1
by brian
clean slate |
1 |
/* Copyright (C) 2000-2004, 2006 MySQL AB
|
2 |
||
3 |
This program is free software; you can redistribute it and/or modify
|
|
4 |
it under the terms of the GNU General Public License as published by
|
|
5 |
the Free Software Foundation; version 2 of the License.
|
|
6 |
||
7 |
This program is distributed in the hope that it will be useful,
|
|
8 |
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
9 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
10 |
GNU General Public License for more details.
|
|
11 |
||
12 |
You should have received a copy of the GNU General Public License
|
|
13 |
along with this program; if not, write to the Free Software
|
|
14 |
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
|
|
15 |
||
16 |
/* Written by Sergei A. Golubchik, who has a shared copyright to this code */
|
|
17 |
||
18 |
/* functions to work with full-text indices */
|
|
19 |
||
20 |
#include "ftdefs.h" |
|
21 |
#include <math.h> |
|
22 |
||
23 |
void _mi_ft_segiterator_init(MI_INFO *info, uint keynr, const uchar *record, |
|
24 |
FT_SEG_ITERATOR *ftsi) |
|
25 |
{
|
|
26 |
DBUG_ENTER("_mi_ft_segiterator_init"); |
|
27 |
||
28 |
ftsi->num=info->s->keyinfo[keynr].keysegs; |
|
29 |
ftsi->seg=info->s->keyinfo[keynr].seg; |
|
30 |
ftsi->rec=record; |
|
31 |
DBUG_VOID_RETURN; |
|
32 |
}
|
|
33 |
||
34 |
void _mi_ft_segiterator_dummy_init(const uchar *record, uint len, |
|
35 |
FT_SEG_ITERATOR *ftsi) |
|
36 |
{
|
|
37 |
DBUG_ENTER("_mi_ft_segiterator_dummy_init"); |
|
38 |
||
39 |
ftsi->num=1; |
|
40 |
ftsi->seg=0; |
|
41 |
ftsi->pos=record; |
|
42 |
ftsi->len=len; |
|
43 |
DBUG_VOID_RETURN; |
|
44 |
}
|
|
45 |
||
46 |
/*
|
|
47 |
This function breaks convention "return 0 in success"
|
|
48 |
but it's easier to use like this
|
|
49 |
||
50 |
while(_mi_ft_segiterator())
|
|
51 |
||
52 |
so "1" means "OK", "0" means "EOF"
|
|
53 |
*/
|
|
54 |
||
55 |
uint _mi_ft_segiterator(register FT_SEG_ITERATOR *ftsi) |
|
56 |
{
|
|
57 |
DBUG_ENTER("_mi_ft_segiterator"); |
|
58 |
||
59 |
if (!ftsi->num) |
|
60 |
DBUG_RETURN(0); |
|
61 |
||
62 |
ftsi->num--; |
|
63 |
if (!ftsi->seg) |
|
64 |
DBUG_RETURN(1); |
|
65 |
||
66 |
ftsi->seg--; |
|
67 |
||
68 |
if (ftsi->seg->null_bit && |
|
69 |
(ftsi->rec[ftsi->seg->null_pos] & ftsi->seg->null_bit)) |
|
70 |
{
|
|
71 |
ftsi->pos=0; |
|
72 |
DBUG_RETURN(1); |
|
73 |
}
|
|
74 |
ftsi->pos= ftsi->rec+ftsi->seg->start; |
|
75 |
if (ftsi->seg->flag & HA_VAR_LENGTH_PART) |
|
76 |
{
|
|
77 |
uint pack_length= (ftsi->seg->bit_start); |
|
78 |
ftsi->len= (pack_length == 1 ? (uint) *(uchar*) ftsi->pos : |
|
79 |
uint2korr(ftsi->pos)); |
|
80 |
ftsi->pos+= pack_length; /* Skip VARCHAR length */ |
|
81 |
DBUG_RETURN(1); |
|
82 |
}
|
|
83 |
if (ftsi->seg->flag & HA_BLOB_PART) |
|
84 |
{
|
|
85 |
ftsi->len=_mi_calc_blob_length(ftsi->seg->bit_start,ftsi->pos); |
|
86 |
memcpy_fixed((char*) &ftsi->pos, ftsi->pos+ftsi->seg->bit_start, |
|
87 |
sizeof(char*)); |
|
88 |
DBUG_RETURN(1); |
|
89 |
}
|
|
90 |
ftsi->len=ftsi->seg->length; |
|
91 |
DBUG_RETURN(1); |
|
92 |
}
|
|
93 |
||
94 |
||
95 |
/* parses a document i.e. calls ft_parse for every keyseg */
|
|
96 |
||
97 |
uint _mi_ft_parse(TREE *parsed, MI_INFO *info, uint keynr, const uchar *record, |
|
98 |
MYSQL_FTPARSER_PARAM *param, MEM_ROOT *mem_root) |
|
99 |
{
|
|
100 |
FT_SEG_ITERATOR ftsi; |
|
101 |
struct st_mysql_ftparser *parser; |
|
102 |
DBUG_ENTER("_mi_ft_parse"); |
|
103 |
||
104 |
_mi_ft_segiterator_init(info, keynr, record, &ftsi); |
|
105 |
||
106 |
ft_parse_init(parsed, info->s->keyinfo[keynr].seg->charset); |
|
107 |
parser= info->s->keyinfo[keynr].parser; |
|
108 |
while (_mi_ft_segiterator(&ftsi)) |
|
109 |
{
|
|
110 |
if (ftsi.pos) |
|
111 |
if (ft_parse(parsed, (uchar *)ftsi.pos, ftsi.len, parser, param, mem_root)) |
|
112 |
DBUG_RETURN(1); |
|
113 |
}
|
|
114 |
DBUG_RETURN(0); |
|
115 |
}
|
|
116 |
||
117 |
FT_WORD *_mi_ft_parserecord(MI_INFO *info, uint keynr, const uchar *record, |
|
118 |
MEM_ROOT *mem_root) |
|
119 |
{
|
|
120 |
TREE ptree; |
|
121 |
MYSQL_FTPARSER_PARAM *param; |
|
122 |
DBUG_ENTER("_mi_ft_parserecord"); |
|
123 |
if (! (param= ftparser_call_initializer(info, keynr, 0))) |
|
124 |
DBUG_RETURN(NULL); |
|
125 |
bzero((char*) &ptree, sizeof(ptree)); |
|
126 |
param->flags= 0; |
|
127 |
if (_mi_ft_parse(&ptree, info, keynr, record, param, mem_root)) |
|
128 |
DBUG_RETURN(NULL); |
|
129 |
||
130 |
DBUG_RETURN(ft_linearize(&ptree, mem_root)); |
|
131 |
}
|
|
132 |
||
133 |
static int _mi_ft_store(MI_INFO *info, uint keynr, uchar *keybuf, |
|
134 |
FT_WORD *wlist, my_off_t filepos) |
|
135 |
{
|
|
136 |
uint key_length; |
|
137 |
DBUG_ENTER("_mi_ft_store"); |
|
138 |
||
139 |
for (; wlist->pos; wlist++) |
|
140 |
{
|
|
141 |
key_length=_ft_make_key(info,keynr,keybuf,wlist,filepos); |
|
142 |
if (_mi_ck_write(info,keynr,(uchar*) keybuf,key_length)) |
|
143 |
DBUG_RETURN(1); |
|
144 |
}
|
|
145 |
DBUG_RETURN(0); |
|
146 |
}
|
|
147 |
||
148 |
static int _mi_ft_erase(MI_INFO *info, uint keynr, uchar *keybuf, |
|
149 |
FT_WORD *wlist, my_off_t filepos) |
|
150 |
{
|
|
151 |
uint key_length, err=0; |
|
152 |
DBUG_ENTER("_mi_ft_erase"); |
|
153 |
||
154 |
for (; wlist->pos; wlist++) |
|
155 |
{
|
|
156 |
key_length=_ft_make_key(info,keynr,keybuf,wlist,filepos); |
|
157 |
if (_mi_ck_delete(info,keynr,(uchar*) keybuf,key_length)) |
|
158 |
err=1; |
|
159 |
}
|
|
160 |
DBUG_RETURN(err); |
|
161 |
}
|
|
162 |
||
163 |
/*
|
|
164 |
Compares an appropriate parts of two WORD_KEY keys directly out of records
|
|
165 |
returns 1 if they are different
|
|
166 |
*/
|
|
167 |
||
168 |
#define THOSE_TWO_DAMN_KEYS_ARE_REALLY_DIFFERENT 1
|
|
169 |
#define GEE_THEY_ARE_ABSOLUTELY_IDENTICAL 0
|
|
170 |
||
171 |
int _mi_ft_cmp(MI_INFO *info, uint keynr, const uchar *rec1, const uchar *rec2) |
|
172 |
{
|
|
173 |
FT_SEG_ITERATOR ftsi1, ftsi2; |
|
174 |
CHARSET_INFO *cs=info->s->keyinfo[keynr].seg->charset; |
|
175 |
DBUG_ENTER("_mi_ft_cmp"); |
|
176 |
_mi_ft_segiterator_init(info, keynr, rec1, &ftsi1); |
|
177 |
_mi_ft_segiterator_init(info, keynr, rec2, &ftsi2); |
|
178 |
||
179 |
while (_mi_ft_segiterator(&ftsi1) && _mi_ft_segiterator(&ftsi2)) |
|
180 |
{
|
|
181 |
if ((ftsi1.pos != ftsi2.pos) && |
|
182 |
(!ftsi1.pos || !ftsi2.pos || |
|
183 |
ha_compare_text(cs, (uchar*) ftsi1.pos,ftsi1.len, |
|
184 |
(uchar*) ftsi2.pos,ftsi2.len,0,0))) |
|
185 |
DBUG_RETURN(THOSE_TWO_DAMN_KEYS_ARE_REALLY_DIFFERENT); |
|
186 |
}
|
|
187 |
DBUG_RETURN(GEE_THEY_ARE_ABSOLUTELY_IDENTICAL); |
|
188 |
}
|
|
189 |
||
190 |
||
191 |
/* update a document entry */
|
|
192 |
||
193 |
int _mi_ft_update(MI_INFO *info, uint keynr, uchar *keybuf, |
|
194 |
const uchar *oldrec, const uchar *newrec, my_off_t pos) |
|
195 |
{
|
|
196 |
int error= -1; |
|
197 |
FT_WORD *oldlist,*newlist, *old_word, *new_word; |
|
198 |
CHARSET_INFO *cs=info->s->keyinfo[keynr].seg->charset; |
|
199 |
uint key_length; |
|
200 |
int cmp, cmp2; |
|
201 |
DBUG_ENTER("_mi_ft_update"); |
|
202 |
||
203 |
if (!(old_word=oldlist=_mi_ft_parserecord(info, keynr, oldrec, |
|
204 |
&info->ft_memroot)) || |
|
205 |
!(new_word=newlist=_mi_ft_parserecord(info, keynr, newrec, |
|
206 |
&info->ft_memroot))) |
|
207 |
goto err; |
|
208 |
||
209 |
error=0; |
|
210 |
while(old_word->pos && new_word->pos) |
|
211 |
{
|
|
212 |
cmp= ha_compare_text(cs, (uchar*) old_word->pos,old_word->len, |
|
213 |
(uchar*) new_word->pos,new_word->len,0,0); |
|
214 |
cmp2= cmp ? 0 : (fabs(old_word->weight - new_word->weight) > 1.e-5); |
|
215 |
||
216 |
if (cmp < 0 || cmp2) |
|
217 |
{
|
|
218 |
key_length=_ft_make_key(info,keynr,keybuf,old_word,pos); |
|
219 |
if ((error=_mi_ck_delete(info,keynr,(uchar*) keybuf,key_length))) |
|
220 |
goto err; |
|
221 |
}
|
|
222 |
if (cmp > 0 || cmp2) |
|
223 |
{
|
|
224 |
key_length=_ft_make_key(info,keynr,keybuf,new_word,pos); |
|
225 |
if ((error=_mi_ck_write(info,keynr,(uchar*) keybuf,key_length))) |
|
226 |
goto err; |
|
227 |
}
|
|
228 |
if (cmp<=0) old_word++; |
|
229 |
if (cmp>=0) new_word++; |
|
230 |
}
|
|
231 |
if (old_word->pos) |
|
232 |
error=_mi_ft_erase(info,keynr,keybuf,old_word,pos); |
|
233 |
else if (new_word->pos) |
|
234 |
error=_mi_ft_store(info,keynr,keybuf,new_word,pos); |
|
235 |
||
236 |
err: |
|
237 |
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE)); |
|
238 |
DBUG_RETURN(error); |
|
239 |
}
|
|
240 |
||
241 |
||
242 |
/* adds a document to the collection */
|
|
243 |
||
244 |
int _mi_ft_add(MI_INFO *info, uint keynr, uchar *keybuf, const uchar *record, |
|
245 |
my_off_t pos) |
|
246 |
{
|
|
247 |
int error= -1; |
|
248 |
FT_WORD *wlist; |
|
249 |
DBUG_ENTER("_mi_ft_add"); |
|
250 |
DBUG_PRINT("enter",("keynr: %d",keynr)); |
|
251 |
||
252 |
if ((wlist=_mi_ft_parserecord(info, keynr, record, &info->ft_memroot))) |
|
253 |
error=_mi_ft_store(info,keynr,keybuf,wlist,pos); |
|
254 |
||
255 |
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE)); |
|
256 |
DBUG_PRINT("exit",("Return: %d",error)); |
|
257 |
DBUG_RETURN(error); |
|
258 |
}
|
|
259 |
||
260 |
||
261 |
/* removes a document from the collection */
|
|
262 |
||
263 |
int _mi_ft_del(MI_INFO *info, uint keynr, uchar *keybuf, const uchar *record, |
|
264 |
my_off_t pos) |
|
265 |
{
|
|
266 |
int error= -1; |
|
267 |
FT_WORD *wlist; |
|
268 |
DBUG_ENTER("_mi_ft_del"); |
|
269 |
DBUG_PRINT("enter",("keynr: %d",keynr)); |
|
270 |
||
271 |
if ((wlist=_mi_ft_parserecord(info, keynr, record, &info->ft_memroot))) |
|
272 |
error=_mi_ft_erase(info,keynr,keybuf,wlist,pos); |
|
273 |
||
274 |
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE)); |
|
275 |
DBUG_PRINT("exit",("Return: %d",error)); |
|
276 |
DBUG_RETURN(error); |
|
277 |
}
|
|
278 |
||
279 |
uint _ft_make_key(MI_INFO *info, uint keynr, uchar *keybuf, FT_WORD *wptr, |
|
280 |
my_off_t filepos) |
|
281 |
{
|
|
282 |
uchar buf[HA_FT_MAXBYTELEN+16]; |
|
283 |
DBUG_ENTER("_ft_make_key"); |
|
284 |
||
285 |
#if HA_FT_WTYPE == HA_KEYTYPE_FLOAT
|
|
286 |
{
|
|
287 |
float weight=(float) ((filepos==HA_OFFSET_ERROR) ? 0 : wptr->weight); |
|
288 |
mi_float4store(buf,weight); |
|
289 |
}
|
|
290 |
#else
|
|
291 |
#error
|
|
292 |
#endif
|
|
293 |
||
294 |
int2store(buf+HA_FT_WLEN,wptr->len); |
|
295 |
memcpy(buf+HA_FT_WLEN+2,wptr->pos,wptr->len); |
|
296 |
DBUG_RETURN(_mi_make_key(info,keynr,(uchar*) keybuf,buf,filepos)); |
|
297 |
}
|
|
298 |
||
299 |
||
300 |
/*
|
|
301 |
convert key value to ft2
|
|
302 |
*/
|
|
303 |
||
304 |
uint _mi_ft_convert_to_ft2(MI_INFO *info, uint keynr, uchar *key) |
|
305 |
{
|
|
306 |
my_off_t root; |
|
307 |
DYNAMIC_ARRAY *da=info->ft1_to_ft2; |
|
308 |
MI_KEYDEF *keyinfo=&info->s->ft2_keyinfo; |
|
309 |
uchar *key_ptr= (uchar*) dynamic_array_ptr(da, 0), *end; |
|
310 |
uint length, key_length; |
|
311 |
DBUG_ENTER("_mi_ft_convert_to_ft2"); |
|
312 |
||
313 |
/* we'll generate one pageful at once, and insert the rest one-by-one */
|
|
314 |
/* calculating the length of this page ...*/
|
|
315 |
length=(keyinfo->block_length-2) / keyinfo->keylength; |
|
316 |
set_if_smaller(length, da->elements); |
|
317 |
length=length * keyinfo->keylength; |
|
318 |
||
319 |
get_key_full_length_rdonly(key_length, key); |
|
320 |
while (_mi_ck_delete(info, keynr, key, key_length) == 0) |
|
321 |
{
|
|
322 |
/*
|
|
323 |
nothing to do here.
|
|
324 |
_mi_ck_delete() will populate info->ft1_to_ft2 with deleted keys
|
|
325 |
*/
|
|
326 |
}
|
|
327 |
||
328 |
/* creating pageful of keys */
|
|
329 |
mi_putint(info->buff,length+2,0); |
|
330 |
memcpy(info->buff+2, key_ptr, length); |
|
331 |
info->buff_used=info->page_changed=1; /* info->buff is used */ |
|
332 |
if ((root= _mi_new(info,keyinfo,DFLT_INIT_HITS)) == HA_OFFSET_ERROR || |
|
333 |
_mi_write_keypage(info,keyinfo,root,DFLT_INIT_HITS,info->buff)) |
|
334 |
DBUG_RETURN(-1); |
|
335 |
||
336 |
/* inserting the rest of key values */
|
|
337 |
end= (uchar*) dynamic_array_ptr(da, da->elements); |
|
338 |
for (key_ptr+=length; key_ptr < end; key_ptr+=keyinfo->keylength) |
|
339 |
if(_mi_ck_real_write_btree(info, keyinfo, key_ptr, 0, &root, SEARCH_SAME)) |
|
340 |
DBUG_RETURN(-1); |
|
341 |
||
342 |
/* now, writing the word key entry */
|
|
343 |
ft_intXstore(key+key_length, - (int) da->elements); |
|
344 |
_mi_dpointer(info, key+key_length+HA_FT_WLEN, root); |
|
345 |
||
346 |
DBUG_RETURN(_mi_ck_real_write_btree(info, |
|
347 |
info->s->keyinfo+keynr, |
|
348 |
key, 0, |
|
349 |
&info->s->state.key_root[keynr], |
|
350 |
SEARCH_SAME)); |
|
351 |
}
|
|
352 |