1PCREPARTIAL(3) Library Functions Manual PCREPARTIAL(3)
2
3
4
6 PCRE - Perl-compatible regular expressions
7
9
10 In normal use of PCRE, if the subject string that is passed to a match‐
11 ing function matches as far as it goes, but is too short to match the
12 entire pattern, PCRE_ERROR_NOMATCH is returned. There are circumstances
13 where it might be helpful to distinguish this case from other cases in
14 which there is no match.
15
16 Consider, for example, an application where a human is required to type
17 in data for a field with specific formatting requirements. An example
18 might be a date in the form ddmmmyy, defined by this pattern:
19
20 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
21
22 If the application sees the user's keystrokes one by one, and can check
23 that what has been typed so far is potentially valid, it is able to
24 raise an error as soon as a mistake is made, by beeping and not
25 reflecting the character that has been typed, for example. This immedi‐
26 ate feedback is likely to be a better user interface than a check that
27 is delayed until the entire string has been entered. Partial matching
28 can also be useful when the subject string is very long and is not all
29 available at once.
30
31 PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
32 PCRE_PARTIAL_HARD options, which can be set when calling any of the
33 matching functions. For backwards compatibility, PCRE_PARTIAL is a syn‐
34 onym for PCRE_PARTIAL_SOFT. The essential difference between the two
35 options is whether or not a partial match is preferred to an alterna‐
36 tive complete match, though the details differ between the two types of
37 matching function. If both options are set, PCRE_PARTIAL_HARD takes
38 precedence.
39
40 If you want to use partial matching with just-in-time optimized code,
41 you must call pcre_study(), pcre16_study() or pcre32_study() with one
42 or both of these options:
43
44 PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
45 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
46
47 PCRE_STUDY_JIT_COMPILE should also be set if you are going to run non-
48 partial matches on the same pattern. If the appropriate JIT study mode
49 has not been set for a match, the interpretive matching code is used.
50
51 Setting a partial matching option disables two of PCRE's standard opti‐
52 mizations. PCRE remembers the last literal data unit in a pattern, and
53 abandons matching immediately if it is not present in the subject
54 string. This optimization cannot be used for a subject string that
55 might match only partially. If the pattern was studied, PCRE knows the
56 minimum length of a matching string, and does not bother to run the
57 matching function on shorter strings. This optimization is also dis‐
58 abled for partial matching.
59
61
62 A partial match occurs during a call to pcre_exec() or
63 pcre[16|32]_exec() when the end of the subject string is reached suc‐
64 cessfully, but matching cannot continue because more characters are
65 needed. However, at least one character in the subject must have been
66 inspected. This character need not form part of the final matched
67 string; lookbehind assertions and the \K escape sequence provide ways
68 of inspecting characters before the start of a matched substring. The
69 requirement for inspecting at least one character exists because an
70 empty string can always be matched; without such a restriction there
71 would always be a partial match of an empty string at the end of the
72 subject.
73
74 If there are at least two slots in the offsets vector when a partial
75 match is returned, the first slot is set to the offset of the earliest
76 character that was inspected. For convenience, the second offset points
77 to the end of the subject so that a substring can easily be identified.
78
79 For the majority of patterns, the first offset identifies the start of
80 the partially matched string. However, for patterns that contain look‐
81 behind assertions, or \K, or begin with \b or \B, earlier characters
82 have been inspected while carrying out the match. For example:
83
84 /(?<=abc)123/
85
86 This pattern matches "123", but only if it is preceded by "abc". If the
87 subject string is "xyzabc12", the offsets after a partial match are for
88 the substring "abc12", because all these characters are needed if
89 another match is tried with extra characters added to the subject.
90
91 What happens when a partial match is identified depends on which of the
92 two partial matching options are set.
93
94 PCRE_PARTIAL_SOFT WITH pcre_exec() OR pcre[16|32]_exec()
95
96 If PCRE_PARTIAL_SOFT is set when pcre_exec() or pcre[16|32]_exec()
97 identifies a partial match, the partial match is remembered, but match‐
98 ing continues as normal, and other alternatives in the pattern are
99 tried. If no complete match can be found, PCRE_ERROR_PARTIAL is
100 returned instead of PCRE_ERROR_NOMATCH.
101
102 This option is "soft" because it prefers a complete match over a par‐
103 tial match. All the various matching items in a pattern behave as if
104 the subject string is potentially complete. For example, \z, \Z, and $
105 match at the end of the subject, as normal, and for \b and \B the end
106 of the subject is treated as a non-alphanumeric.
107
108 If there is more than one partial match, the first one that was found
109 provides the data that is returned. Consider this pattern:
110
111 /123\w+X|dogY/
112
113 If this is matched against the subject string "abc123dog", both alter‐
114 natives fail to match, but the end of the subject is reached during
115 matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3
116 and 9, identifying "123dog" as the first partial match that was found.
117 (In this example, there are two partial matches, because "dog" on its
118 own partially matches the second alternative.)
119
120 PCRE_PARTIAL_HARD WITH pcre_exec() OR pcre[16|32]_exec()
121
122 If PCRE_PARTIAL_HARD is set for pcre_exec() or pcre[16|32]_exec(),
123 PCRE_ERROR_PARTIAL is returned as soon as a partial match is found,
124 without continuing to search for possible complete matches. This option
125 is "hard" because it prefers an earlier partial match over a later com‐
126 plete match. For this reason, the assumption is made that the end of
127 the supplied subject string may not be the true end of the available
128 data, and so, if \z, \Z, \b, \B, or $ are encountered at the end of the
129 subject, the result is PCRE_ERROR_PARTIAL, provided that at least one
130 character in the subject has been inspected.
131
132 Setting PCRE_PARTIAL_HARD also affects the way UTF-8 and UTF-16 subject
133 strings are checked for validity. Normally, an invalid sequence causes
134 the error PCRE_ERROR_BADUTF8 or PCRE_ERROR_BADUTF16. However, in the
135 special case of a truncated character at the end of the subject,
136 PCRE_ERROR_SHORTUTF8 or PCRE_ERROR_SHORTUTF16 is returned when
137 PCRE_PARTIAL_HARD is set.
138
139 Comparing hard and soft partial matching
140
141 The difference between the two partial matching options can be illus‐
142 trated by a pattern such as:
143
144 /dog(sbody)?/
145
146 This matches either "dog" or "dogsbody", greedily (that is, it prefers
147 the longer string if possible). If it is matched against the string
148 "dog" with PCRE_PARTIAL_SOFT, it yields a complete match for "dog".
149 However, if PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL.
150 On the other hand, if the pattern is made ungreedy the result is dif‐
151 ferent:
152
153 /dog(sbody)??/
154
155 In this case the result is always a complete match because that is
156 found first, and matching never continues after finding a complete
157 match. It might be easier to follow this explanation by thinking of the
158 two patterns like this:
159
160 /dog(sbody)?/ is the same as /dogsbody|dog/
161 /dog(sbody)??/ is the same as /dog|dogsbody/
162
163 The second pattern will never match "dogsbody", because it will always
164 find the shorter match first.
165
167
168 The DFA functions move along the subject string character by character,
169 without backtracking, searching for all possible matches simultane‐
170 ously. If the end of the subject is reached before the end of the pat‐
171 tern, there is the possibility of a partial match, again provided that
172 at least one character has been inspected.
173
174 When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if
175 there have been no complete matches. Otherwise, the complete matches
176 are returned. However, if PCRE_PARTIAL_HARD is set, a partial match
177 takes precedence over any complete matches. The portion of the string
178 that was inspected when the longest partial match was found is set as
179 the first matching string, provided there are at least two slots in the
180 offsets vector.
181
182 Because the DFA functions always search for all possible matches, and
183 there is no difference between greedy and ungreedy repetition, their
184 behaviour is different from the standard functions when PCRE_PAR‐
185 TIAL_HARD is set. Consider the string "dog" matched against the
186 ungreedy pattern shown above:
187
188 /dog(sbody)??/
189
190 Whereas the standard functions stop as soon as they find the complete
191 match for "dog", the DFA functions also find the partial match for
192 "dogsbody", and so return that when PCRE_PARTIAL_HARD is set.
193
195
196 If a pattern ends with one of sequences \b or \B, which test for word
197 boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-
198 intuitive results. Consider this pattern:
199
200 /\bcat\b/
201
202 This matches "cat", provided there is a word boundary at either end. If
203 the subject string is "the cat", the comparison of the final "t" with a
204 following character cannot take place, so a partial match is found.
205 However, normal matching carries on, and \b matches at the end of the
206 subject when the last character is a letter, so a complete match is
207 found. The result, therefore, is not PCRE_ERROR_PARTIAL. Using
208 PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL, because
209 then the partial match takes precedence.
210
212
213 For releases of PCRE prior to 8.00, because of the way certain internal
214 optimizations were implemented in the pcre_exec() function, the
215 PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be
216 used with all patterns. From release 8.00 onwards, the restrictions no
217 longer apply, and partial matching with can be requested for any pat‐
218 tern.
219
220 Items that were formerly restricted were repeated single characters and
221 repeated metasequences. If PCRE_PARTIAL was set for a pattern that did
222 not conform to the restrictions, pcre_exec() returned the error code
223 PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
224 PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to find out if a compiled
225 pattern can be used for partial matching now always returns 1.
226
228
229 If the escape sequence \P is present in a pcretest data line, the
230 PCRE_PARTIAL_SOFT option is used for the match. Here is a run of
231 pcretest that uses the date example quoted above:
232
233 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
234 data> 25jun04\P
235 0: 25jun04
236 1: jun
237 data> 25dec3\P
238 Partial match: 23dec3
239 data> 3ju\P
240 Partial match: 3ju
241 data> 3juj\P
242 No match
243 data> j\P
244 No match
245
246 The first data string is matched completely, so pcretest shows the
247 matched substrings. The remaining four strings do not match the com‐
248 plete pattern, but the first two are partial matches. Similar output is
249 obtained if DFA matching is used.
250
251 If the escape sequence \P is present more than once in a pcretest data
252 line, the PCRE_PARTIAL_HARD option is set for the match.
253
255
256 When a partial match has been found using a DFA matching function, it
257 is possible to continue the match by providing additional subject data
258 and calling the function again with the same compiled regular expres‐
259 sion, this time setting the PCRE_DFA_RESTART option. You must pass the
260 same working space as before, because this is where details of the pre‐
261 vious partial match are stored. Here is an example using pcretest,
262 using the \R escape sequence to set the PCRE_DFA_RESTART option (\D
263 specifies the use of the DFA matching function):
264
265 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
266 data> 23ja\P\D
267 Partial match: 23ja
268 data> n05\R\D
269 0: n05
270
271 The first call has "23ja" as the subject, and requests partial match‐
272 ing; the second call has "n05" as the subject for the continued
273 (restarted) match. Notice that when the match is complete, only the
274 last part is shown; PCRE does not retain the previously partially-
275 matched string. It is up to the calling program to do that if it needs
276 to.
277
278 You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
279 PCRE_DFA_RESTART to continue partial matching over multiple segments.
280 This facility can be used to pass very long subject strings to the DFA
281 matching functions.
282
284
285 From release 8.00, the standard matching functions can also be used to
286 do multi-segment matching. Unlike the DFA functions, it is not possible
287 to restart the previous match with a new segment of data. Instead, new
288 data must be added to the previous subject string, and the entire match
289 re-run, starting from the point where the partial match occurred. Ear‐
290 lier data can be discarded.
291
292 It is best to use PCRE_PARTIAL_HARD in this situation, because it does
293 not treat the end of a segment as the end of the subject when matching
294 \z, \Z, \b, \B, and $. Consider an unanchored pattern that matches
295 dates:
296
297 re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
298 data> The date is 23ja\P\P
299 Partial match: 23ja
300
301 At this stage, an application could discard the text preceding "23ja",
302 add on text from the next segment, and call the matching function
303 again. Unlike the DFA matching functions, the entire matching string
304 must always be available, and the complete matching process occurs for
305 each call, so more memory and more processing time is needed.
306
307 Note: If the pattern contains lookbehind assertions, or \K, or starts
308 with \b or \B, the string that is returned for a partial match includes
309 characters that precede the partially matched string itself, because
310 these must be retained when adding on more characters for a subsequent
311 matching attempt. However, in some cases you may need to retain even
312 earlier characters, as discussed in the next section.
313
315
316 Certain types of pattern may give problems with multi-segment matching,
317 whichever matching function is used.
318
319 1. If the pattern contains a test for the beginning of a line, you need
320 to pass the PCRE_NOTBOL option when the subject string for any call
321 does start at the beginning of a line. There is also a PCRE_NOTEOL
322 option, but in practice when doing multi-segment matching you should be
323 using PCRE_PARTIAL_HARD, which includes the effect of PCRE_NOTEOL.
324
325 2. Lookbehind assertions that have already been obeyed are catered for
326 in the offsets that are returned for a partial match. However a lookbe‐
327 hind assertion later in the pattern could require even earlier charac‐
328 ters to be inspected. You can handle this case by using the
329 PCRE_INFO_MAXLOOKBEHIND option of the pcre_fullinfo() or
330 pcre[16|32]_fullinfo() functions to obtain the length of the largest
331 lookbehind in the pattern. This length is given in characters, not
332 bytes. If you always retain at least that many characters before the
333 partially matched string, all should be well. (Of course, near the
334 start of the subject, fewer characters may be present; in that case all
335 characters should be retained.)
336
337 3. Because a partial match must always contain at least one character,
338 what might be considered a partial match of an empty string actually
339 gives a "no match" result. For example:
340
341 re> /c(?<=abc)x/
342 data> ab\P
343 No match
344
345 If the next segment begins "cx", a match should be found, but this will
346 only happen if characters from the previous segment are retained. For
347 this reason, a "no match" result should be interpreted as "partial
348 match of an empty string" when the pattern contains lookbehinds.
349
350 4. Matching a subject string that is split into multiple segments may
351 not always produce exactly the same result as matching over one single
352 long string, especially when PCRE_PARTIAL_SOFT is used. The section
353 "Partial Matching and Word Boundaries" above describes an issue that
354 arises if the pattern ends with \b or \B. Another kind of difference
355 may occur when there are multiple matching possibilities, because (for
356 PCRE_PARTIAL_SOFT) a partial match result is given only when there are
357 no completed matches. This means that as soon as the shortest match has
358 been found, continuation to a new subject segment is no longer possi‐
359 ble. Consider again this pcretest example:
360
361 re> /dog(sbody)?/
362 data> dogsb\P
363 0: dog
364 data> do\P\D
365 Partial match: do
366 data> gsb\R\P\D
367 0: g
368 data> dogsbody\D
369 0: dogsbody
370 1: dog
371
372 The first data line passes the string "dogsb" to a standard matching
373 function, setting the PCRE_PARTIAL_SOFT option. Although the string is
374 a partial match for "dogsbody", the result is not PCRE_ERROR_PARTIAL,
375 because the shorter string "dog" is a complete match. Similarly, when
376 the subject is presented to a DFA matching function in several parts
377 ("do" and "gsb" being the first two) the match stops when "dog" has
378 been found, and it is not possible to continue. On the other hand, if
379 "dogsbody" is presented as a single string, a DFA matching function
380 finds both matches.
381
382 Because of these problems, it is best to use PCRE_PARTIAL_HARD when
383 matching multi-segment data. The example above then behaves differ‐
384 ently:
385
386 re> /dog(sbody)?/
387 data> dogsb\P\P
388 Partial match: dogsb
389 data> do\P\D
390 Partial match: do
391 data> gsb\R\P\P\D
392 Partial match: gsb
393
394 5. Patterns that contain alternatives at the top level which do not all
395 start with the same pattern item may not work as expected when
396 PCRE_DFA_RESTART is used. For example, consider this pattern:
397
398 1234|3789
399
400 If the first part of the subject is "ABC123", a partial match of the
401 first alternative is found at offset 3. There is no partial match for
402 the second alternative, because such a match does not start at the same
403 point in the subject string. Attempting to continue with the string
404 "7890" does not yield a match because only those alternatives that
405 match at one point in the subject are remembered. The problem arises
406 because the start of the second alternative matches within the first
407 alternative. There is no problem with anchored patterns or patterns
408 such as:
409
410 1234|ABCD
411
412 where no string can be a partial match for both alternatives. This is
413 not a problem if a standard matching function is used, because the
414 entire match has to be rerun each time:
415
416 re> /1234|3789/
417 data> ABC123\P\P
418 Partial match: 123
419 data> 1237890
420 0: 3789
421
422 Of course, instead of using PCRE_DFA_RESTART, the same technique of re-
423 running the entire match can also be used with the DFA matching func‐
424 tions. Another possibility is to work with two buffers. If a partial
425 match at offset n in the first buffer is followed by "no match" when
426 PCRE_DFA_RESTART is used on the second buffer, you can then try a new
427 match starting at offset n+1 in the first buffer.
428
430
431 Philip Hazel
432 University Computing Service
433 Cambridge CB2 3QH, England.
434
436
437 Last updated: 24 June 2012
438 Copyright (c) 1997-2012 University of Cambridge.
439
440
441
442PCRE 8.31 24 June 2012 PCREPARTIAL(3)