAL scope, savestack=3
2812 1 <a> <bc> | 4: PLUS
2813 EXACT <b> can match 1 times out of 32767...
2814 Setting an EVAL scope, savestack=3
2815 2 <ab> <c> | 7: EXACT <c>
2816 3 <abc> <> | 9: END
2817 Match successful!
ccf3535a 2818 Freeing REx: 'a*b+c'
47f9c88b
GS
2819
2820If you have gotten this far into the tutorial, you can probably guess
2821what the different parts of the debugging output tell you. The first
2822part
2823
ccf3535a 2824 Compiling REx 'a*b+c'
47f9c88b
GS
2825 size 9 first at 1
2826 1: STAR(4)
2827 2: EXACT <a>(0)
2828 4: PLUS(7)
2829 5: EXACT <b>(0)
2830 7: EXACT <c>(9)
2831 9: END(0)
2832
2833describes the compilation stage. C<STAR(4)> means that there is a
2834starred object, in this case C<'a'>, and if it matches, goto line 4,
2835i.e., C<PLUS(7)>. The middle lines describe some heuristics and
2836optimizations performed before a match:
2837
ccf3535a
JK
2838 floating 'bc' at 0..2147483647 (checking floating) minlen 2
2839 Guessing start of match, REx 'a*b+c' against 'abc'...
2840 Found floating substr 'bc' at offset 1...
47f9c88b
GS
2841 Guessed: match at offset 0
2842
2843Then the match is executed and the remaining lines describe the
2844process:
2845
ccf3535a 2846 Matching REx 'a*b+c' against 'abc'
47f9c88b
GS
2847 Setting an EVAL scope, savestack=3
2848 0 <> <abc> | 1: STAR
2849 EXACT <a> can match 1 times out of 32767...
2850 Setting an EVAL scope, savestack=3
2851 1 <a> <bc> | 4: PLUS
2852 EXACT <b> can match 1 times out of 32767...
2853 Setting an EVAL scope, savestack=3
2854 2 <ab> <c> | 7: EXACT <c>
2855 3 <abc> <> | 9: END
2856 Match successful!
ccf3535a 2857 Freeing REx: 'a*b+c'
47f9c88b 2858
7638d2dc 2859Each step is of the form S<C<< n <x> <y> >>>, with C<< <x> >> the
47f9c88b 2860part of the string matched and C<< <y> >> the part not yet
7638d2dc 2861matched. The S<C<< | 1: STAR >>> says that Perl is at line number 1
39b6ec1a 2862in the compilation list above. See
d9f2b251 2863L<perldebguts/"Debugging Regular Expressions"> for much more detail.
47f9c88b
GS
2864
2865An alternative method of debugging regexps is to embed C<print>
2866statements within the regexp. This provides a blow-by-blow account of
2867the backtracking in an alternation:
2868
2869 "that this" =~ m@(?{print "Start at position ", pos, "\n";})
2870 t(?{print "t1\n";})
2871 h(?{print "h1\n";})
2872 i(?{print "i1\n";})
2873 s(?{print "s1\n";})
2874 |
2875 t(?{print "t2\n";})
2876 h(?{print "h2\n";})
2877 a(?{print "a2\n";})
2878 t(?{print "t2\n";})
2879 (?{print "Done at position ", pos, "\n";})
2880 @x;
2881
2882prints
2883
2884 Start at position 0
2885 t1
2886 h1
2887 t2
2888 h2
2889 a2
2890 t2
2891 Done at position 4
2892
2893=head1 BUGS
2894
2895Code expressions, conditional expressions, and independent expressions
7638d2dc 2896are I<experimental>. Don't use them in production code. Yet.
47f9c88b
GS
2897
2898=head1 SEE ALSO
2899
7638d2dc 2900This is just a tutorial. For the full story on Perl regular
47f9c88b
GS
2901expressions, see the L<perlre> regular expressions reference page.
2902
2903For more information on the matching C<m//> and substitution C<s///>
2904operators, see L<perlop/"Regexp Quote-Like Operators">. For
2905information on the C<split> operation, see L<perlfunc/split>.
2906
2907For an excellent all-around resource on the care and feeding of
2908regular expressions, see the book I<Mastering Regular Expressions> by
2909Jeffrey Friedl (published by O'Reilly, ISBN 1556592-257-3).
2910
2911=head1 AUTHOR AND COPYRIGHT
2912
2913Copyright (c) 2000 Mark Kvale
2914All rights reserved.
2915
2916This document may be distributed under the same terms as Perl itself.
2917
2918=head2 Acknowledgments
2919
2920The inspiration for the stop codon DNA example came from the ZIP
2921code example in chapter 7 of I<Mastering Regular Expressions>.
2922
a6b2f353
GS
2923The author would like to thank Jeff Pinyan, Andrew Johnson, Peter
2924Haworth, Ronald J Kimball, and Joe Smith for all their helpful
2925comments.
47f9c88b
GS
2926
2927=cut
a6b2f353 2928