1PCRE2PARTIAL(3) Library Functions Manual PCRE2PARTIAL(3)
2
6 PCRE2 - Perl-compatible regular expressions
7
9 In normal use of PCRE2, if there is a match up to the end of a subject
11 string, but more characters are needed to match the entire pattern,
12 PCRE2_ERROR_NOMATCH is returned, just like any other failing match.
13 There are circumstances where it might be helpful to distinguish this
14 "partial match" case.
15 One example is an application where the subject string is very long,
17 and not all available at once. The requirement here is to be able to do
18 the matching segment by segment, but special action is needed when a
19 matched substring spans the boundary between two segments.
20 Another example is checking a user input string as it is typed, to
22 ensure that it conforms to a required format. Invalid characters can be
23 immediately diagnosed and rejected, giving instant feedback.
24 Partial matching is a PCRE2-specific feature; it is not Perl-compati‐
26 ble. It is requested by setting one of the PCRE2_PARTIAL_HARD or
27 PCRE2_PARTIAL_SOFT options when calling a matching function. The dif‐
28 ference between the two options is whether or not a partial match is
29 preferred to an alternative complete match, though the details differ
30 between the two types of matching function. If both options are set,
31 PCRE2_PARTIAL_HARD takes precedence.
32 If you want to use partial matching with just-in-time optimized code,
34 as well as setting a partial match option for the matching function,
35 you must also call pcre2_jit_compile() with one or both of these
36 options:
37 PCRE2_JIT_PARTIAL_HARD
39 PCRE2_JIT_PARTIAL_SOFT
40 PCRE2_JIT_COMPLETE should also be set if you are going to run non-par‐
42 tial matches on the same pattern. Separate code is compiled for each
43 mode. If the appropriate JIT mode has not been compiled, interpretive
44 matching code is used.
45 Setting a partial matching option disables two of PCRE2's standard
47 optimization hints. PCRE2 remembers the last literal code unit in a
48 pattern, and abandons matching immediately if it is not present in the
49 subject string. This optimization cannot be used for a subject string
50 that might match only partially. PCRE2 also remembers a minimum length
51 of a matching string, and does not bother to run the matching function
52 on shorter strings. This optimization is also disabled for partial
53 matching.
54
56 A possible partial match occurs during matching when the end of the
58 subject string is reached successfully, but either more characters are
59 needed to complete the match, or the addition of more characters might
60 change what is matched.
61 Example 1: if the pattern is /abc/ and the subject is "ab", more char‐
63 acters are definitely needed to complete a match. In this case both
64 hard and soft matching options yield a partial match.
65 Example 2: if the pattern is /ab+/ and the subject is "ab", a complete
67 match can be found, but the addition of more characters might change
68 what is matched. In this case, only PCRE2_PARTIAL_HARD returns a par‐
69 tial match; PCRE2_PARTIAL_SOFT returns the complete match.
70 On reaching the end of the subject, when PCRE2_PARTIAL_HARD is set, if
72 the next pattern item is \z, \Z, \b, \B, or $ there is always a partial
73 match. Otherwise, for both options, the next pattern item must be one
74 that inspects a character, and at least one of the following must be
75 true:
76 (1) At least one character has already been inspected. An inspected
78 character need not form part of the final matched string; lookbehind
79 assertions and the \K escape sequence provide ways of inspecting char‐
80 acters before the start of a matched string.
81 (2) The pattern contains one or more lookbehind assertions. This condi‐
83 tion exists in case there is a lookbehind that inspects characters
84 before the start of the match.
85 (3) There is a special case when the whole pattern can match an empty
87 string. When the starting point is at the end of the subject, the
88 empty string match is a possibility, and if PCRE2_PARTIAL_SOFT is set
89 and neither of the above conditions is true, it is returned. However,
90 because adding more characters might result in a non-empty match,
91 PCRE2_PARTIAL_HARD returns a partial match, which in this case means
92 "there is going to be a match at this point, but until some more char‐
93 acters are added, we do not know if it will be an empty string or some‐
94 thing longer".
95
97 When a partial matching option is set, the result of calling
99 pcre2_match() can be one of the following:
100 A successful match
102 A complete match has been found, starting and ending within this sub‐
103 ject.
104 PCRE2_ERROR_NOMATCH
106 No match can start anywhere in this subject.
107 PCRE2_ERROR_PARTIAL
109 Adding more characters may result in a complete match that uses one
110 or more characters from the end of this subject.
111 When a partial match is returned, the first two elements in the ovector
113 point to the portion of the subject that was matched, but the values in
114 the rest of the ovector are undefined. The appearance of \K in the pat‐
115 tern has no effect for a partial match. Consider this pattern:
116 /abc\K123/
118 If it is matched against "456abc123xyz" the result is a complete match,
120 and the ovector defines the matched string as "123", because \K resets
121 the "start of match" point. However, if a partial match is requested
122 and the subject string is "456abc12", a partial match is found for the
123 string "abc12", because all these characters are needed for a subse‐
124 quent re-match with additional characters.
125 If there is more than one partial match, the first one that was found
127 provides the data that is returned. Consider this pattern:
128 /123\w+X|dogY/
130 If this is matched against the subject string "abc123dog", both alter‐
132 natives fail to match, but the end of the subject is reached during
133 matching, so PCRE2_ERROR_PARTIAL is returned. The offsets are set to 3
134 and 9, identifying "123dog" as the first partial match. (In this exam‐
135 ple, there are two partial matches, because "dog" on its own partially
136 matches the second alternative.)
137 How a partial match is processed by pcre2_match()
139 What happens when a partial match is identified depends on which of the
141 two partial matching options is set.
142 If PCRE2_PARTIAL_HARD is set, PCRE2_ERROR_PARTIAL is returned as soon
144 as a partial match is found, without continuing to search for possible
145 complete matches. This option is "hard" because it prefers an earlier
146 partial match over a later complete match. For this reason, the assump‐
147 tion is made that the end of the supplied subject string is not the
148 true end of the available data, which is why \z, \Z, \b, \B, and $
149 always give a partial match.
150 If PCRE2_PARTIAL_SOFT is set, the partial match is remembered, but
152 matching continues as normal, and other alternatives in the pattern are
153 tried. If no complete match can be found, PCRE2_ERROR_PARTIAL is
154 returned instead of PCRE2_ERROR_NOMATCH. This option is "soft" because
155 it prefers a complete match over a partial match. All the various
156 matching items in a pattern behave as if the subject string is poten‐
157 tially complete; \z, \Z, and $ match at the end of the subject, as nor‐
158 mal, and for \b and \B the end of the subject is treated as a non-
159 alphanumeric.
160 The difference between the two partial matching options can be illus‐
162 trated by a pattern such as:
163 /dog(sbody)?/
165 This matches either "dog" or "dogsbody", greedily (that is, it prefers
167 the longer string if possible). If it is matched against the string
168 "dog" with PCRE2_PARTIAL_SOFT, it yields a complete match for "dog".
169 However, if PCRE2_PARTIAL_HARD is set, the result is PCRE2_ERROR_PAR‐
170 TIAL. On the other hand, if the pattern is made ungreedy the result is
171 different:
172 /dog(sbody)??/
174 In this case the result is always a complete match because that is
176 found first, and matching never continues after finding a complete
177 match. It might be easier to follow this explanation by thinking of the
178 two patterns like this:
179 /dog(sbody)?/ is the same as /dogsbody|dog/
181 /dog(sbody)??/ is the same as /dog|dogsbody/
182 The second pattern will never match "dogsbody", because it will always
184 find the shorter match first.
185 Example of partial matching using pcre2test
187 The pcre2test data modifiers partial_hard (or ph) and partial_soft (or
189 ps) set PCRE2_PARTIAL_HARD and PCRE2_PARTIAL_SOFT, respectively, when
190 calling pcre2_match(). Here is a run of pcre2test using a pattern that
191 matches the whole subject in the form of a date:
192 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
194 data> 25dec3\=ph
195 Partial match: 23dec3
196 data> 3ju\=ph
197 Partial match: 3ju
198 data> 3juj\=ph
199 No match
200 This example gives the same results for both hard and soft partial
202 matching options. Here is an example where there is a difference:
203 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
205 data> 25jun04\=ps
206 0: 25jun04
207 1: jun
208 data> 25jun04\=ph
209 Partial match: 25jun04
210 With PCRE2_PARTIAL_SOFT, the subject is matched completely. For
212 PCRE2_PARTIAL_HARD, however, the subject is assumed not to be complete,
213 so there is only a partial match.
214
216 PCRE was not originally designed with multi-segment matching in mind.
218 However, over time, features (including partial matching) that make
219 multi-segment matching possible have been added. A very long string can
220 be searched segment by segment by calling pcre2_match() repeatedly,
221 with the aim of achieving the same results that would happen if the
222 entire string was available for searching all the time. Normally, the
223 strings that are being sought are much shorter than each individual
224 segment, and are in the middle of very long strings, so the pattern is
225 normally not anchored.
226 Special logic must be implemented to handle a matched substring that
228 spans a segment boundary. PCRE2_PARTIAL_HARD should be used, because it
229 returns a partial match at the end of a segment whenever there is the
230 possibility of changing the match by adding more characters. The
231 PCRE2_NOTBOL option should also be set for all but the first segment.
232 When a partial match occurs, the next segment must be added to the cur‐
234 rent subject and the match re-run, using the startoffset argument of
235 pcre2_match() to begin at the point where the partial match started.
236 For example:
237 re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
239 data> ...the date is 23ja\=ph
240 Partial match: 23ja
241 data> ...the date is 23jan19 and on that day...\=offset=15
242 0: 23jan19
243 1: jan
244 Note the use of the offset modifier to start the new match where the
246 partial match was found. In this example, the next segment was added to
247 the one in which the partial match was found. This is the most
248 straightforward approach, typically using a memory buffer that is twice
249 the size of each segment. After a partial match, the first half of the
250 buffer is discarded, the second half is moved to the start of the buf‐
251 fer, and a new segment is added before repeating the match as in the
252 example above. After a no match, the entire buffer can be discarded.
253 If there are memory constraints, you may want to discard text that pre‐
255 cedes a partial match before adding the next segment. Unfortunately,
256 this is not at present straightforward. In cases such as the above,
257 where the pattern does not contain any lookbehinds, it is sufficient to
258 retain only the partially matched substring. However, if the pattern
259 contains a lookbehind assertion, characters that precede the start of
260 the partial match may have been inspected during the matching process.
261 When pcre2test displays a partial match, it indicates these characters
262 with '<' if the allusedtext modifier is set:
263 re> "(?<=123)abc"
265 data> xx123ab\=ph,allusedtext
266 Partial match: 123ab
267 <<<
268 However, the allusedtext modifier is not available for JIT matching,
270 because JIT matching does not record the first (or last) consulted
271 characters. For this reason, this information is not available via the
272 API. It is therefore not possible in general to obtain the exact number
273 of characters that must be retained in order to get the right match
274 result. If you cannot retain the entire segment, you must find some
275 heuristic way of choosing.
276 If you know the approximate length of the matching substrings, you can
278 use that to decide how much text to retain. The only lookbehind infor‐
279 mation that is currently available via the API is the length of the
280 longest individual lookbehind in a pattern, but this can be misleading
281 if there are nested lookbehinds. The value returned by calling
282 pcre2_pattern_info() with the PCRE2_INFO_MAXLOOKBEHIND option is the
283 maximum number of characters (not code units) that any individual look‐
284 behind moves back when it is processed. A pattern such as
285 "(?<=(?<!b)a)" has a maximum lookbehind value of one, but inspects two
286 characters before its starting point.
287 In a non-UTF or a 32-bit case, moving back is just a subtraction, but
289 in UTF-8 or UTF-16 you have to count characters while moving back
290 through the code units.
291
293 The DFA function moves along the subject string character by character,
295 without backtracking, searching for all possible matches simultane‐
296 ously. If the end of the subject is reached before the end of the pat‐
297 tern, there is the possibility of a partial match.
298 When PCRE2_PARTIAL_SOFT is set, PCRE2_ERROR_PARTIAL is returned only if
300 there have been no complete matches. Otherwise, the complete matches
301 are returned. If PCRE2_PARTIAL_HARD is set, a partial match takes
302 precedence over any complete matches. The portion of the string that
303 was matched when the longest partial match was found is set as the
304 first matching string.
305 Because the DFA function always searches for all possible matches, and
307 there is no difference between greedy and ungreedy repetition, its be‐
308 haviour is different from the pcre2_match(). Consider the string "dog"
309 matched against this ungreedy pattern:
310 /dog(sbody)??/
312 Whereas the standard function stops as soon as it finds the complete
314 match for "dog", the DFA function also finds the partial match for
315 "dogsbody", and so returns that when PCRE2_PARTIAL_HARD is set.
316
318 When a partial match has been found using the DFA matching function, it
320 is possible to continue the match by providing additional subject data
321 and calling the function again with the same compiled regular expres‐
322 sion, this time setting the PCRE2_DFA_RESTART option. You must pass the
323 same working space as before, because this is where details of the pre‐
324 vious partial match are stored. You can set the PCRE2_PARTIAL_SOFT or
325 PCRE2_PARTIAL_HARD options with PCRE2_DFA_RESTART to continue partial
326 matching over multiple segments. Here is an example using pcre2test:
327 re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
329 data> 23ja\=dfa,ps
330 Partial match: 23ja
331 data> n05\=dfa,dfa_restart
332 0: n05
333 The first call has "23ja" as the subject, and requests partial match‐
335 ing; the second call has "n05" as the subject for the continued
336 (restarted) match. Notice that when the match is complete, only the
337 last part is shown; PCRE2 does not retain the previously partially-
338 matched string. It is up to the calling program to do that if it needs
339 to. This means that, for an unanchored pattern, if a continued match
340 fails, it is not possible to try again at a new starting point. All
341 this facility is capable of doing is continuing with the previous match
342 attempt. For example, consider this pattern:
343 1234|3789
345 If the first part of the subject is "ABC123", a partial match of the
347 first alternative is found at offset 3. There is no partial match for
348 the second alternative, because such a match does not start at the same
349 point in the subject string. Attempting to continue with the string
350 "7890" does not yield a match because only those alternatives that
351 match at one point in the subject are remembered. Depending on the
352 application, this may or may not be what you want.
353 If you do want to allow for starting again at the next character, one
355 way of doing it is to retain some or all of the segment and try a new
356 complete match, as described for pcre2_match() above. Another possibil‐
357 ity is to work with two buffers. If a partial match at offset n in the
358 first buffer is followed by "no match" when PCRE2_DFA_RESTART is used
359 on the second buffer, you can then try a new match starting at offset
360 n+1 in the first buffer.
361
363 Philip Hazel
365 University Computing Service
366 Cambridge, England.
367
369 Last updated: 04 September 2019
371 Copyright (c) 1997-2019 University of Cambridge.
372PCRE2 10.34 04 September 2019 PCRE2PARTIAL(3)