1PCRE2PARTIAL(3) Library Functions Manual PCRE2PARTIAL(3)
2
6 PCRE2 - Perl-compatible regular expressions
7
9 In normal use of PCRE2, if the subject string that is passed to a
11 matching function matches as far as it goes, but is too short to match
12 the entire pattern, PCRE2_ERROR_NOMATCH is returned. There are circum‐
13 stances where it might be helpful to distinguish this case from other
14 cases in which there is no match.
15 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 ^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
21 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 PCRE2 supports partial matching by means of the PCRE2_PARTIAL_SOFT and
32 PCRE2_PARTIAL_HARD options, which can be set when calling a matching
33 function. The difference between the two options is whether or not a
34 partial match is preferred to an alternative complete match, though the
35 details differ between the two types of matching function. If both
36 options are set, PCRE2_PARTIAL_HARD takes precedence.
37 If you want to use partial matching with just-in-time optimized code,
39 you must call pcre2_jit_compile() with one or both of these options:
40 PCRE2_JIT_PARTIAL_SOFT
42 PCRE2_JIT_PARTIAL_HARD
43 PCRE2_JIT_COMPLETE should also be set if you are going to run non-par‐
45 tial matches on the same pattern. If the appropriate JIT mode has not
46 been compiled, interpretive matching code is used.
47 Setting a partial matching option disables two of PCRE2's standard
49 optimizations. PCRE2 remembers the last literal code unit in a pattern,
50 and abandons matching immediately if it is not present in the subject
51 string. This optimization cannot be used for a subject string that
52 might match only partially. PCRE2 also knows the minimum length of a
53 matching string, and does not bother to run the matching function on
54 shorter strings. This optimization is also disabled for partial match‐
55 ing.
56
58 A partial match occurs during a call to pcre2_match() when the end of
60 the subject string is reached successfully, but matching cannot con‐
61 tinue because more characters are needed. However, at least one charac‐
62 ter in the subject must have been inspected. This character need not
63 form part of the final matched string; lookbehind assertions and the \K
64 escape sequence provide ways of inspecting characters before the start
65 of a matched string. The requirement for inspecting at least one char‐
66 acter exists because an empty string can always be matched; without
67 such a restriction there would always be a partial match of an empty
68 string at the end of the subject.
69 When a partial match is returned, the first two elements in the ovector
71 point to the portion of the subject that was matched, but the values in
72 the rest of the ovector are undefined. The appearance of \K in the pat‐
73 tern has no effect for a partial match. Consider this pattern:
74 /abc\K123/
76 If it is matched against "456abc123xyz" the result is a complete match,
78 and the ovector defines the matched string as "123", because \K resets
79 the "start of match" point. However, if a partial match is requested
80 and the subject string is "456abc12", a partial match is found for the
81 string "abc12", because all these characters are needed for a subse‐
82 quent re-match with additional characters.
83 What happens when a partial match is identified depends on which of the
85 two partial matching options are set.
86 PCRE2_PARTIAL_SOFT WITH pcre2_match()
88 If PCRE2_PARTIAL_SOFT is set when pcre2_match() identifies a partial
90 match, the partial match is remembered, but matching continues as nor‐
91 mal, and other alternatives in the pattern are tried. If no complete
92 match can be found, PCRE2_ERROR_PARTIAL is returned instead of
93 PCRE2_ERROR_NOMATCH.
94 This option is "soft" because it prefers a complete match over a par‐
96 tial match. All the various matching items in a pattern behave as if
97 the subject string is potentially complete. For example, \z, \Z, and $
98 match at the end of the subject, as normal, and for \b and \B the end
99 of the subject is treated as a non-alphanumeric.
100 If there is more than one partial match, the first one that was found
102 provides the data that is returned. Consider this pattern:
103 /123\w+X|dogY/
105 If this is matched against the subject string "abc123dog", both alter‐
107 natives fail to match, but the end of the subject is reached during
108 matching, so PCRE2_ERROR_PARTIAL is returned. The offsets are set to 3
109 and 9, identifying "123dog" as the first partial match that was found.
110 (In this example, there are two partial matches, because "dog" on its
111 own partially matches the second alternative.)
112 PCRE2_PARTIAL_HARD WITH pcre2_match()
114 If PCRE2_PARTIAL_HARD is set for pcre2_match(), PCRE2_ERROR_PARTIAL is
116 returned as soon as a partial match is found, without continuing to
117 search for possible complete matches. This option is "hard" because it
118 prefers an earlier partial match over a later complete match. For this
119 reason, the assumption is made that the end of the supplied subject
120 string may not be the true end of the available data, and so, if \z,
121 \Z, \b, \B, or $ are encountered at the end of the subject, the result
122 is PCRE2_ERROR_PARTIAL, provided that at least one character in the
123 subject has been inspected.
124 Comparing hard and soft partial matching
126 The difference between the two partial matching options can be illus‐
128 trated by a pattern such as:
129 /dog(sbody)?/
131 This matches either "dog" or "dogsbody", greedily (that is, it prefers
133 the longer string if possible). If it is matched against the string
134 "dog" with PCRE2_PARTIAL_SOFT, it yields a complete match for "dog".
135 However, if PCRE2_PARTIAL_HARD is set, the result is PCRE2_ERROR_PAR‐
136 TIAL. On the other hand, if the pattern is made ungreedy the result is
137 different:
138 /dog(sbody)??/
140 In this case the result is always a complete match because that is
142 found first, and matching never continues after finding a complete
143 match. It might be easier to follow this explanation by thinking of the
144 two patterns like this:
145 /dog(sbody)?/ is the same as /dogsbody|dog/
147 /dog(sbody)??/ is the same as /dog|dogsbody/
148 The second pattern will never match "dogsbody", because it will always
150 find the shorter match first.
151
153 The DFA functions move along the subject string character by character,
155 without backtracking, searching for all possible matches simultane‐
156 ously. If the end of the subject is reached before the end of the pat‐
157 tern, there is the possibility of a partial match, again provided that
158 at least one character has been inspected.
159 When PCRE2_PARTIAL_SOFT is set, PCRE2_ERROR_PARTIAL is returned only if
161 there have been no complete matches. Otherwise, the complete matches
162 are returned. However, if PCRE2_PARTIAL_HARD is set, a partial match
163 takes precedence over any complete matches. The portion of the string
164 that was matched when the longest partial match was found is set as the
165 first matching string.
166 Because the DFA functions always search for all possible matches, and
168 there is no difference between greedy and ungreedy repetition, their
169 behaviour is different from the standard functions when PCRE2_PAR‐
170 TIAL_HARD is set. Consider the string "dog" matched against the
171 ungreedy pattern shown above:
172 /dog(sbody)??/
174 Whereas the standard function stops as soon as it finds the complete
176 match for "dog", the DFA function also finds the partial match for
177 "dogsbody", and so returns that when PCRE2_PARTIAL_HARD is set.
178
180 If a pattern ends with one of sequences \b or \B, which test for word
182 boundaries, partial matching with PCRE2_PARTIAL_SOFT can give counter-
183 intuitive results. Consider this pattern:
184 /\bcat\b/
186 This matches "cat", provided there is a word boundary at either end. If
188 the subject string is "the cat", the comparison of the final "t" with a
189 following character cannot take place, so a partial match is found.
190 However, normal matching carries on, and \b matches at the end of the
191 subject when the last character is a letter, so a complete match is
192 found. The result, therefore, is not PCRE2_ERROR_PARTIAL. Using
193 PCRE2_PARTIAL_HARD in this case does yield PCRE2_ERROR_PARTIAL, because
194 then the partial match takes precedence.
195
197 If the partial_soft (or ps) modifier is present on a pcre2test data
199 line, the PCRE2_PARTIAL_SOFT option is used for the match. Here is a
200 run of pcre2test that uses the date example quoted above:
201 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
203 data> 25jun04\=ps
204 0: 25jun04
205 1: jun
206 data> 25dec3\=ps
207 Partial match: 23dec3
208 data> 3ju\=ps
209 Partial match: 3ju
210 data> 3juj\=ps
211 No match
212 data> j\=ps
213 No match
214 The first data string is matched completely, so pcre2test shows the
216 matched substrings. The remaining four strings do not match the com‐
217 plete pattern, but the first two are partial matches. Similar output is
218 obtained if DFA matching is used.
219 If the partial_hard (or ph) modifier is present on a pcre2test data
221 line, the PCRE2_PARTIAL_HARD option is set for the match.
222
224 When a partial match has been found using a DFA matching function, it
226 is possible to continue the match by providing additional subject data
227 and calling the function again with the same compiled regular expres‐
228 sion, this time setting the PCRE2_DFA_RESTART option. You must pass the
229 same working space as before, because this is where details of the pre‐
230 vious partial match are stored. Here is an example using pcre2test:
231 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
233 data> 23ja\=dfa,ps
234 Partial match: 23ja
235 data> n05\=dfa,dfa_restart
236 0: n05
237 The first call has "23ja" as the subject, and requests partial match‐
239 ing; the second call has "n05" as the subject for the continued
240 (restarted) match. Notice that when the match is complete, only the
241 last part is shown; PCRE2 does not retain the previously partially-
242 matched string. It is up to the calling program to do that if it needs
243 to.
244 That means that, for an unanchored pattern, if a continued match fails,
246 it is not possible to try again at a new starting point. All this
247 facility is capable of doing is continuing with the previous match
248 attempt. In the previous example, if the second set of data is "ug23"
249 the result is no match, even though there would be a match for "aug23"
250 if the entire string were given at once. Depending on the application,
251 this may or may not be what you want. The only way to allow for start‐
252 ing again at the next character is to retain the matched part of the
253 subject and try a new complete match.
254 You can set the PCRE2_PARTIAL_SOFT or PCRE2_PARTIAL_HARD options with
256 PCRE2_DFA_RESTART to continue partial matching over multiple segments.
257 This facility can be used to pass very long subject strings to the DFA
258 matching functions.
259
261 Unlike the DFA function, it is not possible to restart the previous
263 match with a new segment of data when using pcre2_match(). Instead, new
264 data must be added to the previous subject string, and the entire match
265 re-run, starting from the point where the partial match occurred. Ear‐
266 lier data can be discarded.
267 It is best to use PCRE2_PARTIAL_HARD in this situation, because it does
269 not treat the end of a segment as the end of the subject when matching
270 \z, \Z, \b, \B, and $. Consider an unanchored pattern that matches
271 dates:
272 re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
274 data> The date is 23ja\=ph
275 Partial match: 23ja
276 At this stage, an application could discard the text preceding "23ja",
278 add on text from the next segment, and call the matching function
279 again. Unlike the DFA matching function, the entire matching string
280 must always be available, and the complete matching process occurs for
281 each call, so more memory and more processing time is needed.
282
284 Certain types of pattern may give problems with multi-segment matching,
286 whichever matching function is used.
287 1. If the pattern contains a test for the beginning of a line, you need
289 to pass the PCRE2_NOTBOL option when the subject string for any call
290 does start at the beginning of a line. There is also a PCRE2_NOTEOL
291 option, but in practice when doing multi-segment matching you should be
292 using PCRE2_PARTIAL_HARD, which includes the effect of PCRE2_NOTEOL.
293 2. If a pattern contains a lookbehind assertion, characters that pre‐
295 cede the start of the partial match may have been inspected during the
296 matching process. When using pcre2_match(), sufficient characters must
297 be retained for the next match attempt. You can ensure that enough
298 characters are retained by doing the following:
299 Before doing any matching, find the length of the longest lookbehind in
301 the pattern by calling pcre2_pattern_info() with the
302 PCRE2_INFO_MAXLOOKBEHIND option. Note that the resulting count is in
303 characters, not code units. After a partial match, moving back from the
304 ovector[0] offset in the subject by the number of characters given for
305 the maximum lookbehind gets you to the earliest character that must be
306 retained. In a non-UTF or a 32-bit situation, moving back is just a
307 subtraction, but in UTF-8 or UTF-16 you have to count characters while
308 moving back through the code units.
309 Characters before the point you have now reached can be discarded, and
311 after the next segment has been added to what is retained, you should
312 run the next match with the startoffset argument set so that the match
313 begins at the same point as before.
314 For example, if the pattern "(?<=123)abc" is partially matched against
316 the string "xx123ab", the ovector offsets are 5 and 7 ("ab"). The maxi‐
317 mum lookbehind count is 3, so all characters before offset 2 can be
318 discarded. The value of startoffset for the next match should be 3.
319 When pcre2test displays a partial match, it indicates the lookbehind
320 characters with '<' characters:
321 re> "(?<=123)abc"
323 data> xx123ab\=ph
324 Partial match: 123ab
325 <<<
326 3. Because a partial match must always contain at least one character,
328 what might be considered a partial match of an empty string actually
329 gives a "no match" result. For example:
330 re> /c(?<=abc)x/
332 data> ab\=ps
333 No match
334 If the next segment begins "cx", a match should be found, but this will
336 only happen if characters from the previous segment are retained. For
337 this reason, a "no match" result should be interpreted as "partial
338 match of an empty string" when the pattern contains lookbehinds.
339 4. Matching a subject string that is split into multiple segments may
341 not always produce exactly the same result as matching over one single
342 long string, especially when PCRE2_PARTIAL_SOFT is used. The section
343 "Partial Matching and Word Boundaries" above describes an issue that
344 arises if the pattern ends with \b or \B. Another kind of difference
345 may occur when there are multiple matching possibilities, because (for
346 PCRE2_PARTIAL_SOFT) a partial match result is given only when there are
347 no completed matches. This means that as soon as the shortest match has
348 been found, continuation to a new subject segment is no longer possi‐
349 ble. Consider this pcre2test example:
350 re> /dog(sbody)?/
352 data> dogsb\=ps
353 0: dog
354 data> do\=ps,dfa
355 Partial match: do
356 data> gsb\=ps,dfa,dfa_restart
357 0: g
358 data> dogsbody\=dfa
359 0: dogsbody
360 1: dog
361 The first data line passes the string "dogsb" to a standard matching
363 function, setting the PCRE2_PARTIAL_SOFT option. Although the string is
364 a partial match for "dogsbody", the result is not PCRE2_ERROR_PARTIAL,
365 because the shorter string "dog" is a complete match. Similarly, when
366 the subject is presented to a DFA matching function in several parts
367 ("do" and "gsb" being the first two) the match stops when "dog" has
368 been found, and it is not possible to continue. On the other hand, if
369 "dogsbody" is presented as a single string, a DFA matching function
370 finds both matches.
371 Because of these problems, it is best to use PCRE2_PARTIAL_HARD when
373 matching multi-segment data. The example above then behaves differ‐
374 ently:
375 re> /dog(sbody)?/
377 data> dogsb\=ph
378 Partial match: dogsb
379 data> do\=ps,dfa
380 Partial match: do
381 data> gsb\=ph,dfa,dfa_restart
382 Partial match: gsb
383 5. Patterns that contain alternatives at the top level which do not all
385 start with the same pattern item may not work as expected when
386 PCRE2_DFA_RESTART is used. For example, consider this pattern:
387 1234|3789
389 If the first part of the subject is "ABC123", a partial match of the
391 first alternative is found at offset 3. There is no partial match for
392 the second alternative, because such a match does not start at the same
393 point in the subject string. Attempting to continue with the string
394 "7890" does not yield a match because only those alternatives that
395 match at one point in the subject are remembered. The problem arises
396 because the start of the second alternative matches within the first
397 alternative. There is no problem with anchored patterns or patterns
398 such as:
399 1234|ABCD
401 where no string can be a partial match for both alternatives. This is
403 not a problem if a standard matching function is used, because the
404 entire match has to be rerun each time:
405 re> /1234|3789/
407 data> ABC123\=ph
408 Partial match: 123
409 data> 1237890
410 0: 3789
411 Of course, instead of using PCRE2_DFA_RESTART, the same technique of
413 re-running the entire match can also be used with the DFA matching
414 function. Another possibility is to work with two buffers. If a partial
415 match at offset n in the first buffer is followed by "no match" when
416 PCRE2_DFA_RESTART is used on the second buffer, you can then try a new
417 match starting at offset n+1 in the first buffer.
418
420 Philip Hazel
422 University Computing Service
423 Cambridge, England.
424
426 Last updated: 22 December 2014
428 Copyright (c) 1997-2014 University of Cambridge.
429PCRE2 10.00 22 December 2014 PCRE2PARTIAL(3)