PostgreSQL Source Code  git master
citext.c
Go to the documentation of this file.
1 /*
2  * contrib/citext/citext.c
3  */
4 #include "postgres.h"
5 
6 #include "catalog/pg_collation.h"
7 #include "utils/builtins.h"
8 #include "utils/formatting.h"
9 #include "utils/hashutils.h"
10 #include "utils/varlena.h"
11 
13 
14 /*
15  * ====================
16  * FORWARD DECLARATIONS
17  * ====================
18  */
19 
20 static int32 citextcmp(text *left, text *right, Oid collid);
21 static int32 internal_citext_pattern_cmp(text *left, text *right, Oid collid);
22 
23 /*
24  * =================
25  * UTILITY FUNCTIONS
26  * =================
27  */
28 
29 /*
30  * citextcmp()
31  * Internal comparison function for citext strings.
32  * Returns int32 negative, zero, or positive.
33  */
34 static int32
35 citextcmp(text *left, text *right, Oid collid)
36 {
37  char *lcstr,
38  *rcstr;
39  int32 result;
40 
41  /*
42  * We must do our str_tolower calls with DEFAULT_COLLATION_OID, not the
43  * input collation as you might expect. This is so that the behavior of
44  * citext's equality and hashing functions is not collation-dependent. We
45  * should change this once the core infrastructure is able to cope with
46  * collation-dependent equality and hashing functions.
47  */
48 
49  lcstr = str_tolower(VARDATA_ANY(left), VARSIZE_ANY_EXHDR(left), DEFAULT_COLLATION_OID);
50  rcstr = str_tolower(VARDATA_ANY(right), VARSIZE_ANY_EXHDR(right), DEFAULT_COLLATION_OID);
51 
52  result = varstr_cmp(lcstr, strlen(lcstr),
53  rcstr, strlen(rcstr),
54  collid);
55 
56  pfree(lcstr);
57  pfree(rcstr);
58 
59  return result;
60 }
61 
62 /*
63  * citext_pattern_cmp()
64  * Internal character-by-character comparison function for citext strings.
65  * Returns int32 negative, zero, or positive.
66  */
67 static int32
68 internal_citext_pattern_cmp(text *left, text *right, Oid collid)
69 {
70  char *lcstr,
71  *rcstr;
72  int llen,
73  rlen;
74  int32 result;
75 
76  lcstr = str_tolower(VARDATA_ANY(left), VARSIZE_ANY_EXHDR(left), DEFAULT_COLLATION_OID);
77  rcstr = str_tolower(VARDATA_ANY(right), VARSIZE_ANY_EXHDR(right), DEFAULT_COLLATION_OID);
78 
79  llen = strlen(lcstr);
80  rlen = strlen(rcstr);
81 
82  result = memcmp((void *) lcstr, (void *) rcstr, Min(llen, rlen));
83  if (result == 0)
84  {
85  if (llen < rlen)
86  result = -1;
87  else if (llen > rlen)
88  result = 1;
89  }
90 
91  pfree(lcstr);
92  pfree(rcstr);
93 
94  return result;
95 }
96 
97 /*
98  * ==================
99  * INDEXING FUNCTIONS
100  * ==================
101  */
102 
104 
105 Datum
107 {
108  text *left = PG_GETARG_TEXT_PP(0);
109  text *right = PG_GETARG_TEXT_PP(1);
110  int32 result;
111 
112  result = citextcmp(left, right, PG_GET_COLLATION());
113 
114  PG_FREE_IF_COPY(left, 0);
115  PG_FREE_IF_COPY(right, 1);
116 
117  PG_RETURN_INT32(result);
118 }
119 
121 
122 Datum
124 {
125  text *left = PG_GETARG_TEXT_PP(0);
126  text *right = PG_GETARG_TEXT_PP(1);
127  int32 result;
128 
129  result = internal_citext_pattern_cmp(left, right, PG_GET_COLLATION());
130 
131  PG_FREE_IF_COPY(left, 0);
132  PG_FREE_IF_COPY(right, 1);
133 
134  PG_RETURN_INT32(result);
135 }
136 
138 
139 Datum
141 {
142  text *txt = PG_GETARG_TEXT_PP(0);
143  char *str;
144  Datum result;
145 
146  str = str_tolower(VARDATA_ANY(txt), VARSIZE_ANY_EXHDR(txt), DEFAULT_COLLATION_OID);
147  result = hash_any((unsigned char *) str, strlen(str));
148  pfree(str);
149 
150  /* Avoid leaking memory for toasted inputs */
151  PG_FREE_IF_COPY(txt, 0);
152 
153  PG_RETURN_DATUM(result);
154 }
155 
157 
158 Datum
160 {
161  text *txt = PG_GETARG_TEXT_PP(0);
162  uint64 seed = PG_GETARG_INT64(1);
163  char *str;
164  Datum result;
165 
166  str = str_tolower(VARDATA_ANY(txt), VARSIZE_ANY_EXHDR(txt), DEFAULT_COLLATION_OID);
167  result = hash_any_extended((unsigned char *) str, strlen(str), seed);
168  pfree(str);
169 
170  /* Avoid leaking memory for toasted inputs */
171  PG_FREE_IF_COPY(txt, 0);
172 
173  PG_RETURN_DATUM(result);
174 }
175 
176 /*
177  * ==================
178  * OPERATOR FUNCTIONS
179  * ==================
180  */
181 
183 
184 Datum
186 {
187  text *left = PG_GETARG_TEXT_PP(0);
188  text *right = PG_GETARG_TEXT_PP(1);
189  char *lcstr,
190  *rcstr;
191  bool result;
192 
193  /* We can't compare lengths in advance of downcasing ... */
194 
195  lcstr = str_tolower(VARDATA_ANY(left), VARSIZE_ANY_EXHDR(left), DEFAULT_COLLATION_OID);
196  rcstr = str_tolower(VARDATA_ANY(right), VARSIZE_ANY_EXHDR(right), DEFAULT_COLLATION_OID);
197 
198  /*
199  * Since we only care about equality or not-equality, we can avoid all the
200  * expense of strcoll() here, and just do bitwise comparison.
201  */
202  result = (strcmp(lcstr, rcstr) == 0);
203 
204  pfree(lcstr);
205  pfree(rcstr);
206  PG_FREE_IF_COPY(left, 0);
207  PG_FREE_IF_COPY(right, 1);
208 
209  PG_RETURN_BOOL(result);
210 }
211 
213 
214 Datum
216 {
217  text *left = PG_GETARG_TEXT_PP(0);
218  text *right = PG_GETARG_TEXT_PP(1);
219  char *lcstr,
220  *rcstr;
221  bool result;
222 
223  /* We can't compare lengths in advance of downcasing ... */
224 
225  lcstr = str_tolower(VARDATA_ANY(left), VARSIZE_ANY_EXHDR(left), DEFAULT_COLLATION_OID);
226  rcstr = str_tolower(VARDATA_ANY(right), VARSIZE_ANY_EXHDR(right), DEFAULT_COLLATION_OID);
227 
228  /*
229  * Since we only care about equality or not-equality, we can avoid all the
230  * expense of strcoll() here, and just do bitwise comparison.
231  */
232  result = (strcmp(lcstr, rcstr) != 0);
233 
234  pfree(lcstr);
235  pfree(rcstr);
236  PG_FREE_IF_COPY(left, 0);
237  PG_FREE_IF_COPY(right, 1);
238 
239  PG_RETURN_BOOL(result);
240 }
241 
243 
244 Datum
246 {
247  text *left = PG_GETARG_TEXT_PP(0);
248  text *right = PG_GETARG_TEXT_PP(1);
249  bool result;
250 
251  result = citextcmp(left, right, PG_GET_COLLATION()) < 0;
252 
253  PG_FREE_IF_COPY(left, 0);
254  PG_FREE_IF_COPY(right, 1);
255 
256  PG_RETURN_BOOL(result);
257 }
258 
260 
261 Datum
263 {
264  text *left = PG_GETARG_TEXT_PP(0);
265  text *right = PG_GETARG_TEXT_PP(1);
266  bool result;
267 
268  result = citextcmp(left, right, PG_GET_COLLATION()) <= 0;
269 
270  PG_FREE_IF_COPY(left, 0);
271  PG_FREE_IF_COPY(right, 1);
272 
273  PG_RETURN_BOOL(result);
274 }
275 
277 
278 Datum
280 {
281  text *left = PG_GETARG_TEXT_PP(0);
282  text *right = PG_GETARG_TEXT_PP(1);
283  bool result;
284 
285  result = citextcmp(left, right, PG_GET_COLLATION()) > 0;
286 
287  PG_FREE_IF_COPY(left, 0);
288  PG_FREE_IF_COPY(right, 1);
289 
290  PG_RETURN_BOOL(result);
291 }
292 
294 
295 Datum
297 {
298  text *left = PG_GETARG_TEXT_PP(0);
299  text *right = PG_GETARG_TEXT_PP(1);
300  bool result;
301 
302  result = citextcmp(left, right, PG_GET_COLLATION()) >= 0;
303 
304  PG_FREE_IF_COPY(left, 0);
305  PG_FREE_IF_COPY(right, 1);
306 
307  PG_RETURN_BOOL(result);
308 }
309 
311 
312 Datum
314 {
315  text *left = PG_GETARG_TEXT_PP(0);
316  text *right = PG_GETARG_TEXT_PP(1);
317  bool result;
318 
319  result = internal_citext_pattern_cmp(left, right, PG_GET_COLLATION()) < 0;
320 
321  PG_FREE_IF_COPY(left, 0);
322  PG_FREE_IF_COPY(right, 1);
323 
324  PG_RETURN_BOOL(result);
325 }
326 
328 
329 Datum
331 {
332  text *left = PG_GETARG_TEXT_PP(0);
333  text *right = PG_GETARG_TEXT_PP(1);
334  bool result;
335 
336  result = internal_citext_pattern_cmp(left, right, PG_GET_COLLATION()) <= 0;
337 
338  PG_FREE_IF_COPY(left, 0);
339  PG_FREE_IF_COPY(right, 1);
340 
341  PG_RETURN_BOOL(result);
342 }
343 
345 
346 Datum
348 {
349  text *left = PG_GETARG_TEXT_PP(0);
350  text *right = PG_GETARG_TEXT_PP(1);
351  bool result;
352 
353  result = internal_citext_pattern_cmp(left, right, PG_GET_COLLATION()) > 0;
354 
355  PG_FREE_IF_COPY(left, 0);
356  PG_FREE_IF_COPY(right, 1);
357 
358  PG_RETURN_BOOL(result);
359 }
360 
362 
363 Datum
365 {
366  text *left = PG_GETARG_TEXT_PP(0);
367  text *right = PG_GETARG_TEXT_PP(1);
368  bool result;
369 
370  result = internal_citext_pattern_cmp(left, right, PG_GET_COLLATION()) >= 0;
371 
372  PG_FREE_IF_COPY(left, 0);
373  PG_FREE_IF_COPY(right, 1);
374 
375  PG_RETURN_BOOL(result);
376 }
377 
378 /*
379  * ===================
380  * AGGREGATE FUNCTIONS
381  * ===================
382  */
383 
385 
386 Datum
388 {
389  text *left = PG_GETARG_TEXT_PP(0);
390  text *right = PG_GETARG_TEXT_PP(1);
391  text *result;
392 
393  result = citextcmp(left, right, PG_GET_COLLATION()) < 0 ? left : right;
394  PG_RETURN_TEXT_P(result);
395 }
396 
398 
399 Datum
401 {
402  text *left = PG_GETARG_TEXT_PP(0);
403  text *right = PG_GETARG_TEXT_PP(1);
404  text *result;
405 
406  result = citextcmp(left, right, PG_GET_COLLATION()) > 0 ? left : right;
407  PG_RETURN_TEXT_P(result);
408 }
#define VARDATA_ANY(PTR)
Definition: postgres.h:348
static int32 internal_citext_pattern_cmp(text *left, text *right, Oid collid)
Definition: citext.c:68
Datum citext_smaller(PG_FUNCTION_ARGS)
Definition: citext.c:387
Datum hash_any(const unsigned char *k, int keylen)
Definition: hashfn.c:148
char * str_tolower(const char *buff, size_t nbytes, Oid collid)
Definition: formatting.c:1645
PG_FUNCTION_INFO_V1(citext_cmp)
Datum citext_pattern_lt(PG_FUNCTION_ARGS)
Definition: citext.c:313
#define Min(x, y)
Definition: c.h:904
#define PG_RETURN_INT32(x)
Definition: fmgr.h:344
static int32 citextcmp(text *left, text *right, Oid collid)
Definition: citext.c:35
Datum citext_hash(PG_FUNCTION_ARGS)
Definition: citext.c:140
Datum hash_any_extended(const unsigned char *k, int keylen, uint64 seed)
Definition: hashfn.c:374
unsigned int Oid
Definition: postgres_ext.h:31
#define PG_GET_COLLATION()
Definition: fmgr.h:193
Datum citext_ge(PG_FUNCTION_ARGS)
Definition: citext.c:296
signed int int32
Definition: c.h:346
#define PG_GETARG_TEXT_PP(n)
Definition: fmgr.h:303
void pfree(void *pointer)
Definition: mcxt.c:1056
int varstr_cmp(const char *arg1, int len1, const char *arg2, int len2, Oid collid)
Definition: varlena.c:1474
Datum citext_larger(PG_FUNCTION_ARGS)
Definition: citext.c:400
Datum citext_pattern_ge(PG_FUNCTION_ARGS)
Definition: citext.c:364
Datum citext_cmp(PG_FUNCTION_ARGS)
Definition: citext.c:106
PG_MODULE_MAGIC
Definition: citext.c:12
Datum citext_hash_extended(PG_FUNCTION_ARGS)
Definition: citext.c:159
Datum citext_pattern_cmp(PG_FUNCTION_ARGS)
Definition: citext.c:123
Datum citext_pattern_gt(PG_FUNCTION_ARGS)
Definition: citext.c:347
#define PG_RETURN_BOOL(x)
Definition: fmgr.h:349
uintptr_t Datum
Definition: postgres.h:367
#define PG_RETURN_DATUM(x)
Definition: fmgr.h:343
Datum citext_gt(PG_FUNCTION_ARGS)
Definition: citext.c:279
#define PG_RETURN_TEXT_P(x)
Definition: fmgr.h:361
Datum citext_ne(PG_FUNCTION_ARGS)
Definition: citext.c:215
Datum citext_le(PG_FUNCTION_ARGS)
Definition: citext.c:262
#define PG_FREE_IF_COPY(ptr, n)
Definition: fmgr.h:255
Datum citext_eq(PG_FUNCTION_ARGS)
Definition: citext.c:185
Datum citext_lt(PG_FUNCTION_ARGS)
Definition: citext.c:245
#define VARSIZE_ANY_EXHDR(PTR)
Definition: postgres.h:341
Definition: c.h:549
#define PG_FUNCTION_ARGS
Definition: fmgr.h:188
#define PG_GETARG_INT64(n)
Definition: fmgr.h:277
Datum citext_pattern_le(PG_FUNCTION_ARGS)
Definition: citext.c:330