1 /*
2 * fuzzystrmatch.c
3 *
4 * Functions for "fuzzy" comparison of strings
5 *
6 * Joe Conway <mail@joeconway.com>
7 *
8 * contrib/fuzzystrmatch/fuzzystrmatch.c
9 * Copyright (c) 2001-2018, PostgreSQL Global Development Group
10 * ALL RIGHTS RESERVED;
11 *
12 * metaphone()
13 * -----------
14 * Modified for PostgreSQL by Joe Conway.
15 * Based on CPAN's "Text-Metaphone-1.96" by Michael G Schwern <schwern@pobox.com>
16 * Code slightly modified for use as PostgreSQL function (palloc, elog, etc).
17 * Metaphone was originally created by Lawrence Philips and presented in article
18 * in "Computer Language" December 1990 issue.
19 *
20 * Permission to use, copy, modify, and distribute this software and its
21 * documentation for any purpose, without fee, and without a written agreement
22 * is hereby granted, provided that the above copyright notice and this
23 * paragraph and the following two paragraphs appear in all copies.
24 *
25 * IN NO EVENT SHALL THE AUTHORS OR DISTRIBUTORS BE LIABLE TO ANY PARTY FOR
26 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, INCLUDING
27 * LOST PROFITS, ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS
28 * DOCUMENTATION, EVEN IF THE AUTHOR OR DISTRIBUTORS HAVE BEEN ADVISED OF THE
29 * POSSIBILITY OF SUCH DAMAGE.
30 *
31 * THE AUTHORS AND DISTRIBUTORS SPECIFICALLY DISCLAIM ANY WARRANTIES,
32 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
33 * AND FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
34 * ON AN "AS IS" BASIS, AND THE AUTHOR AND DISTRIBUTORS HAS NO OBLIGATIONS TO
35 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
36 *
37 */
38
39 #include "postgres.h"
40
41 #include <ctype.h>
42
43 #include "mb/pg_wchar.h"
44 #include "utils/builtins.h"
45 #include "utils/varlena.h"
46
47 PG_MODULE_MAGIC;
48
49 /*
50 * Soundex
51 */
52 static void _soundex(const char *instr, char *outstr);
53
54 #define SOUNDEX_LEN 4
55
56 /* ABCDEFGHIJKLMNOPQRSTUVWXYZ */
57 static const char *soundex_table = "01230120022455012623010202";
58
59 static char
soundex_code(char letter)60 soundex_code(char letter)
61 {
62 letter = toupper((unsigned char) letter);
63 /* Defend against non-ASCII letters */
64 if (letter >= 'A' && letter <= 'Z')
65 return soundex_table[letter - 'A'];
66 return letter;
67 }
68
69 /*
70 * Metaphone
71 */
72 #define MAX_METAPHONE_STRLEN 255
73
74 /*
75 * Original code by Michael G Schwern starts here.
76 * Code slightly modified for use as PostgreSQL function.
77 */
78
79
80 /**************************************************************************
81 metaphone -- Breaks english phrases down into their phonemes.
82
83 Input
84 word -- An english word to be phonized
85 max_phonemes -- How many phonemes to calculate. If 0, then it
86 will phonize the entire phrase.
87 phoned_word -- The final phonized word. (We'll allocate the
88 memory.)
89 Output
90 error -- A simple error flag, returns true or false
91
92 NOTES: ALL non-alpha characters are ignored, this includes whitespace,
93 although non-alpha characters will break up phonemes.
94 ****************************************************************************/
95
96
97 /* I add modifications to the traditional metaphone algorithm that you
98 might find in books. Define this if you want metaphone to behave
99 traditionally */
100 #undef USE_TRADITIONAL_METAPHONE
101
102 /* Special encodings */
103 #define SH 'X'
104 #define TH '0'
105
106 static char Lookahead(char *word, int how_far);
107 static void _metaphone(char *word, int max_phonemes, char **phoned_word);
108
109 /* Metachar.h ... little bits about characters for metaphone */
110
111
112 /*-- Character encoding array & accessing macros --*/
113 /* Stolen directly out of the book... */
114 static const char _codes[26] = {
115 1, 16, 4, 16, 9, 2, 4, 16, 9, 2, 0, 2, 2, 2, 1, 4, 0, 2, 4, 4, 1, 0, 0, 0, 8, 0
116 /* a b c d e f g h i j k l m n o p q r s t u v w x y z */
117 };
118
119 static int
getcode(char c)120 getcode(char c)
121 {
122 if (isalpha((unsigned char) c))
123 {
124 c = toupper((unsigned char) c);
125 /* Defend against non-ASCII letters */
126 if (c >= 'A' && c <= 'Z')
127 return _codes[c - 'A'];
128 }
129 return 0;
130 }
131
132 #define isvowel(c) (getcode(c) & 1) /* AEIOU */
133
134 /* These letters are passed through unchanged */
135 #define NOCHANGE(c) (getcode(c) & 2) /* FJMNR */
136
137 /* These form diphthongs when preceding H */
138 #define AFFECTH(c) (getcode(c) & 4) /* CGPST */
139
140 /* These make C and G soft */
141 #define MAKESOFT(c) (getcode(c) & 8) /* EIY */
142
143 /* These prevent GH from becoming F */
144 #define NOGHTOF(c) (getcode(c) & 16) /* BDH */
145
146 PG_FUNCTION_INFO_V1(levenshtein_with_costs);
147 Datum
levenshtein_with_costs(PG_FUNCTION_ARGS)148 levenshtein_with_costs(PG_FUNCTION_ARGS)
149 {
150 text *src = PG_GETARG_TEXT_PP(0);
151 text *dst = PG_GETARG_TEXT_PP(1);
152 int ins_c = PG_GETARG_INT32(2);
153 int del_c = PG_GETARG_INT32(3);
154 int sub_c = PG_GETARG_INT32(4);
155 const char *s_data;
156 const char *t_data;
157 int s_bytes,
158 t_bytes;
159
160 /* Extract a pointer to the actual character data */
161 s_data = VARDATA_ANY(src);
162 t_data = VARDATA_ANY(dst);
163 /* Determine length of each string in bytes */
164 s_bytes = VARSIZE_ANY_EXHDR(src);
165 t_bytes = VARSIZE_ANY_EXHDR(dst);
166
167 PG_RETURN_INT32(varstr_levenshtein(s_data, s_bytes, t_data, t_bytes,
168 ins_c, del_c, sub_c, false));
169 }
170
171
172 PG_FUNCTION_INFO_V1(levenshtein);
173 Datum
levenshtein(PG_FUNCTION_ARGS)174 levenshtein(PG_FUNCTION_ARGS)
175 {
176 text *src = PG_GETARG_TEXT_PP(0);
177 text *dst = PG_GETARG_TEXT_PP(1);
178 const char *s_data;
179 const char *t_data;
180 int s_bytes,
181 t_bytes;
182
183 /* Extract a pointer to the actual character data */
184 s_data = VARDATA_ANY(src);
185 t_data = VARDATA_ANY(dst);
186 /* Determine length of each string in bytes */
187 s_bytes = VARSIZE_ANY_EXHDR(src);
188 t_bytes = VARSIZE_ANY_EXHDR(dst);
189
190 PG_RETURN_INT32(varstr_levenshtein(s_data, s_bytes, t_data, t_bytes,
191 1, 1, 1, false));
192 }
193
194
195 PG_FUNCTION_INFO_V1(levenshtein_less_equal_with_costs);
196 Datum
levenshtein_less_equal_with_costs(PG_FUNCTION_ARGS)197 levenshtein_less_equal_with_costs(PG_FUNCTION_ARGS)
198 {
199 text *src = PG_GETARG_TEXT_PP(0);
200 text *dst = PG_GETARG_TEXT_PP(1);
201 int ins_c = PG_GETARG_INT32(2);
202 int del_c = PG_GETARG_INT32(3);
203 int sub_c = PG_GETARG_INT32(4);
204 int max_d = PG_GETARG_INT32(5);
205 const char *s_data;
206 const char *t_data;
207 int s_bytes,
208 t_bytes;
209
210 /* Extract a pointer to the actual character data */
211 s_data = VARDATA_ANY(src);
212 t_data = VARDATA_ANY(dst);
213 /* Determine length of each string in bytes */
214 s_bytes = VARSIZE_ANY_EXHDR(src);
215 t_bytes = VARSIZE_ANY_EXHDR(dst);
216
217 PG_RETURN_INT32(varstr_levenshtein_less_equal(s_data, s_bytes,
218 t_data, t_bytes,
219 ins_c, del_c, sub_c,
220 max_d, false));
221 }
222
223
224 PG_FUNCTION_INFO_V1(levenshtein_less_equal);
225 Datum
levenshtein_less_equal(PG_FUNCTION_ARGS)226 levenshtein_less_equal(PG_FUNCTION_ARGS)
227 {
228 text *src = PG_GETARG_TEXT_PP(0);
229 text *dst = PG_GETARG_TEXT_PP(1);
230 int max_d = PG_GETARG_INT32(2);
231 const char *s_data;
232 const char *t_data;
233 int s_bytes,
234 t_bytes;
235
236 /* Extract a pointer to the actual character data */
237 s_data = VARDATA_ANY(src);
238 t_data = VARDATA_ANY(dst);
239 /* Determine length of each string in bytes */
240 s_bytes = VARSIZE_ANY_EXHDR(src);
241 t_bytes = VARSIZE_ANY_EXHDR(dst);
242
243 PG_RETURN_INT32(varstr_levenshtein_less_equal(s_data, s_bytes,
244 t_data, t_bytes,
245 1, 1, 1,
246 max_d, false));
247 }
248
249
250 /*
251 * Calculates the metaphone of an input string.
252 * Returns number of characters requested
253 * (suggested value is 4)
254 */
255 PG_FUNCTION_INFO_V1(metaphone);
256 Datum
metaphone(PG_FUNCTION_ARGS)257 metaphone(PG_FUNCTION_ARGS)
258 {
259 char *str_i = TextDatumGetCString(PG_GETARG_DATUM(0));
260 size_t str_i_len = strlen(str_i);
261 int reqlen;
262 char *metaph;
263
264 /* return an empty string if we receive one */
265 if (!(str_i_len > 0))
266 PG_RETURN_TEXT_P(cstring_to_text(""));
267
268 if (str_i_len > MAX_METAPHONE_STRLEN)
269 ereport(ERROR,
270 (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
271 errmsg("argument exceeds the maximum length of %d bytes",
272 MAX_METAPHONE_STRLEN)));
273
274 reqlen = PG_GETARG_INT32(1);
275 if (reqlen > MAX_METAPHONE_STRLEN)
276 ereport(ERROR,
277 (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
278 errmsg("output exceeds the maximum length of %d bytes",
279 MAX_METAPHONE_STRLEN)));
280
281 if (!(reqlen > 0))
282 ereport(ERROR,
283 (errcode(ERRCODE_ZERO_LENGTH_CHARACTER_STRING),
284 errmsg("output cannot be empty string")));
285
286 _metaphone(str_i, reqlen, &metaph);
287 PG_RETURN_TEXT_P(cstring_to_text(metaph));
288 }
289
290
291 /*
292 * Original code by Michael G Schwern starts here.
293 * Code slightly modified for use as PostgreSQL
294 * function (palloc, etc).
295 */
296
297 /* I suppose I could have been using a character pointer instead of
298 * accessing the array directly... */
299
300 /* Look at the next letter in the word */
301 #define Next_Letter (toupper((unsigned char) word[w_idx+1]))
302 /* Look at the current letter in the word */
303 #define Curr_Letter (toupper((unsigned char) word[w_idx]))
304 /* Go N letters back. */
305 #define Look_Back_Letter(n) \
306 (w_idx >= (n) ? toupper((unsigned char) word[w_idx-(n)]) : '\0')
307 /* Previous letter. I dunno, should this return null on failure? */
308 #define Prev_Letter (Look_Back_Letter(1))
309 /* Look two letters down. It makes sure you don't walk off the string. */
310 #define After_Next_Letter \
311 (Next_Letter != '\0' ? toupper((unsigned char) word[w_idx+2]) : '\0')
312 #define Look_Ahead_Letter(n) toupper((unsigned char) Lookahead(word+w_idx, n))
313
314
315 /* Allows us to safely look ahead an arbitrary # of letters */
316 /* I probably could have just used strlen... */
317 static char
Lookahead(char * word,int how_far)318 Lookahead(char *word, int how_far)
319 {
320 char letter_ahead = '\0'; /* null by default */
321 int idx;
322
323 for (idx = 0; word[idx] != '\0' && idx < how_far; idx++);
324 /* Edge forward in the string... */
325
326 letter_ahead = word[idx]; /* idx will be either == to how_far or at the
327 * end of the string */
328 return letter_ahead;
329 }
330
331
332 /* phonize one letter */
333 #define Phonize(c) do {(*phoned_word)[p_idx++] = c;} while (0)
334 /* Slap a null character on the end of the phoned word */
335 #define End_Phoned_Word do {(*phoned_word)[p_idx] = '\0';} while (0)
336 /* How long is the phoned word? */
337 #define Phone_Len (p_idx)
338
339 /* Note is a letter is a 'break' in the word */
340 #define Isbreak(c) (!isalpha((unsigned char) (c)))
341
342
343 static void
_metaphone(char * word,int max_phonemes,char ** phoned_word)344 _metaphone(char *word, /* IN */
345 int max_phonemes,
346 char **phoned_word) /* OUT */
347 {
348 int w_idx = 0; /* point in the phonization we're at. */
349 int p_idx = 0; /* end of the phoned phrase */
350
351 /*-- Parameter checks --*/
352
353 /*
354 * Shouldn't be necessary, but left these here anyway jec Aug 3, 2001
355 */
356
357 /* Negative phoneme length is meaningless */
358 if (!(max_phonemes > 0))
359 /* internal error */
360 elog(ERROR, "metaphone: Requested output length must be > 0");
361
362 /* Empty/null string is meaningless */
363 if ((word == NULL) || !(strlen(word) > 0))
364 /* internal error */
365 elog(ERROR, "metaphone: Input string length must be > 0");
366
367 /*-- Allocate memory for our phoned_phrase --*/
368 if (max_phonemes == 0)
369 { /* Assume largest possible */
370 *phoned_word = palloc(sizeof(char) * strlen(word) + 1);
371 }
372 else
373 {
374 *phoned_word = palloc(sizeof(char) * max_phonemes + 1);
375 }
376
377 /*-- The first phoneme has to be processed specially. --*/
378 /* Find our first letter */
379 for (; !isalpha((unsigned char) (Curr_Letter)); w_idx++)
380 {
381 /* On the off chance we were given nothing but crap... */
382 if (Curr_Letter == '\0')
383 {
384 End_Phoned_Word;
385 return;
386 }
387 }
388
389 switch (Curr_Letter)
390 {
391 /* AE becomes E */
392 case 'A':
393 if (Next_Letter == 'E')
394 {
395 Phonize('E');
396 w_idx += 2;
397 }
398 /* Remember, preserve vowels at the beginning */
399 else
400 {
401 Phonize('A');
402 w_idx++;
403 }
404 break;
405 /* [GKP]N becomes N */
406 case 'G':
407 case 'K':
408 case 'P':
409 if (Next_Letter == 'N')
410 {
411 Phonize('N');
412 w_idx += 2;
413 }
414 break;
415
416 /*
417 * WH becomes H, WR becomes R W if followed by a vowel
418 */
419 case 'W':
420 if (Next_Letter == 'H' ||
421 Next_Letter == 'R')
422 {
423 Phonize(Next_Letter);
424 w_idx += 2;
425 }
426 else if (isvowel(Next_Letter))
427 {
428 Phonize('W');
429 w_idx += 2;
430 }
431 /* else ignore */
432 break;
433 /* X becomes S */
434 case 'X':
435 Phonize('S');
436 w_idx++;
437 break;
438 /* Vowels are kept */
439
440 /*
441 * We did A already case 'A': case 'a':
442 */
443 case 'E':
444 case 'I':
445 case 'O':
446 case 'U':
447 Phonize(Curr_Letter);
448 w_idx++;
449 break;
450 default:
451 /* do nothing */
452 break;
453 }
454
455
456
457 /* On to the metaphoning */
458 for (; Curr_Letter != '\0' &&
459 (max_phonemes == 0 || Phone_Len < max_phonemes);
460 w_idx++)
461 {
462 /*
463 * How many letters to skip because an earlier encoding handled
464 * multiple letters
465 */
466 unsigned short int skip_letter = 0;
467
468
469 /*
470 * THOUGHT: It would be nice if, rather than having things like...
471 * well, SCI. For SCI you encode the S, then have to remember to skip
472 * the C. So the phonome SCI invades both S and C. It would be
473 * better, IMHO, to skip the C from the S part of the encoding. Hell,
474 * I'm trying it.
475 */
476
477 /* Ignore non-alphas */
478 if (!isalpha((unsigned char) (Curr_Letter)))
479 continue;
480
481 /* Drop duplicates, except CC */
482 if (Curr_Letter == Prev_Letter &&
483 Curr_Letter != 'C')
484 continue;
485
486 switch (Curr_Letter)
487 {
488 /* B -> B unless in MB */
489 case 'B':
490 if (Prev_Letter != 'M')
491 Phonize('B');
492 break;
493
494 /*
495 * 'sh' if -CIA- or -CH, but not SCH, except SCHW. (SCHW is
496 * handled in S) S if -CI-, -CE- or -CY- dropped if -SCI-,
497 * SCE-, -SCY- (handed in S) else K
498 */
499 case 'C':
500 if (MAKESOFT(Next_Letter))
501 { /* C[IEY] */
502 if (After_Next_Letter == 'A' &&
503 Next_Letter == 'I')
504 { /* CIA */
505 Phonize(SH);
506 }
507 /* SC[IEY] */
508 else if (Prev_Letter == 'S')
509 {
510 /* Dropped */
511 }
512 else
513 Phonize('S');
514 }
515 else if (Next_Letter == 'H')
516 {
517 #ifndef USE_TRADITIONAL_METAPHONE
518 if (After_Next_Letter == 'R' ||
519 Prev_Letter == 'S')
520 { /* Christ, School */
521 Phonize('K');
522 }
523 else
524 Phonize(SH);
525 #else
526 Phonize(SH);
527 #endif
528 skip_letter++;
529 }
530 else
531 Phonize('K');
532 break;
533
534 /*
535 * J if in -DGE-, -DGI- or -DGY- else T
536 */
537 case 'D':
538 if (Next_Letter == 'G' &&
539 MAKESOFT(After_Next_Letter))
540 {
541 Phonize('J');
542 skip_letter++;
543 }
544 else
545 Phonize('T');
546 break;
547
548 /*
549 * F if in -GH and not B--GH, D--GH, -H--GH, -H---GH else
550 * dropped if -GNED, -GN, else dropped if -DGE-, -DGI- or
551 * -DGY- (handled in D) else J if in -GE-, -GI, -GY and not GG
552 * else K
553 */
554 case 'G':
555 if (Next_Letter == 'H')
556 {
557 if (!(NOGHTOF(Look_Back_Letter(3)) ||
558 Look_Back_Letter(4) == 'H'))
559 {
560 Phonize('F');
561 skip_letter++;
562 }
563 else
564 {
565 /* silent */
566 }
567 }
568 else if (Next_Letter == 'N')
569 {
570 if (Isbreak(After_Next_Letter) ||
571 (After_Next_Letter == 'E' &&
572 Look_Ahead_Letter(3) == 'D'))
573 {
574 /* dropped */
575 }
576 else
577 Phonize('K');
578 }
579 else if (MAKESOFT(Next_Letter) &&
580 Prev_Letter != 'G')
581 Phonize('J');
582 else
583 Phonize('K');
584 break;
585 /* H if before a vowel and not after C,G,P,S,T */
586 case 'H':
587 if (isvowel(Next_Letter) &&
588 !AFFECTH(Prev_Letter))
589 Phonize('H');
590 break;
591
592 /*
593 * dropped if after C else K
594 */
595 case 'K':
596 if (Prev_Letter != 'C')
597 Phonize('K');
598 break;
599
600 /*
601 * F if before H else P
602 */
603 case 'P':
604 if (Next_Letter == 'H')
605 Phonize('F');
606 else
607 Phonize('P');
608 break;
609
610 /*
611 * K
612 */
613 case 'Q':
614 Phonize('K');
615 break;
616
617 /*
618 * 'sh' in -SH-, -SIO- or -SIA- or -SCHW- else S
619 */
620 case 'S':
621 if (Next_Letter == 'I' &&
622 (After_Next_Letter == 'O' ||
623 After_Next_Letter == 'A'))
624 Phonize(SH);
625 else if (Next_Letter == 'H')
626 {
627 Phonize(SH);
628 skip_letter++;
629 }
630 #ifndef USE_TRADITIONAL_METAPHONE
631 else if (Next_Letter == 'C' &&
632 Look_Ahead_Letter(2) == 'H' &&
633 Look_Ahead_Letter(3) == 'W')
634 {
635 Phonize(SH);
636 skip_letter += 2;
637 }
638 #endif
639 else
640 Phonize('S');
641 break;
642
643 /*
644 * 'sh' in -TIA- or -TIO- else 'th' before H else T
645 */
646 case 'T':
647 if (Next_Letter == 'I' &&
648 (After_Next_Letter == 'O' ||
649 After_Next_Letter == 'A'))
650 Phonize(SH);
651 else if (Next_Letter == 'H')
652 {
653 Phonize(TH);
654 skip_letter++;
655 }
656 else
657 Phonize('T');
658 break;
659 /* F */
660 case 'V':
661 Phonize('F');
662 break;
663 /* W before a vowel, else dropped */
664 case 'W':
665 if (isvowel(Next_Letter))
666 Phonize('W');
667 break;
668 /* KS */
669 case 'X':
670 Phonize('K');
671 if (max_phonemes == 0 || Phone_Len < max_phonemes)
672 Phonize('S');
673 break;
674 /* Y if followed by a vowel */
675 case 'Y':
676 if (isvowel(Next_Letter))
677 Phonize('Y');
678 break;
679 /* S */
680 case 'Z':
681 Phonize('S');
682 break;
683 /* No transformation */
684 case 'F':
685 case 'J':
686 case 'L':
687 case 'M':
688 case 'N':
689 case 'R':
690 Phonize(Curr_Letter);
691 break;
692 default:
693 /* nothing */
694 break;
695 } /* END SWITCH */
696
697 w_idx += skip_letter;
698 } /* END FOR */
699
700 End_Phoned_Word;
701
702 return;
703 } /* END metaphone */
704
705
706 /*
707 * SQL function: soundex(text) returns text
708 */
709 PG_FUNCTION_INFO_V1(soundex);
710
711 Datum
soundex(PG_FUNCTION_ARGS)712 soundex(PG_FUNCTION_ARGS)
713 {
714 char outstr[SOUNDEX_LEN + 1];
715 char *arg;
716
717 arg = text_to_cstring(PG_GETARG_TEXT_PP(0));
718
719 _soundex(arg, outstr);
720
721 PG_RETURN_TEXT_P(cstring_to_text(outstr));
722 }
723
724 static void
_soundex(const char * instr,char * outstr)725 _soundex(const char *instr, char *outstr)
726 {
727 int count;
728
729 AssertArg(instr);
730 AssertArg(outstr);
731
732 outstr[SOUNDEX_LEN] = '\0';
733
734 /* Skip leading non-alphabetic characters */
735 while (!isalpha((unsigned char) instr[0]) && instr[0])
736 ++instr;
737
738 /* No string left */
739 if (!instr[0])
740 {
741 outstr[0] = (char) 0;
742 return;
743 }
744
745 /* Take the first letter as is */
746 *outstr++ = (char) toupper((unsigned char) *instr++);
747
748 count = 1;
749 while (*instr && count < SOUNDEX_LEN)
750 {
751 if (isalpha((unsigned char) *instr) &&
752 soundex_code(*instr) != soundex_code(*(instr - 1)))
753 {
754 *outstr = soundex_code(instr[0]);
755 if (*outstr != '0')
756 {
757 ++outstr;
758 ++count;
759 }
760 }
761 ++instr;
762 }
763
764 /* Fill with 0's */
765 while (count < SOUNDEX_LEN)
766 {
767 *outstr = '0';
768 ++outstr;
769 ++count;
770 }
771 }
772
773 PG_FUNCTION_INFO_V1(difference);
774
775 Datum
difference(PG_FUNCTION_ARGS)776 difference(PG_FUNCTION_ARGS)
777 {
778 char sndx1[SOUNDEX_LEN + 1],
779 sndx2[SOUNDEX_LEN + 1];
780 int i,
781 result;
782
783 _soundex(text_to_cstring(PG_GETARG_TEXT_PP(0)), sndx1);
784 _soundex(text_to_cstring(PG_GETARG_TEXT_PP(1)), sndx2);
785
786 result = 0;
787 for (i = 0; i < SOUNDEX_LEN; i++)
788 {
789 if (sndx1[i] == sndx2[i])
790 result++;
791 }
792
793 PG_RETURN_INT32(result);
794 }
795