1
/* Copyright (C) 2006 MySQL AB & MySQL Finland AB & TCX DataKonsult AB
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.
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.
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 */
16
/* Written by Sergei A. Golubchik, who has a shared copyright to this code */
18
/* functions to work with full-text indices */
20
#include "ma_ftdefs.h"
23
void _ma_ft_segiterator_init(MARIA_HA *info, uint keynr, const uchar *record,
24
FT_SEG_ITERATOR *ftsi)
26
DBUG_ENTER("_ma_ft_segiterator_init");
28
ftsi->num=info->s->keyinfo[keynr].keysegs;
29
ftsi->seg=info->s->keyinfo[keynr].seg;
34
void _ma_ft_segiterator_dummy_init(const uchar *record, uint len,
35
FT_SEG_ITERATOR *ftsi)
37
DBUG_ENTER("_ma_ft_segiterator_dummy_init");
47
This function breaks convention "return 0 in success"
48
but it's easier to use like this
50
while(_ma_ft_segiterator())
52
so "1" means "OK", "0" means "EOF"
55
uint _ma_ft_segiterator(register FT_SEG_ITERATOR *ftsi)
57
DBUG_ENTER("_ma_ft_segiterator");
68
if (ftsi->seg->null_bit &&
69
(ftsi->rec[ftsi->seg->null_pos] & ftsi->seg->null_bit))
74
ftsi->pos= ftsi->rec+ftsi->seg->start;
75
if (ftsi->seg->flag & HA_VAR_LENGTH_PART)
77
uint pack_length= (ftsi->seg->bit_start);
78
ftsi->len= (pack_length == 1 ? (uint) * ftsi->pos :
79
uint2korr(ftsi->pos));
80
ftsi->pos+= pack_length; /* Skip VARCHAR length */
83
if (ftsi->seg->flag & HA_BLOB_PART)
85
ftsi->len= _ma_calc_blob_length(ftsi->seg->bit_start,ftsi->pos);
86
memcpy(&ftsi->pos, ftsi->pos+ftsi->seg->bit_start, sizeof(char*));
89
ftsi->len=ftsi->seg->length;
94
/* parses a document i.e. calls maria_ft_parse for every keyseg */
96
uint _ma_ft_parse(TREE *parsed, MARIA_HA *info, uint keynr, const uchar *record,
97
MYSQL_FTPARSER_PARAM *param, MEM_ROOT *mem_root)
100
struct st_mysql_ftparser *parser;
101
DBUG_ENTER("_ma_ft_parse");
103
_ma_ft_segiterator_init(info, keynr, record, &ftsi);
105
maria_ft_parse_init(parsed, info->s->keyinfo[keynr].seg->charset);
106
parser= info->s->keyinfo[keynr].parser;
107
while (_ma_ft_segiterator(&ftsi))
109
/** @todo this casts ftsi.pos (const) to non-const */
111
if (maria_ft_parse(parsed, (uchar *)ftsi.pos, ftsi.len, parser, param,
118
FT_WORD * _ma_ft_parserecord(MARIA_HA *info, uint keynr, const uchar *record,
122
MYSQL_FTPARSER_PARAM *param;
123
DBUG_ENTER("_ma_ft_parserecord");
124
if (! (param= maria_ftparser_call_initializer(info, keynr, 0)))
126
bzero((char*) &ptree, sizeof(ptree));
128
if (_ma_ft_parse(&ptree, info, keynr, record, param, mem_root))
131
DBUG_RETURN(maria_ft_linearize(&ptree, mem_root));
134
static int _ma_ft_store(MARIA_HA *info, uint keynr, uchar *keybuf,
135
FT_WORD *wlist, my_off_t filepos)
137
DBUG_ENTER("_ma_ft_store");
139
for (; wlist->pos; wlist++)
142
_ma_ft_make_key(info, &key, keynr, keybuf, wlist, filepos);
143
if (_ma_ck_write(info, &key))
149
static int _ma_ft_erase(MARIA_HA *info, uint keynr, uchar *keybuf,
150
FT_WORD *wlist, my_off_t filepos)
153
DBUG_ENTER("_ma_ft_erase");
155
for (; wlist->pos; wlist++)
158
_ma_ft_make_key(info, &key, keynr, keybuf, wlist, filepos);
159
if (_ma_ck_delete(info, &key))
166
Compares an appropriate parts of two WORD_KEY keys directly out of records
167
returns 1 if they are different
170
#define THOSE_TWO_DAMN_KEYS_ARE_REALLY_DIFFERENT 1
171
#define GEE_THEY_ARE_ABSOLUTELY_IDENTICAL 0
173
int _ma_ft_cmp(MARIA_HA *info, uint keynr, const uchar *rec1, const uchar *rec2)
175
FT_SEG_ITERATOR ftsi1, ftsi2;
176
CHARSET_INFO *cs=info->s->keyinfo[keynr].seg->charset;
177
DBUG_ENTER("_ma_ft_cmp");
179
_ma_ft_segiterator_init(info, keynr, rec1, &ftsi1);
180
_ma_ft_segiterator_init(info, keynr, rec2, &ftsi2);
182
while (_ma_ft_segiterator(&ftsi1) && _ma_ft_segiterator(&ftsi2))
184
if ((ftsi1.pos != ftsi2.pos) &&
185
(!ftsi1.pos || !ftsi2.pos ||
186
ha_compare_text(cs, ftsi1.pos,ftsi1.len,
187
ftsi2.pos,ftsi2.len,0,0)))
188
DBUG_RETURN(THOSE_TWO_DAMN_KEYS_ARE_REALLY_DIFFERENT);
190
DBUG_RETURN(GEE_THEY_ARE_ABSOLUTELY_IDENTICAL);
194
/* update a document entry */
196
int _ma_ft_update(MARIA_HA *info, uint keynr, uchar *keybuf,
197
const uchar *oldrec, const uchar *newrec, my_off_t pos)
200
FT_WORD *oldlist,*newlist, *old_word, *new_word;
201
CHARSET_INFO *cs=info->s->keyinfo[keynr].seg->charset;
203
DBUG_ENTER("_ma_ft_update");
205
if (!(old_word=oldlist=_ma_ft_parserecord(info, keynr, oldrec,
206
&info->ft_memroot)) ||
207
!(new_word=newlist=_ma_ft_parserecord(info, keynr, newrec,
212
while(old_word->pos && new_word->pos)
214
cmp= ha_compare_text(cs, (uchar*) old_word->pos,old_word->len,
215
(uchar*) new_word->pos,new_word->len,0,0);
216
cmp2= cmp ? 0 : (fabs(old_word->weight - new_word->weight) > 1.e-5);
221
_ma_ft_make_key(info, &key, keynr, keybuf, old_word, pos);
222
if (_ma_ck_delete(info, &key))
231
_ma_ft_make_key(info, &key, keynr, keybuf, new_word,pos);
232
if ((error= _ma_ck_write(info, &key)))
235
if (cmp<=0) old_word++;
236
if (cmp>=0) new_word++;
239
error= _ma_ft_erase(info,keynr,keybuf,old_word,pos);
240
else if (new_word->pos)
241
error= _ma_ft_store(info,keynr,keybuf,new_word,pos);
244
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE));
249
/* adds a document to the collection */
251
int _ma_ft_add(MARIA_HA *info, uint keynr, uchar *keybuf, const uchar *record,
256
DBUG_ENTER("_ma_ft_add");
257
DBUG_PRINT("enter",("keynr: %d",keynr));
259
if ((wlist= _ma_ft_parserecord(info, keynr, record, &info->ft_memroot)))
260
error= _ma_ft_store(info,keynr,keybuf,wlist,pos);
261
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE));
262
DBUG_PRINT("exit",("Return: %d",error));
267
/* removes a document from the collection */
269
int _ma_ft_del(MARIA_HA *info, uint keynr, uchar *keybuf, const uchar *record,
274
DBUG_ENTER("_ma_ft_del");
275
DBUG_PRINT("enter",("keynr: %d",keynr));
277
if ((wlist= _ma_ft_parserecord(info, keynr, record, &info->ft_memroot)))
278
error= _ma_ft_erase(info,keynr,keybuf,wlist,pos);
279
free_root(&info->ft_memroot, MYF(MY_MARK_BLOCKS_FREE));
280
DBUG_PRINT("exit",("Return: %d",error));
285
MARIA_KEY *_ma_ft_make_key(MARIA_HA *info, MARIA_KEY *key, uint keynr,
287
FT_WORD *wptr, my_off_t filepos)
289
uchar buf[HA_FT_MAXBYTELEN+16];
290
DBUG_ENTER("_ma_ft_make_key");
292
#if HA_FT_WTYPE == HA_KEYTYPE_FLOAT
294
float weight=(float) ((filepos==HA_OFFSET_ERROR) ? 0 : wptr->weight);
295
mi_float4store(buf,weight);
301
int2store(buf+HA_FT_WLEN,wptr->len);
302
memcpy(buf+HA_FT_WLEN+2,wptr->pos,wptr->len);
303
/* Can't be spatial so it's ok to call _ma_make_key directly here */
304
DBUG_RETURN(_ma_make_key(info, key, keynr, keybuf, buf, filepos, 0));
309
convert key value to ft2
312
my_bool _ma_ft_convert_to_ft2(MARIA_HA *info, MARIA_KEY *key)
314
MARIA_SHARE *share= info->s;
316
DYNAMIC_ARRAY *da=info->ft1_to_ft2;
317
MARIA_KEYDEF *keyinfo=&share->ft2_keyinfo;
318
uchar *key_ptr= (uchar*) dynamic_array_ptr(da, 0), *end;
319
uint length, key_length;
320
MARIA_PINNED_PAGE tmp_page_link, *page_link= &tmp_page_link;
323
DBUG_ENTER("_ma_ft_convert_to_ft2");
325
/* we'll generate one pageful at once, and insert the rest one-by-one */
326
/* calculating the length of this page ...*/
327
length=(keyinfo->block_length-2) / keyinfo->keylength;
328
set_if_smaller(length, da->elements);
329
length=length * keyinfo->keylength;
331
get_key_full_length_rdonly(key_length, key->data);
332
while (_ma_ck_delete(info, key) == 0)
336
_ma_ck_delete() will populate info->ft1_to_ft2 with deleted keys
340
/* creating pageful of keys */
341
bzero(info->buff, share->keypage_header);
342
_ma_store_keynr(share, info->buff, keyinfo->key_nr);
343
_ma_store_page_used(share, info->buff, length + share->keypage_header);
344
memcpy(info->buff + share->keypage_header, key_ptr, length);
345
info->keyread_buff_used= info->page_changed=1; /* info->buff is used */
347
@todo RECOVERY BUG this is not logged yet. Ok as this code is never
348
called, but soon it will be.
350
if ((root= _ma_new(info, DFLT_INIT_HITS, &page_link)) == HA_OFFSET_ERROR)
353
_ma_page_setup(&page, info, keyinfo, root, info->buff);
354
if (_ma_write_keypage(&page, page_link->write_lock, DFLT_INIT_HITS))
357
/* inserting the rest of key values */
358
end= (uchar*) dynamic_array_ptr(da, da->elements);
359
tmp_key.keyinfo= keyinfo;
360
tmp_key.data_length= keyinfo->keylength;
361
tmp_key.ref_length= 0;
363
for (key_ptr+=length; key_ptr < end; key_ptr+=keyinfo->keylength)
365
tmp_key.data= key_ptr;
366
if (_ma_ck_real_write_btree(info, &tmp_key, &root, SEARCH_SAME))
370
/* now, writing the word key entry */
371
ft_intXstore(key->data + key_length, - (int) da->elements);
372
_ma_dpointer(share, key->data + key_length + HA_FT_WLEN, root);
374
DBUG_RETURN(_ma_ck_real_write_btree(info, key,
375
&share->state.key_root[key->keyinfo->