1 /* $Id: trunc.c,v 1.56 2005-04-20 10:21:29 adam Exp $
2 Copyright (C) 1995-2005
5 This file is part of the Zebra server.
7 Zebra is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with Zebra; see the file LICENSE.zebra. If not, write to the
19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
35 int (*cmp)(const void *p1, const void *p2);
42 static void heap_swap(struct trunc_info *ti, int i1, int i2)
47 ti->ptr[i1] = ti->ptr[i2];
51 static void heap_delete(struct trunc_info *ti)
53 int cur = 1, child = 2;
55 heap_swap(ti, 1, ti->heapnum--);
56 while (child <= ti->heapnum) {
57 if (child < ti->heapnum &&
58 (*ti->cmp)(ti->heap[ti->ptr[child]],
59 ti->heap[ti->ptr[1+child]]) > 0)
61 if ((*ti->cmp)(ti->heap[ti->ptr[cur]],
62 ti->heap[ti->ptr[child]]) > 0)
64 heap_swap(ti, cur, child);
73 static void heap_insert(struct trunc_info *ti, const char *buf, int indx)
77 cur = ++(ti->heapnum);
78 memcpy(ti->heap[ti->ptr[cur]], buf, ti->keysize);
79 ti->indx[ti->ptr[cur]] = indx;
81 while (parent && (*ti->cmp)(ti->heap[ti->ptr[parent]],
82 ti->heap[ti->ptr[cur]]) > 0)
84 heap_swap(ti, cur, parent);
90 static struct trunc_info *heap_init(int size, int key_size,
91 int (*cmp)(const void *p1,
94 struct trunc_info *ti = (struct trunc_info *) xmalloc(sizeof(*ti));
99 ti->keysize = key_size;
101 ti->indx = (int *) xmalloc(size * sizeof(*ti->indx));
102 ti->heap = (char **) xmalloc(size * sizeof(*ti->heap));
103 ti->ptr = (int *) xmalloc(size * sizeof(*ti->ptr));
104 ti->swapbuf = (char *) xmalloc(ti->keysize);
105 ti->tmpbuf = (char *) xmalloc(ti->keysize);
106 ti->buf = (char *) xmalloc(size * ti->keysize);
107 for (i = size; --i >= 0; )
110 ti->heap[i] = ti->buf + ti->keysize * i;
115 static void heap_close(struct trunc_info *ti)
126 static RSET rset_trunc_r(ZebraHandle zi, const char *term, int length,
127 const char *flags, ISAM_P *isam_p, int from, int to,
128 int merge_chunk, int preserve_position,
129 int term_type, NMEM rset_nmem,
130 const struct key_control *kctrl, int scope,
137 result = rstemp_create(rset_nmem, kctrl, scope,
138 res_get(zi->res, "setTmpDir"), termid);
139 result_rsfd = rset_open(result, RSETF_WRITE);
141 if (to - from > merge_chunk)
145 int i, i_add = (to-from)/merge_chunk + 1;
146 struct trunc_info *ti;
148 int rsmax = (to-from)/i_add + 1;
150 rset = (RSET *) xmalloc(sizeof(*rset) * rsmax);
151 rsfd = (RSFD *) xmalloc(sizeof(*rsfd) * rsmax);
153 for (i = from; i < to; i += i_add)
156 rset[rscur] = rset_trunc_r(zi, term, length, flags,
158 merge_chunk, preserve_position,
159 term_type, rset_nmem,
162 rset[rscur] = rset_trunc_r(zi, term, length, flags,
164 merge_chunk, preserve_position,
165 term_type, rset_nmem,
169 ti = heap_init (rscur, sizeof(struct it_key), key_compare_it);
170 for (i = rscur; --i >= 0; )
172 rsfd[i] = rset_open(rset[i], RSETF_READ);
173 if (rset_read(rsfd[i], ti->tmpbuf, 0))
174 heap_insert(ti, ti->tmpbuf, i);
178 rset_delete(rset[i]);
183 int n = ti->indx[ti->ptr[1]];
185 rset_write(result_rsfd, ti->heap[ti->ptr[1]]);
190 if(!rset_read (rsfd[n], ti->tmpbuf,0))
194 rset_delete(rset[n]);
197 if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
200 heap_insert(ti, ti->tmpbuf, n);
209 else if (zi->reg->isamc)
213 struct trunc_info *ti;
215 ispt = (ISAMC_PP *) xmalloc(sizeof(*ispt) * (to-from));
217 ti = heap_init(to-from, sizeof(struct it_key),
219 for (i = to-from; --i >= 0; )
221 ispt[i] = isamc_pp_open(zi->reg->isamc, isam_p[from+i]);
222 if (isamc_pp_read(ispt[i], ti->tmpbuf))
223 heap_insert(ti, ti->tmpbuf, i);
225 isamc_pp_close(ispt[i]);
229 int n = ti->indx[ti->ptr[1]];
231 rset_write(result_rsfd, ti->heap[ti->ptr[1]]);
233 if (preserve_position)
236 if (isamc_pp_read(ispt[n], ti->tmpbuf))
237 heap_insert(ti, ti->tmpbuf, n);
239 isamc_pp_close(ispt[n]);
245 if (!isamc_pp_read(ispt[n], ti->tmpbuf))
248 isamc_pp_close(ispt[n]);
251 if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
254 heap_insert(ti, ti->tmpbuf, n);
263 else if (zi->reg->isams)
267 struct trunc_info *ti;
270 ispt = (ISAMS_PP *) xmalloc(sizeof(*ispt) * (to-from));
272 ti = heap_init(to-from, sizeof(struct it_key),
274 for (i = to-from; --i >= 0; )
276 ispt[i] = isams_pp_open(zi->reg->isams, isam_p[from+i]);
277 if (isams_pp_read(ispt[i], ti->tmpbuf))
278 heap_insert(ti, ti->tmpbuf, i);
280 isams_pp_close(ispt[i]);
284 int n = ti->indx[ti->ptr[1]];
286 rset_write(result_rsfd, ti->heap[ti->ptr[1]]);
290 if (!isams_pp_read(ispt[n], ti->tmpbuf))
293 isams_pp_close(ispt[n]);
296 if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
299 heap_insert(ti, ti->tmpbuf, n);
307 else if (zi->reg->isamb)
311 struct trunc_info *ti;
313 ispt = (ISAMB_PP *) xmalloc(sizeof(*ispt) * (to-from));
315 ti = heap_init(to-from, sizeof(struct it_key),
317 for (i = to-from; --i >= 0; )
319 if (isam_p[from+i]) {
320 ispt[i] = isamb_pp_open(zi->reg->isamb, isam_p[from+i], scope);
321 if (isamb_pp_read(ispt[i], ti->tmpbuf))
322 heap_insert(ti, ti->tmpbuf, i);
324 isamb_pp_close(ispt[i]);
329 int n = ti->indx[ti->ptr[1]];
331 rset_write(result_rsfd, ti->heap[ti->ptr[1]]);
334 if (preserve_position)
337 if (isamb_pp_read(ispt[n], ti->tmpbuf))
338 heap_insert(ti, ti->tmpbuf, n);
340 isamb_pp_close(ispt[n]);
346 if (!isamb_pp_read(ispt[n], ti->tmpbuf))
349 isamb_pp_close(ispt[n]);
352 if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
355 heap_insert(ti, ti->tmpbuf, n);
365 yaz_log(YLOG_WARN, "Unknown isam set in rset_trunc_r");
367 rset_close(result_rsfd);
371 static int isams_trunc_cmp(const void *p1, const void *p2)
373 ISAM_P i1 = *(ISAM_P*) p1;
374 ISAM_P i2 = *(ISAM_P*) p2;
383 static int isamc_trunc_cmp(const void *p1, const void *p2)
385 ISAM_P i1 = *(ISAM_P*) p1;
386 ISAM_P i2 = *(ISAM_P*) p2;
389 d = (isamc_type(i1) - isamc_type(i2));
391 d = isamc_block(i1) - isamc_block(i2);
399 RSET rset_trunc(ZebraHandle zi, ISAM_P *isam_p, int no,
400 const char *term, int length, const char *flags,
401 int preserve_position, int term_type, NMEM rset_nmem,
402 const struct key_control *kctrl, int scope)
408 return rsnull_create(rset_nmem,kctrl);
409 termid = rset_term_create(term, length, flags, term_type,rset_nmem);
413 return rsisams_create(rset_nmem, kctrl, scope,
414 zi->reg->isams, *isam_p, termid);
415 qsort(isam_p, no, sizeof(*isam_p), isams_trunc_cmp);
417 else if (zi->reg->isamc)
420 return rsisamc_create(rset_nmem, kctrl, scope,
421 zi->reg->isamc, *isam_p, termid);
422 qsort(isam_p, no, sizeof(*isam_p), isamc_trunc_cmp);
424 else if (zi->reg->isamb)
426 int trunc_limit = atoi(res_get_def(zi->res, "trunclimit", "10000"));
428 return rsisamb_create(rset_nmem,kctrl, scope,
429 zi->reg->isamb, *isam_p, termid);
430 else if (no < trunc_limit)
433 RSET *rsets = xmalloc(no*sizeof(RSET)); /* use nmem! */
435 for (i = 0; i<no; i++)
436 rsets[i] = rsisamb_create(rset_nmem, kctrl, scope,
437 zi->reg->isamb, isam_p[i], termid);
438 r = rsmulti_or_create( rset_nmem, kctrl, scope, no, rsets);
442 qsort(isam_p, no, sizeof(*isam_p), isamc_trunc_cmp);
446 yaz_log(YLOG_WARN, "Unknown isam set in rset_trunc");
447 return rsnull_create(rset_nmem, kctrl);
449 trunc_chunk = atoi(res_get_def(zi->res, "truncchunk", "100"));
450 result = rset_trunc_r(zi, term, length, flags, isam_p, 0, no, trunc_chunk,
451 preserve_position, term_type, rset_nmem, kctrl, scope,