Commit | Line | Data |
68dc0745 |
1 | =head1 NAME |
2 | |
e573f903 |
3 | perlfaq6 - Regular Expressions ($Revision: 7910 $) |
68dc0745 |
4 | |
5 | =head1 DESCRIPTION |
6 | |
7 | This section is surprisingly small because the rest of the FAQ is |
8 | littered with answers involving regular expressions. For example, |
9 | decoding a URL and checking whether something is a number are handled |
10 | with regular expressions, but those answers are found elsewhere in |
b432a672 |
11 | this document (in L<perlfaq9>: "How do I decode or create those %-encodings |
12 | on the web" and L<perlfaq4>: "How do I determine whether a scalar is |
13 | a number/whole/integer/float", to be precise). |
68dc0745 |
14 | |
54310121 |
15 | =head2 How can I hope to use regular expressions without creating illegible and unmaintainable code? |
d74e8afc |
16 | X<regex, legibility> X<regexp, legibility> |
17 | X<regular expression, legibility> X</x> |
68dc0745 |
18 | |
19 | Three techniques can make regular expressions maintainable and |
20 | understandable. |
21 | |
22 | =over 4 |
23 | |
d92eb7b0 |
24 | =item Comments Outside the Regex |
68dc0745 |
25 | |
26 | Describe what you're doing and how you're doing it, using normal Perl |
27 | comments. |
28 | |
ac9dac7f |
29 | # turn the line into the first word, a colon, and the |
30 | # number of characters on the rest of the line |
31 | s/^(\w+)(.*)/ lc($1) . ":" . length($2) /meg; |
68dc0745 |
32 | |
d92eb7b0 |
33 | =item Comments Inside the Regex |
68dc0745 |
34 | |
d92eb7b0 |
35 | The C</x> modifier causes whitespace to be ignored in a regex pattern |
68dc0745 |
36 | (except in a character class), and also allows you to use normal |
37 | comments there, too. As you can imagine, whitespace and comments help |
38 | a lot. |
39 | |
40 | C</x> lets you turn this: |
41 | |
ac9dac7f |
42 | s{<(?:[^>'"]*|".*?"|'.*?')+>}{}gs; |
68dc0745 |
43 | |
44 | into this: |
45 | |
ac9dac7f |
46 | s{ < # opening angle bracket |
47 | (?: # Non-backreffing grouping paren |
48 | [^>'"] * # 0 or more things that are neither > nor ' nor " |
49 | | # or else |
50 | ".*?" # a section between double quotes (stingy match) |
51 | | # or else |
52 | '.*?' # a section between single quotes (stingy match) |
53 | ) + # all occurring one or more times |
54 | > # closing angle bracket |
55 | }{}gsx; # replace with nothing, i.e. delete |
68dc0745 |
56 | |
57 | It's still not quite so clear as prose, but it is very useful for |
58 | describing the meaning of each part of the pattern. |
59 | |
60 | =item Different Delimiters |
61 | |
62 | While we normally think of patterns as being delimited with C</> |
63 | characters, they can be delimited by almost any character. L<perlre> |
64 | describes this. For example, the C<s///> above uses braces as |
65 | delimiters. Selecting another delimiter can avoid quoting the |
66 | delimiter within the pattern: |
67 | |
ac9dac7f |
68 | s/\/usr\/local/\/usr\/share/g; # bad delimiter choice |
69 | s#/usr/local#/usr/share#g; # better |
68dc0745 |
70 | |
71 | =back |
72 | |
73 | =head2 I'm having trouble matching over more than one line. What's wrong? |
d74e8afc |
74 | X<regex, multiline> X<regexp, multiline> X<regular expression, multiline> |
68dc0745 |
75 | |
3392b9ec |
76 | Either you don't have more than one line in the string you're looking |
77 | at (probably), or else you aren't using the correct modifier(s) on |
78 | your pattern (possibly). |
68dc0745 |
79 | |
80 | There are many ways to get multiline data into a string. If you want |
81 | it to happen automatically while reading input, you'll want to set $/ |
82 | (probably to '' for paragraphs or C<undef> for the whole file) to |
83 | allow you to read more than one line at a time. |
84 | |
85 | Read L<perlre> to help you decide which of C</s> and C</m> (or both) |
86 | you might want to use: C</s> allows dot to include newline, and C</m> |
87 | allows caret and dollar to match next to a newline, not just at the |
88 | end of the string. You do need to make sure that you've actually |
89 | got a multiline string in there. |
90 | |
91 | For example, this program detects duplicate words, even when they span |
92 | line breaks (but not paragraph ones). For this example, we don't need |
93 | C</s> because we aren't using dot in a regular expression that we want |
94 | to cross line boundaries. Neither do we need C</m> because we aren't |
95 | wanting caret or dollar to match at any point inside the record next |
96 | to newlines. But it's imperative that $/ be set to something other |
97 | than the default, or else we won't actually ever have a multiline |
98 | record read in. |
99 | |
ac9dac7f |
100 | $/ = ''; # read in more whole paragraph, not just one line |
101 | while ( <> ) { |
102 | while ( /\b([\w'-]+)(\s+\1)+\b/gi ) { # word starts alpha |
103 | print "Duplicate $1 at paragraph $.\n"; |
104 | } |
54310121 |
105 | } |
68dc0745 |
106 | |
107 | Here's code that finds sentences that begin with "From " (which would |
108 | be mangled by many mailers): |
109 | |
ac9dac7f |
110 | $/ = ''; # read in more whole paragraph, not just one line |
111 | while ( <> ) { |
112 | while ( /^From /gm ) { # /m makes ^ match next to \n |
113 | print "leading from in paragraph $.\n"; |
114 | } |
68dc0745 |
115 | } |
68dc0745 |
116 | |
117 | Here's code that finds everything between START and END in a paragraph: |
118 | |
ac9dac7f |
119 | undef $/; # read in whole file, not just one line or paragraph |
120 | while ( <> ) { |
121 | while ( /START(.*?)END/sgm ) { # /s makes . cross line boundaries |
122 | print "$1\n"; |
123 | } |
68dc0745 |
124 | } |
68dc0745 |
125 | |
126 | =head2 How can I pull out lines between two patterns that are themselves on different lines? |
d74e8afc |
127 | X<..> |
68dc0745 |
128 | |
129 | You can use Perl's somewhat exotic C<..> operator (documented in |
130 | L<perlop>): |
131 | |
ac9dac7f |
132 | perl -ne 'print if /START/ .. /END/' file1 file2 ... |
68dc0745 |
133 | |
134 | If you wanted text and not lines, you would use |
135 | |
ac9dac7f |
136 | perl -0777 -ne 'print "$1\n" while /START(.*?)END/gs' file1 file2 ... |
68dc0745 |
137 | |
138 | But if you want nested occurrences of C<START> through C<END>, you'll |
139 | run up against the problem described in the question in this section |
140 | on matching balanced text. |
141 | |
5a964f20 |
142 | Here's another example of using C<..>: |
143 | |
ac9dac7f |
144 | while (<>) { |
145 | $in_header = 1 .. /^$/; |
e573f903 |
146 | $in_body = /^$/ .. eof; |
5a964f20 |
147 | # now choose between them |
ac9dac7f |
148 | } continue { |
e573f903 |
149 | $. = 0 if eof; # fix $. |
ac9dac7f |
150 | } |
5a964f20 |
151 | |
68dc0745 |
152 | =head2 I put a regular expression into $/ but it didn't work. What's wrong? |
d74e8afc |
153 | X<$/, regexes in> X<$INPUT_RECORD_SEPARATOR, regexes in> |
154 | X<$RS, regexes in> |
68dc0745 |
155 | |
197aec24 |
156 | Up to Perl 5.8.0, $/ has to be a string. This may change in 5.10, |
49d635f9 |
157 | but don't get your hopes up. Until then, you can use these examples |
158 | if you really need to do this. |
159 | |
28b41a80 |
160 | If you have File::Stream, this is easy. |
161 | |
ac9dac7f |
162 | use File::Stream; |
163 | |
164 | my $stream = File::Stream->new( |
165 | $filehandle, |
166 | separator => qr/\s*,\s*/, |
167 | ); |
28b41a80 |
168 | |
ac9dac7f |
169 | print "$_\n" while <$stream>; |
28b41a80 |
170 | |
171 | If you don't have File::Stream, you have to do a little more work. |
172 | |
173 | You can use the four argument form of sysread to continually add to |
197aec24 |
174 | a buffer. After you add to the buffer, you check if you have a |
49d635f9 |
175 | complete line (using your regular expression). |
176 | |
ac9dac7f |
177 | local $_ = ""; |
178 | while( sysread FH, $_, 8192, length ) { |
179 | while( s/^((?s).*?)your_pattern/ ) { |
180 | my $record = $1; |
181 | # do stuff here. |
182 | } |
183 | } |
197aec24 |
184 | |
49d635f9 |
185 | You can do the same thing with foreach and a match using the |
186 | c flag and the \G anchor, if you do not mind your entire file |
187 | being in memory at the end. |
197aec24 |
188 | |
ac9dac7f |
189 | local $_ = ""; |
190 | while( sysread FH, $_, 8192, length ) { |
191 | foreach my $record ( m/\G((?s).*?)your_pattern/gc ) { |
192 | # do stuff here. |
193 | } |
194 | substr( $_, 0, pos ) = "" if pos; |
195 | } |
68dc0745 |
196 | |
3fe9a6f1 |
197 | |
a6dd486b |
198 | =head2 How do I substitute case insensitively on the LHS while preserving case on the RHS? |
d74e8afc |
199 | X<replace, case preserving> X<substitute, case preserving> |
200 | X<substitution, case preserving> X<s, case preserving> |
68dc0745 |
201 | |
d92eb7b0 |
202 | Here's a lovely Perlish solution by Larry Rosler. It exploits |
203 | properties of bitwise xor on ASCII strings. |
204 | |
ac9dac7f |
205 | $_= "this is a TEsT case"; |
d92eb7b0 |
206 | |
ac9dac7f |
207 | $old = 'test'; |
208 | $new = 'success'; |
d92eb7b0 |
209 | |
ac9dac7f |
210 | s{(\Q$old\E)} |
211 | { uc $new | (uc $1 ^ $1) . |
212 | (uc(substr $1, -1) ^ substr $1, -1) x |
213 | (length($new) - length $1) |
214 | }egi; |
d92eb7b0 |
215 | |
ac9dac7f |
216 | print; |
d92eb7b0 |
217 | |
8305e449 |
218 | And here it is as a subroutine, modeled after the above: |
d92eb7b0 |
219 | |
ac9dac7f |
220 | sub preserve_case($$) { |
221 | my ($old, $new) = @_; |
222 | my $mask = uc $old ^ $old; |
d92eb7b0 |
223 | |
ac9dac7f |
224 | uc $new | $mask . |
225 | substr($mask, -1) x (length($new) - length($old)) |
d92eb7b0 |
226 | } |
227 | |
ac9dac7f |
228 | $a = "this is a TEsT case"; |
229 | $a =~ s/(test)/preserve_case($1, "success")/egi; |
230 | print "$a\n"; |
d92eb7b0 |
231 | |
232 | This prints: |
233 | |
ac9dac7f |
234 | this is a SUcCESS case |
d92eb7b0 |
235 | |
74b9445a |
236 | As an alternative, to keep the case of the replacement word if it is |
237 | longer than the original, you can use this code, by Jeff Pinyan: |
238 | |
ac9dac7f |
239 | sub preserve_case { |
240 | my ($from, $to) = @_; |
241 | my ($lf, $lt) = map length, @_; |
7207e29d |
242 | |
ac9dac7f |
243 | if ($lt < $lf) { $from = substr $from, 0, $lt } |
244 | else { $from .= substr $to, $lf } |
7207e29d |
245 | |
ac9dac7f |
246 | return uc $to | ($from ^ uc $from); |
247 | } |
74b9445a |
248 | |
249 | This changes the sentence to "this is a SUcCess case." |
250 | |
d92eb7b0 |
251 | Just to show that C programmers can write C in any programming language, |
252 | if you prefer a more C-like solution, the following script makes the |
253 | substitution have the same case, letter by letter, as the original. |
254 | (It also happens to run about 240% slower than the Perlish solution runs.) |
255 | If the substitution has more characters than the string being substituted, |
256 | the case of the last character is used for the rest of the substitution. |
68dc0745 |
257 | |
ac9dac7f |
258 | # Original by Nathan Torkington, massaged by Jeffrey Friedl |
259 | # |
260 | sub preserve_case($$) |
261 | { |
262 | my ($old, $new) = @_; |
263 | my ($state) = 0; # 0 = no change; 1 = lc; 2 = uc |
264 | my ($i, $oldlen, $newlen, $c) = (0, length($old), length($new)); |
265 | my ($len) = $oldlen < $newlen ? $oldlen : $newlen; |
266 | |
267 | for ($i = 0; $i < $len; $i++) { |
268 | if ($c = substr($old, $i, 1), $c =~ /[\W\d_]/) { |
269 | $state = 0; |
270 | } elsif (lc $c eq $c) { |
271 | substr($new, $i, 1) = lc(substr($new, $i, 1)); |
272 | $state = 1; |
273 | } else { |
274 | substr($new, $i, 1) = uc(substr($new, $i, 1)); |
275 | $state = 2; |
276 | } |
277 | } |
278 | # finish up with any remaining new (for when new is longer than old) |
279 | if ($newlen > $oldlen) { |
280 | if ($state == 1) { |
281 | substr($new, $oldlen) = lc(substr($new, $oldlen)); |
282 | } elsif ($state == 2) { |
283 | substr($new, $oldlen) = uc(substr($new, $oldlen)); |
284 | } |
285 | } |
286 | return $new; |
287 | } |
68dc0745 |
288 | |
5a964f20 |
289 | =head2 How can I make C<\w> match national character sets? |
d74e8afc |
290 | X<\w> |
68dc0745 |
291 | |
49d635f9 |
292 | Put C<use locale;> in your script. The \w character class is taken |
293 | from the current locale. |
294 | |
295 | See L<perllocale> for details. |
68dc0745 |
296 | |
297 | =head2 How can I match a locale-smart version of C</[a-zA-Z]/>? |
d74e8afc |
298 | X<alpha> |
68dc0745 |
299 | |
49d635f9 |
300 | You can use the POSIX character class syntax C</[[:alpha:]]/> |
301 | documented in L<perlre>. |
302 | |
303 | No matter which locale you are in, the alphabetic characters are |
304 | the characters in \w without the digits and the underscore. |
305 | As a regex, that looks like C</[^\W\d_]/>. Its complement, |
197aec24 |
306 | the non-alphabetics, is then everything in \W along with |
307 | the digits and the underscore, or C</[\W\d_]/>. |
68dc0745 |
308 | |
d92eb7b0 |
309 | =head2 How can I quote a variable to use in a regex? |
d74e8afc |
310 | X<regex, escaping> X<regexp, escaping> X<regular expression, escaping> |
68dc0745 |
311 | |
312 | The Perl parser will expand $variable and @variable references in |
313 | regular expressions unless the delimiter is a single quote. Remember, |
79a522f5 |
314 | too, that the right-hand side of a C<s///> substitution is considered |
68dc0745 |
315 | a double-quoted string (see L<perlop> for more details). Remember |
d92eb7b0 |
316 | also that any regex special characters will be acted on unless you |
68dc0745 |
317 | precede the substitution with \Q. Here's an example: |
318 | |
ac9dac7f |
319 | $string = "Placido P. Octopus"; |
320 | $regex = "P."; |
68dc0745 |
321 | |
ac9dac7f |
322 | $string =~ s/$regex/Polyp/; |
323 | # $string is now "Polypacido P. Octopus" |
68dc0745 |
324 | |
c83084d1 |
325 | Because C<.> is special in regular expressions, and can match any |
326 | single character, the regex C<P.> here has matched the <Pl> in the |
327 | original string. |
328 | |
329 | To escape the special meaning of C<.>, we use C<\Q>: |
330 | |
ac9dac7f |
331 | $string = "Placido P. Octopus"; |
332 | $regex = "P."; |
c83084d1 |
333 | |
ac9dac7f |
334 | $string =~ s/\Q$regex/Polyp/; |
335 | # $string is now "Placido Polyp Octopus" |
c83084d1 |
336 | |
337 | The use of C<\Q> causes the <.> in the regex to be treated as a |
338 | regular character, so that C<P.> matches a C<P> followed by a dot. |
68dc0745 |
339 | |
340 | =head2 What is C</o> really for? |
d74e8afc |
341 | X</o> |
68dc0745 |
342 | |
46fc3d4c |
343 | Using a variable in a regular expression match forces a re-evaluation |
a6dd486b |
344 | (and perhaps recompilation) each time the regular expression is |
345 | encountered. The C</o> modifier locks in the regex the first time |
346 | it's used. This always happens in a constant regular expression, and |
347 | in fact, the pattern was compiled into the internal format at the same |
348 | time your entire program was. |
68dc0745 |
349 | |
350 | Use of C</o> is irrelevant unless variable interpolation is used in |
d92eb7b0 |
351 | the pattern, and if so, the regex engine will neither know nor care |
68dc0745 |
352 | whether the variables change after the pattern is evaluated the I<very |
353 | first> time. |
354 | |
355 | C</o> is often used to gain an extra measure of efficiency by not |
356 | performing subsequent evaluations when you know it won't matter |
357 | (because you know the variables won't change), or more rarely, when |
d92eb7b0 |
358 | you don't want the regex to notice if they do. |
68dc0745 |
359 | |
360 | For example, here's a "paragrep" program: |
361 | |
ac9dac7f |
362 | $/ = ''; # paragraph mode |
363 | $pat = shift; |
364 | while (<>) { |
365 | print if /$pat/o; |
366 | } |
68dc0745 |
367 | |
368 | =head2 How do I use a regular expression to strip C style comments from a file? |
369 | |
370 | While this actually can be done, it's much harder than you'd think. |
371 | For example, this one-liner |
372 | |
ac9dac7f |
373 | perl -0777 -pe 's{/\*.*?\*/}{}gs' foo.c |
68dc0745 |
374 | |
375 | will work in many but not all cases. You see, it's too simple-minded for |
376 | certain kinds of C programs, in particular, those with what appear to be |
377 | comments in quoted strings. For that, you'd need something like this, |
d92eb7b0 |
378 | created by Jeffrey Friedl and later modified by Fred Curtis. |
68dc0745 |
379 | |
ac9dac7f |
380 | $/ = undef; |
381 | $_ = <>; |
382 | s#/\*[^*]*\*+([^/*][^*]*\*+)*/|("(\\.|[^"\\])*"|'(\\.|[^'\\])*'|.[^/"'\\]*)#defined $2 ? $2 : ""#gse; |
383 | print; |
68dc0745 |
384 | |
385 | This could, of course, be more legibly written with the C</x> modifier, adding |
d92eb7b0 |
386 | whitespace and comments. Here it is expanded, courtesy of Fred Curtis. |
387 | |
388 | s{ |
389 | /\* ## Start of /* ... */ comment |
390 | [^*]*\*+ ## Non-* followed by 1-or-more *'s |
391 | ( |
392 | [^/*][^*]*\*+ |
393 | )* ## 0-or-more things which don't start with / |
394 | ## but do end with '*' |
395 | / ## End of /* ... */ comment |
396 | |
397 | | ## OR various things which aren't comments: |
398 | |
399 | ( |
400 | " ## Start of " ... " string |
401 | ( |
402 | \\. ## Escaped char |
403 | | ## OR |
404 | [^"\\] ## Non "\ |
405 | )* |
406 | " ## End of " ... " string |
407 | |
408 | | ## OR |
409 | |
410 | ' ## Start of ' ... ' string |
411 | ( |
412 | \\. ## Escaped char |
413 | | ## OR |
414 | [^'\\] ## Non '\ |
415 | )* |
416 | ' ## End of ' ... ' string |
417 | |
418 | | ## OR |
419 | |
420 | . ## Anything other char |
421 | [^/"'\\]* ## Chars which doesn't start a comment, string or escape |
422 | ) |
c98c5709 |
423 | }{defined $2 ? $2 : ""}gxse; |
d92eb7b0 |
424 | |
e573f903 |
425 | A slight modification also removes C++ comments, as long as they are not |
426 | spread over multiple lines using a continuation character): |
d92eb7b0 |
427 | |
ac9dac7f |
428 | s#/\*[^*]*\*+([^/*][^*]*\*+)*/|//[^\n]*|("(\\.|[^"\\])*"|'(\\.|[^'\\])*'|.[^/"'\\]*)#defined $2 ? $2 : ""#gse; |
68dc0745 |
429 | |
430 | =head2 Can I use Perl regular expressions to match balanced text? |
d74e8afc |
431 | X<regex, matching balanced test> X<regexp, matching balanced test> |
432 | X<regular expression, matching balanced test> |
68dc0745 |
433 | |
8305e449 |
434 | Historically, Perl regular expressions were not capable of matching |
435 | balanced text. As of more recent versions of perl including 5.6.1 |
436 | experimental features have been added that make it possible to do this. |
437 | Look at the documentation for the (??{ }) construct in recent perlre manual |
438 | pages to see an example of matching balanced parentheses. Be sure to take |
439 | special notice of the warnings present in the manual before making use |
440 | of this feature. |
441 | |
442 | CPAN contains many modules that can be useful for matching text |
443 | depending on the context. Damian Conway provides some useful |
444 | patterns in Regexp::Common. The module Text::Balanced provides a |
445 | general solution to this problem. |
446 | |
447 | One of the common applications of balanced text matching is working |
448 | with XML and HTML. There are many modules available that support |
449 | these needs. Two examples are HTML::Parser and XML::Parser. There |
450 | are many others. |
68dc0745 |
451 | |
452 | An elaborate subroutine (for 7-bit ASCII only) to pull out balanced |
453 | and possibly nested single chars, like C<`> and C<'>, C<{> and C<}>, |
454 | or C<(> and C<)> can be found in |
a93751fa |
455 | http://www.cpan.org/authors/id/TOMC/scripts/pull_quotes.gz . |
68dc0745 |
456 | |
8305e449 |
457 | The C::Scan module from CPAN also contains such subs for internal use, |
68dc0745 |
458 | but they are undocumented. |
459 | |
d92eb7b0 |
460 | =head2 What does it mean that regexes are greedy? How can I get around it? |
d74e8afc |
461 | X<greedy> X<greediness> |
68dc0745 |
462 | |
d92eb7b0 |
463 | Most people mean that greedy regexes match as much as they can. |
68dc0745 |
464 | Technically speaking, it's actually the quantifiers (C<?>, C<*>, C<+>, |
465 | C<{}>) that are greedy rather than the whole pattern; Perl prefers local |
466 | greed and immediate gratification to overall greed. To get non-greedy |
467 | versions of the same quantifiers, use (C<??>, C<*?>, C<+?>, C<{}?>). |
468 | |
469 | An example: |
470 | |
ac9dac7f |
471 | $s1 = $s2 = "I am very very cold"; |
472 | $s1 =~ s/ve.*y //; # I am cold |
473 | $s2 =~ s/ve.*?y //; # I am very cold |
68dc0745 |
474 | |
475 | Notice how the second substitution stopped matching as soon as it |
476 | encountered "y ". The C<*?> quantifier effectively tells the regular |
477 | expression engine to find a match as quickly as possible and pass |
478 | control on to whatever is next in line, like you would if you were |
479 | playing hot potato. |
480 | |
f9ac83b8 |
481 | =head2 How do I process each word on each line? |
d74e8afc |
482 | X<word> |
68dc0745 |
483 | |
484 | Use the split function: |
485 | |
ac9dac7f |
486 | while (<>) { |
487 | foreach $word ( split ) { |
488 | # do something with $word here |
489 | } |
197aec24 |
490 | } |
68dc0745 |
491 | |
54310121 |
492 | Note that this isn't really a word in the English sense; it's just |
493 | chunks of consecutive non-whitespace characters. |
68dc0745 |
494 | |
f1cbbd6e |
495 | To work with only alphanumeric sequences (including underscores), you |
496 | might consider |
68dc0745 |
497 | |
ac9dac7f |
498 | while (<>) { |
499 | foreach $word (m/(\w+)/g) { |
500 | # do something with $word here |
501 | } |
68dc0745 |
502 | } |
68dc0745 |
503 | |
504 | =head2 How can I print out a word-frequency or line-frequency summary? |
505 | |
506 | To do this, you have to parse out each word in the input stream. We'll |
54310121 |
507 | pretend that by word you mean chunk of alphabetics, hyphens, or |
508 | apostrophes, rather than the non-whitespace chunk idea of a word given |
68dc0745 |
509 | in the previous question: |
510 | |
ac9dac7f |
511 | while (<>) { |
512 | while ( /(\b[^\W_\d][\w'-]+\b)/g ) { # misses "`sheep'" |
513 | $seen{$1}++; |
514 | } |
54310121 |
515 | } |
ac9dac7f |
516 | |
517 | while ( ($word, $count) = each %seen ) { |
518 | print "$count $word\n"; |
519 | } |
68dc0745 |
520 | |
521 | If you wanted to do the same thing for lines, you wouldn't need a |
522 | regular expression: |
523 | |
ac9dac7f |
524 | while (<>) { |
525 | $seen{$_}++; |
526 | } |
527 | |
528 | while ( ($line, $count) = each %seen ) { |
529 | print "$count $line"; |
530 | } |
68dc0745 |
531 | |
b432a672 |
532 | If you want these output in a sorted order, see L<perlfaq4>: "How do I |
533 | sort a hash (optionally by value instead of key)?". |
68dc0745 |
534 | |
535 | =head2 How can I do approximate matching? |
d74e8afc |
536 | X<match, approximate> X<matching, approximate> |
68dc0745 |
537 | |
538 | See the module String::Approx available from CPAN. |
539 | |
540 | =head2 How do I efficiently match many regular expressions at once? |
d74e8afc |
541 | X<regex, efficiency> X<regexp, efficiency> |
542 | X<regular expression, efficiency> |
68dc0745 |
543 | |
7678cced |
544 | ( contributed by brian d foy ) |
545 | |
6670e5e7 |
546 | Avoid asking Perl to compile a regular expression every time |
7678cced |
547 | you want to match it. In this example, perl must recompile |
548 | the regular expression for every iteration of the foreach() |
549 | loop since it has no way to know what $pattern will be. |
550 | |
ac9dac7f |
551 | @patterns = qw( foo bar baz ); |
6670e5e7 |
552 | |
ac9dac7f |
553 | LINE: while( <DATA> ) |
554 | { |
6670e5e7 |
555 | foreach $pattern ( @patterns ) |
7678cced |
556 | { |
ac9dac7f |
557 | if( /\b$pattern\b/i ) |
558 | { |
559 | print; |
560 | next LINE; |
561 | } |
562 | } |
7678cced |
563 | } |
68dc0745 |
564 | |
7678cced |
565 | The qr// operator showed up in perl 5.005. It compiles a |
566 | regular expression, but doesn't apply it. When you use the |
567 | pre-compiled version of the regex, perl does less work. In |
568 | this example, I inserted a map() to turn each pattern into |
569 | its pre-compiled form. The rest of the script is the same, |
570 | but faster. |
571 | |
ac9dac7f |
572 | @patterns = map { qr/\b$_\b/i } qw( foo bar baz ); |
7678cced |
573 | |
ac9dac7f |
574 | LINE: while( <> ) |
575 | { |
6670e5e7 |
576 | foreach $pattern ( @patterns ) |
7678cced |
577 | { |
ac9dac7f |
578 | print if /\b$pattern\b/i; |
579 | next LINE; |
580 | } |
7678cced |
581 | } |
6670e5e7 |
582 | |
7678cced |
583 | In some cases, you may be able to make several patterns into |
584 | a single regular expression. Beware of situations that require |
585 | backtracking though. |
65acb1b1 |
586 | |
7678cced |
587 | $regex = join '|', qw( foo bar baz ); |
588 | |
ac9dac7f |
589 | LINE: while( <> ) |
590 | { |
7678cced |
591 | print if /\b(?:$regex)\b/i; |
592 | } |
593 | |
594 | For more details on regular expression efficiency, see Mastering |
595 | Regular Expressions by Jeffrey Freidl. He explains how regular |
596 | expressions engine work and why some patterns are surprisingly |
6670e5e7 |
597 | inefficient. Once you understand how perl applies regular |
7678cced |
598 | expressions, you can tune them for individual situations. |
68dc0745 |
599 | |
600 | =head2 Why don't word-boundary searches with C<\b> work for me? |
d74e8afc |
601 | X<\b> |
68dc0745 |
602 | |
7678cced |
603 | (contributed by brian d foy) |
604 | |
605 | Ensure that you know what \b really does: it's the boundary between a |
606 | word character, \w, and something that isn't a word character. That |
607 | thing that isn't a word character might be \W, but it can also be the |
608 | start or end of the string. |
609 | |
610 | It's not (not!) the boundary between whitespace and non-whitespace, |
611 | and it's not the stuff between words we use to create sentences. |
612 | |
613 | In regex speak, a word boundary (\b) is a "zero width assertion", |
614 | meaning that it doesn't represent a character in the string, but a |
615 | condition at a certain position. |
616 | |
617 | For the regular expression, /\bPerl\b/, there has to be a word |
618 | boundary before the "P" and after the "l". As long as something other |
619 | than a word character precedes the "P" and succeeds the "l", the |
620 | pattern will match. These strings match /\bPerl\b/. |
621 | |
622 | "Perl" # no word char before P or after l |
623 | "Perl " # same as previous (space is not a word char) |
624 | "'Perl'" # the ' char is not a word char |
625 | "Perl's" # no word char before P, non-word char after "l" |
626 | |
627 | These strings do not match /\bPerl\b/. |
628 | |
629 | "Perl_" # _ is a word char! |
630 | "Perler" # no word char before P, but one after l |
6670e5e7 |
631 | |
7678cced |
632 | You don't have to use \b to match words though. You can look for |
d7f8936a |
633 | non-word characters surrounded by word characters. These strings |
7678cced |
634 | match the pattern /\b'\b/. |
635 | |
636 | "don't" # the ' char is surrounded by "n" and "t" |
637 | "qep'a'" # the ' char is surrounded by "p" and "a" |
6670e5e7 |
638 | |
7678cced |
639 | These strings do not match /\b'\b/. |
68dc0745 |
640 | |
7678cced |
641 | "foo'" # there is no word char after non-word ' |
6670e5e7 |
642 | |
7678cced |
643 | You can also use the complement of \b, \B, to specify that there |
644 | should not be a word boundary. |
68dc0745 |
645 | |
7678cced |
646 | In the pattern /\Bam\B/, there must be a word character before the "a" |
647 | and after the "m". These patterns match /\Bam\B/: |
68dc0745 |
648 | |
7678cced |
649 | "llama" # "am" surrounded by word chars |
650 | "Samuel" # same |
6670e5e7 |
651 | |
7678cced |
652 | These strings do not match /\Bam\B/ |
68dc0745 |
653 | |
7678cced |
654 | "Sam" # no word boundary before "a", but one after "m" |
655 | "I am Sam" # "am" surrounded by non-word chars |
68dc0745 |
656 | |
68dc0745 |
657 | |
658 | =head2 Why does using $&, $`, or $' slow my program down? |
d74e8afc |
659 | X<$MATCH> X<$&> X<$POSTMATCH> X<$'> X<$PREMATCH> X<$`> |
68dc0745 |
660 | |
571e049f |
661 | (contributed by Anno Siegel) |
68dc0745 |
662 | |
571e049f |
663 | Once Perl sees that you need one of these variables anywhere in the |
b68463f7 |
664 | program, it provides them on each and every pattern match. That means |
665 | that on every pattern match the entire string will be copied, part of it |
666 | to $`, part to $&, and part to $'. Thus the penalty is most severe with |
667 | long strings and patterns that match often. Avoid $&, $', and $` if you |
668 | can, but if you can't, once you've used them at all, use them at will |
669 | because you've already paid the price. Remember that some algorithms |
670 | really appreciate them. As of the 5.005 release, the $& variable is no |
671 | longer "expensive" the way the other two are. |
672 | |
673 | Since Perl 5.6.1 the special variables @- and @+ can functionally replace |
674 | $`, $& and $'. These arrays contain pointers to the beginning and end |
675 | of each match (see perlvar for the full story), so they give you |
676 | essentially the same information, but without the risk of excessive |
677 | string copying. |
6670e5e7 |
678 | |
68dc0745 |
679 | =head2 What good is C<\G> in a regular expression? |
d74e8afc |
680 | X<\G> |
68dc0745 |
681 | |
49d635f9 |
682 | You use the C<\G> anchor to start the next match on the same |
683 | string where the last match left off. The regular |
684 | expression engine cannot skip over any characters to find |
685 | the next match with this anchor, so C<\G> is similar to the |
686 | beginning of string anchor, C<^>. The C<\G> anchor is typically |
687 | used with the C<g> flag. It uses the value of pos() |
688 | as the position to start the next match. As the match |
689 | operator makes successive matches, it updates pos() with the |
690 | position of the next character past the last match (or the |
691 | first character of the next match, depending on how you like |
692 | to look at it). Each string has its own pos() value. |
693 | |
694 | Suppose you want to match all of consective pairs of digits |
695 | in a string like "1122a44" and stop matching when you |
696 | encounter non-digits. You want to match C<11> and C<22> but |
697 | the letter <a> shows up between C<22> and C<44> and you want |
698 | to stop at C<a>. Simply matching pairs of digits skips over |
699 | the C<a> and still matches C<44>. |
700 | |
701 | $_ = "1122a44"; |
702 | my @pairs = m/(\d\d)/g; # qw( 11 22 44 ) |
703 | |
704 | If you use the \G anchor, you force the match after C<22> to |
705 | start with the C<a>. The regular expression cannot match |
706 | there since it does not find a digit, so the next match |
707 | fails and the match operator returns the pairs it already |
708 | found. |
709 | |
710 | $_ = "1122a44"; |
711 | my @pairs = m/\G(\d\d)/g; # qw( 11 22 ) |
712 | |
713 | You can also use the C<\G> anchor in scalar context. You |
714 | still need the C<g> flag. |
715 | |
716 | $_ = "1122a44"; |
717 | while( m/\G(\d\d)/g ) |
718 | { |
719 | print "Found $1\n"; |
720 | } |
197aec24 |
721 | |
49d635f9 |
722 | After the match fails at the letter C<a>, perl resets pos() |
723 | and the next match on the same string starts at the beginning. |
724 | |
725 | $_ = "1122a44"; |
726 | while( m/\G(\d\d)/g ) |
727 | { |
728 | print "Found $1\n"; |
729 | } |
730 | |
731 | print "Found $1 after while" if m/(\d\d)/g; # finds "11" |
732 | |
733 | You can disable pos() resets on fail with the C<c> flag. |
734 | Subsequent matches start where the last successful match |
735 | ended (the value of pos()) even if a match on the same |
736 | string as failed in the meantime. In this case, the match |
737 | after the while() loop starts at the C<a> (where the last |
738 | match stopped), and since it does not use any anchor it can |
739 | skip over the C<a> to find "44". |
740 | |
741 | $_ = "1122a44"; |
742 | while( m/\G(\d\d)/gc ) |
743 | { |
744 | print "Found $1\n"; |
745 | } |
746 | |
747 | print "Found $1 after while" if m/(\d\d)/g; # finds "44" |
748 | |
749 | Typically you use the C<\G> anchor with the C<c> flag |
750 | when you want to try a different match if one fails, |
751 | such as in a tokenizer. Jeffrey Friedl offers this example |
752 | which works in 5.004 or later. |
68dc0745 |
753 | |
ac9dac7f |
754 | while (<>) { |
755 | chomp; |
756 | PARSER: { |
757 | m/ \G( \d+\b )/gcx && do { print "number: $1\n"; redo; }; |
758 | m/ \G( \w+ )/gcx && do { print "word: $1\n"; redo; }; |
759 | m/ \G( \s+ )/gcx && do { print "space: $1\n"; redo; }; |
760 | m/ \G( [^\w\d]+ )/gcx && do { print "other: $1\n"; redo; }; |
761 | } |
762 | } |
68dc0745 |
763 | |
49d635f9 |
764 | For each line, the PARSER loop first tries to match a series |
765 | of digits followed by a word boundary. This match has to |
766 | start at the place the last match left off (or the beginning |
197aec24 |
767 | of the string on the first match). Since C<m/ \G( \d+\b |
49d635f9 |
768 | )/gcx> uses the C<c> flag, if the string does not match that |
769 | regular expression, perl does not reset pos() and the next |
770 | match starts at the same position to try a different |
771 | pattern. |
68dc0745 |
772 | |
d92eb7b0 |
773 | =head2 Are Perl regexes DFAs or NFAs? Are they POSIX compliant? |
d74e8afc |
774 | X<DFA> X<NFA> X<POSIX> |
68dc0745 |
775 | |
776 | While it's true that Perl's regular expressions resemble the DFAs |
777 | (deterministic finite automata) of the egrep(1) program, they are in |
46fc3d4c |
778 | fact implemented as NFAs (non-deterministic finite automata) to allow |
68dc0745 |
779 | backtracking and backreferencing. And they aren't POSIX-style either, |
780 | because those guarantee worst-case behavior for all cases. (It seems |
781 | that some people prefer guarantees of consistency, even when what's |
782 | guaranteed is slowness.) See the book "Mastering Regular Expressions" |
783 | (from O'Reilly) by Jeffrey Friedl for all the details you could ever |
784 | hope to know on these matters (a full citation appears in |
785 | L<perlfaq2>). |
786 | |
788611b6 |
787 | =head2 What's wrong with using grep in a void context? |
d74e8afc |
788 | X<grep> |
68dc0745 |
789 | |
788611b6 |
790 | The problem is that grep builds a return list, regardless of the context. |
791 | This means you're making Perl go to the trouble of building a list that |
792 | you then just throw away. If the list is large, you waste both time and space. |
793 | If your intent is to iterate over the list, then use a for loop for this |
f05bbc40 |
794 | purpose. |
68dc0745 |
795 | |
788611b6 |
796 | In perls older than 5.8.1, map suffers from this problem as well. |
797 | But since 5.8.1, this has been fixed, and map is context aware - in void |
798 | context, no lists are constructed. |
799 | |
54310121 |
800 | =head2 How can I match strings with multibyte characters? |
d74e8afc |
801 | X<regex, and multibyte characters> X<regexp, and multibyte characters> |
ac9dac7f |
802 | X<regular expression, and multibyte characters> X<martian> X<encoding, Martian> |
68dc0745 |
803 | |
d9d154f2 |
804 | Starting from Perl 5.6 Perl has had some level of multibyte character |
805 | support. Perl 5.8 or later is recommended. Supported multibyte |
fe854a6f |
806 | character repertoires include Unicode, and legacy encodings |
d9d154f2 |
807 | through the Encode module. See L<perluniintro>, L<perlunicode>, |
808 | and L<Encode>. |
809 | |
810 | If you are stuck with older Perls, you can do Unicode with the |
811 | C<Unicode::String> module, and character conversions using the |
812 | C<Unicode::Map8> and C<Unicode::Map> modules. If you are using |
813 | Japanese encodings, you might try using the jperl 5.005_03. |
814 | |
815 | Finally, the following set of approaches was offered by Jeffrey |
816 | Friedl, whose article in issue #5 of The Perl Journal talks about |
817 | this very matter. |
68dc0745 |
818 | |
fc36a67e |
819 | Let's suppose you have some weird Martian encoding where pairs of |
820 | ASCII uppercase letters encode single Martian letters (i.e. the two |
821 | bytes "CV" make a single Martian letter, as do the two bytes "SG", |
822 | "VS", "XX", etc.). Other bytes represent single characters, just like |
823 | ASCII. |
68dc0745 |
824 | |
fc36a67e |
825 | So, the string of Martian "I am CVSGXX!" uses 12 bytes to encode the |
826 | nine characters 'I', ' ', 'a', 'm', ' ', 'CV', 'SG', 'XX', '!'. |
68dc0745 |
827 | |
828 | Now, say you want to search for the single character C</GX/>. Perl |
fc36a67e |
829 | doesn't know about Martian, so it'll find the two bytes "GX" in the "I |
830 | am CVSGXX!" string, even though that character isn't there: it just |
831 | looks like it is because "SG" is next to "XX", but there's no real |
832 | "GX". This is a big problem. |
68dc0745 |
833 | |
834 | Here are a few ways, all painful, to deal with it: |
835 | |
ac9dac7f |
836 | # Make sure adjacent "martian" bytes are no longer adjacent. |
837 | $martian =~ s/([A-Z][A-Z])/ $1 /g; |
838 | |
839 | print "found GX!\n" if $martian =~ /GX/; |
68dc0745 |
840 | |
841 | Or like this: |
842 | |
ac9dac7f |
843 | @chars = $martian =~ m/([A-Z][A-Z]|[^A-Z])/g; |
844 | # above is conceptually similar to: @chars = $text =~ m/(.)/g; |
845 | # |
846 | foreach $char (@chars) { |
847 | print "found GX!\n", last if $char eq 'GX'; |
848 | } |
68dc0745 |
849 | |
850 | Or like this: |
851 | |
ac9dac7f |
852 | while ($martian =~ m/\G([A-Z][A-Z]|.)/gs) { # \G probably unneeded |
853 | print "found GX!\n", last if $1 eq 'GX'; |
854 | } |
68dc0745 |
855 | |
49d635f9 |
856 | Here's another, slightly less painful, way to do it from Benjamin |
c98c5709 |
857 | Goldberg, who uses a zero-width negative look-behind assertion. |
49d635f9 |
858 | |
c98c5709 |
859 | print "found GX!\n" if $martian =~ m/ |
ac9dac7f |
860 | (?<![A-Z]) |
861 | (?:[A-Z][A-Z])*? |
862 | GX |
c98c5709 |
863 | /x; |
197aec24 |
864 | |
49d635f9 |
865 | This succeeds if the "martian" character GX is in the string, and fails |
c98c5709 |
866 | otherwise. If you don't like using (?<!), a zero-width negative |
867 | look-behind assertion, you can replace (?<![A-Z]) with (?:^|[^A-Z]). |
49d635f9 |
868 | |
869 | It does have the drawback of putting the wrong thing in $-[0] and $+[0], |
870 | but this usually can be worked around. |
68dc0745 |
871 | |
ac9dac7f |
872 | =head2 How do I match a regular expression that's in a variable? |
873 | X<regex, in variable> X<eval> X<regex> X<quotemeta> X<\Q, regex> |
874 | X<\E, regex>, X<qr//> |
65acb1b1 |
875 | |
ac9dac7f |
876 | (contributed by brian d foy) |
65acb1b1 |
877 | |
ac9dac7f |
878 | We don't have to hard-code patterns into the match operator (or |
879 | anything else that works with regular expressions). We can put the |
880 | pattern in a variable for later use. |
65acb1b1 |
881 | |
ac9dac7f |
882 | The match operator is a double quote context, so you can interpolate |
883 | your variable just like a double quoted string. In this case, you |
884 | read the regular expression as user input and store it in C<$regex>. |
885 | Once you have the pattern in C<$regex>, you use that variable in the |
886 | match operator. |
65acb1b1 |
887 | |
ac9dac7f |
888 | chomp( my $regex = <STDIN> ); |
65acb1b1 |
889 | |
ac9dac7f |
890 | if( $string =~ m/$regex/ ) { ... } |
65acb1b1 |
891 | |
ac9dac7f |
892 | Any regular expression special characters in C<$regex> are still |
893 | special, and the pattern still has to be valid or Perl will complain. |
894 | For instance, in this pattern there is an unpaired parenthesis. |
65acb1b1 |
895 | |
ac9dac7f |
896 | my $regex = "Unmatched ( paren"; |
897 | |
898 | "Two parens to bind them all" =~ m/$regex/; |
899 | |
900 | When Perl compiles the regular expression, it treats the parenthesis |
901 | as the start of a memory match. When it doesn't find the closing |
902 | parenthesis, it complains: |
903 | |
904 | Unmatched ( in regex; marked by <-- HERE in m/Unmatched ( <-- HERE paren/ at script line 3. |
905 | |
906 | You can get around this in several ways depending on our situation. |
907 | First, if you don't want any of the characters in the string to be |
908 | special, you can escape them with C<quotemeta> before you use the string. |
909 | |
910 | chomp( my $regex = <STDIN> ); |
911 | $regex = quotemeta( $regex ); |
912 | |
913 | if( $string =~ m/$regex/ ) { ... } |
914 | |
915 | You can also do this directly in the match operator using the C<\Q> |
916 | and C<\E> sequences. The C<\Q> tells Perl where to start escaping |
917 | special characters, and the C<\E> tells it where to stop (see L<perlop> |
918 | for more details). |
919 | |
920 | chomp( my $regex = <STDIN> ); |
921 | |
922 | if( $string =~ m/\Q$regex\E/ ) { ... } |
923 | |
924 | Alternately, you can use C<qr//>, the regular expression quote operator (see |
925 | L<perlop> for more details). It quotes and perhaps compiles the pattern, |
926 | and you can apply regular expression flags to the pattern. |
927 | |
928 | chomp( my $input = <STDIN> ); |
929 | |
930 | my $regex = qr/$input/is; |
931 | |
932 | $string =~ m/$regex/ # same as m/$input/is; |
933 | |
934 | You might also want to trap any errors by wrapping an C<eval> block |
935 | around the whole thing. |
936 | |
937 | chomp( my $input = <STDIN> ); |
938 | |
939 | eval { |
940 | if( $string =~ m/\Q$input\E/ ) { ... } |
941 | }; |
942 | warn $@ if $@; |
943 | |
944 | Or... |
945 | |
946 | my $regex = eval { qr/$input/is }; |
947 | if( defined $regex ) { |
948 | $string =~ m/$regex/; |
949 | } |
950 | else { |
951 | warn $@; |
952 | } |
65acb1b1 |
953 | |
500071f4 |
954 | =head1 REVISION |
955 | |
e573f903 |
956 | Revision: $Revision: 7910 $ |
500071f4 |
957 | |
e573f903 |
958 | Date: $Date: 2006-10-07 22:38:54 +0200 (sam, 07 oct 2006) $ |
500071f4 |
959 | |
960 | See L<perlfaq> for source control details and availability. |
961 | |
68dc0745 |
962 | =head1 AUTHOR AND COPYRIGHT |
963 | |
58103a2e |
964 | Copyright (c) 1997-2006 Tom Christiansen, Nathan Torkington, and |
7678cced |
965 | other authors as noted. All rights reserved. |
5a964f20 |
966 | |
5a7beb56 |
967 | This documentation is free; you can redistribute it and/or modify it |
968 | under the same terms as Perl itself. |
5a964f20 |
969 | |
970 | Irrespective of its distribution, all code examples in this file |
971 | are hereby placed into the public domain. You are permitted and |
972 | encouraged to use this code in your own programs for fun |
973 | or for profit as you see fit. A simple comment in the code giving |
974 | credit would be courteous but is not required. |