-/*
- * Copyright (C) 1994-1996, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: trunc.c,v $
- * Revision 1.2 1996-11-08 11:10:28 adam
- * Buffers used during file match got bigger.
- * Compressed ISAM support everywhere.
- * Bug fixes regarding masking characters in queries.
- * Redesigned Regexp-2 queries.
- *
- * Revision 1.1 1996/11/04 14:07:40 adam
- * Moved truncation code to trunc.c.
- *
- */
+/* $Id: trunc.c,v 1.51 2005-01-15 20:47:15 adam Exp $
+ Copyright (C) 1995-2005
+ Index Data ApS
+
+This file is part of the Zebra server.
+
+Zebra is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Zebra; see the file LICENSE.zebra. If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
+*/
+
+
#include <stdio.h>
#include <assert.h>
-#include "zserver.h"
-#include <rstemp.h>
-#include <rsisam.h>
-#include <rsisamc.h>
-#include <rsnull.h>
+#include "index.h"
+#include <rset.h>
struct trunc_info {
int *ptr;
}
}
-static
-struct trunc_info *heap_init (int size, int key_size,
- int (*cmp)(const void *p1, const void *p2))
+static struct trunc_info *heap_init (int size, int key_size,
+ int (*cmp)(const void *p1,
+ const void *p2))
{
- struct trunc_info *ti = xmalloc (sizeof(*ti));
+ struct trunc_info *ti = (struct trunc_info *) xmalloc(sizeof(*ti));
int i;
++size;
ti->heapnum = 0;
ti->keysize = key_size;
ti->cmp = cmp;
- ti->indx = xmalloc (size * sizeof(*ti->indx));
- ti->heap = xmalloc (size * sizeof(*ti->heap));
- ti->ptr = xmalloc (size * sizeof(*ti->ptr));
- ti->swapbuf = xmalloc (ti->keysize);
- ti->tmpbuf = xmalloc (ti->keysize);
- ti->buf = xmalloc (size * ti->keysize);
+ ti->indx = (int *) xmalloc(size * sizeof(*ti->indx));
+ ti->heap = (char **) xmalloc(size * sizeof(*ti->heap));
+ ti->ptr = (int *) xmalloc(size * sizeof(*ti->ptr));
+ ti->swapbuf = (char *) xmalloc(ti->keysize);
+ ti->tmpbuf = (char *) xmalloc(ti->keysize);
+ ti->buf = (char *) xmalloc(size * ti->keysize);
for (i = size; --i >= 0; )
{
ti->ptr[i] = i;
static void heap_close (struct trunc_info *ti)
{
- xfree (ti->ptr);
- xfree (ti->indx);
- xfree (ti->heap);
- xfree (ti->swapbuf);
- xfree (ti->tmpbuf);
- xfree (ti);
+ xfree(ti->ptr);
+ xfree(ti->indx);
+ xfree(ti->heap);
+ xfree(ti->swapbuf);
+ xfree(ti->tmpbuf);
+ xfree(ti->buf);
+ xfree(ti);
}
-static RSET rset_trunc_r (ZServerInfo *zi, ISAM_P *isam_p, int from, int to,
- int merge_chunk)
+static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length,
+ const char *flags, ISAMS_P *isam_p, int from, int to,
+ int merge_chunk, int preserve_position,
+ int term_type, NMEM rset_nmem,
+ const struct key_control *kctrl, int scope,
+ TERMID termid)
{
RSET result;
RSFD result_rsfd;
- rset_temp_parms parms;
+ int nn = 0;
+ /*
+ rset_temp_parms parms;
+ parms.cmp = key_compare_it;
parms.key_size = sizeof(struct it_key);
+ parms.temp_path = res_get (zi->res, "setTmpDir");
result = rset_create (rset_kind_temp, &parms);
- result_rsfd = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO);
+ */
+ result = rstemp_create( rset_nmem,kctrl, scope,
+ res_get (zi->res, "setTmpDir"), termid);
+ result_rsfd = rset_open (result, RSETF_WRITE);
if (to - from > merge_chunk)
{
int rscur = 0;
int rsmax = (to-from)/i_add + 1;
- rset = xmalloc (sizeof(*rset) * rsmax);
- rsfd = xmalloc (sizeof(*rsfd) * rsmax);
+ rset = (RSET *) xmalloc(sizeof(*rset) * rsmax);
+ rsfd = (RSFD *) xmalloc(sizeof(*rsfd) * rsmax);
for (i = from; i < to; i += i_add)
{
if (i_add <= to - i)
- rset[rscur] = rset_trunc_r (zi, isam_p, i, i+i_add,
- merge_chunk);
+ rset[rscur] = rset_trunc_r (zi, term, length, flags,
+ isam_p, i, i+i_add,
+ merge_chunk, preserve_position,
+ term_type, rset_nmem,
+ kctrl, scope,termid);
else
- rset[rscur] = rset_trunc_r (zi, isam_p, i, to,
- merge_chunk);
+ rset[rscur] = rset_trunc_r (zi, term, length, flags,
+ isam_p, i, to,
+ merge_chunk, preserve_position,
+ term_type, rset_nmem,
+ kctrl, scope,termid);
rscur++;
}
- ti = heap_init (rscur, sizeof(struct it_key), key_compare);
+ ti = heap_init (rscur, sizeof(struct it_key), key_compare_it);
for (i = rscur; --i >= 0; )
{
- rsfd[i] = rset_open (rset[i], RSETF_READ|RSETF_SORT_SYSNO);
- if (rset_read (rset[i], rsfd[i], ti->tmpbuf))
+ rsfd[i] = rset_open (rset[i], RSETF_READ);
+ if (rset_read(rsfd[i], ti->tmpbuf,0))
heap_insert (ti, ti->tmpbuf, i);
else
{
- rset_close (rset[i], rsfd[i]);
+ rset_close (rsfd[i]);
rset_delete (rset[i]);
}
}
{
int n = ti->indx[ti->ptr[1]];
- rset_write (result, result_rsfd, ti->heap[ti->ptr[1]]);
+ rset_write (result_rsfd, ti->heap[ti->ptr[1]]);
+ nn++;
while (1)
{
- if (!rset_read (rset[n], rsfd[n], ti->tmpbuf))
+ if (!rset_read (rsfd[n], ti->tmpbuf,0))
{
heap_delete (ti);
- rset_close (rset[n], rsfd[n]);
+ rset_close (rsfd[n]);
rset_delete (rset[n]);
break;
}
}
}
}
- xfree (rset);
- xfree (rsfd);
+ xfree(rset);
+ xfree(rsfd);
heap_close (ti);
}
- else if (zi->isam)
+ else if (zi->reg->isamc)
{
- ISPT *ispt;
+ ISAMC_PP *ispt;
int i;
struct trunc_info *ti;
- ispt = xmalloc (sizeof(*ispt) * (to-from));
+ ispt = (ISAMC_PP *) xmalloc(sizeof(*ispt) * (to-from));
ti = heap_init (to-from, sizeof(struct it_key),
- key_compare);
+ key_compare_it);
for (i = to-from; --i >= 0; )
{
- ispt[i] = is_position (zi->isam, isam_p[from+i]);
- if (is_readkey (ispt[i], ti->tmpbuf))
+ ispt[i] = isc_pp_open (zi->reg->isamc, isam_p[from+i]);
+ if (isc_pp_read (ispt[i], ti->tmpbuf))
heap_insert (ti, ti->tmpbuf, i);
else
- is_pt_free (ispt[i]);
+ isc_pp_close (ispt[i]);
}
while (ti->heapnum)
{
int n = ti->indx[ti->ptr[1]];
- rset_write (result, result_rsfd, ti->heap[ti->ptr[1]]);
-#if 1
-/* section that preserve all keys */
- heap_delete (ti);
- if (is_readkey (ispt[n], ti->tmpbuf))
- heap_insert (ti, ti->tmpbuf, n);
+ rset_write (result_rsfd, ti->heap[ti->ptr[1]]);
+ nn++;
+ if (preserve_position)
+ {
+ heap_delete (ti);
+ if (isc_pp_read (ispt[n], ti->tmpbuf))
+ heap_insert (ti, ti->tmpbuf, n);
+ else
+ isc_pp_close (ispt[n]);
+ }
else
- is_pt_free (ispt[n]);
-#else
-/* section that preserve all keys with unique sysnos */
- while (1)
{
- if (!is_readkey (ispt[n], ti->tmpbuf))
- {
- heap_delete (ti);
- is_pt_free (ispt[n]);
- break;
- }
- if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
+ while (1)
{
- heap_delete (ti);
- heap_insert (ti, ti->tmpbuf, n);
- break;
+ if (!isc_pp_read (ispt[n], ti->tmpbuf))
+ {
+ heap_delete (ti);
+ isc_pp_close (ispt[n]);
+ break;
+ }
+ if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
+ {
+ heap_delete (ti);
+ heap_insert (ti, ti->tmpbuf, n);
+ break;
+ }
}
}
-#endif
}
heap_close (ti);
- xfree (ispt);
+ xfree(ispt);
}
- else
+ else if (zi->reg->isams)
{
- ISAMC_PP *ispt;
+ ISAMS_PP *ispt;
int i;
struct trunc_info *ti;
+ int nn = 0;
- ispt = xmalloc (sizeof(*ispt) * (to-from));
+ ispt = (ISAMS_PP *) xmalloc(sizeof(*ispt) * (to-from));
ti = heap_init (to-from, sizeof(struct it_key),
- key_compare);
+ key_compare_it);
for (i = to-from; --i >= 0; )
{
- ispt[i] = isc_pp_open (zi->isamc, isam_p[from+i]);
- if (isc_pp_read (ispt[i], ti->tmpbuf))
+ ispt[i] = isams_pp_open (zi->reg->isams, isam_p[from+i]);
+ if (isams_pp_read (ispt[i], ti->tmpbuf))
heap_insert (ti, ti->tmpbuf, i);
else
- isc_pp_close (ispt[i]);
+ isams_pp_close (ispt[i]);
}
while (ti->heapnum)
{
int n = ti->indx[ti->ptr[1]];
- rset_write (result, result_rsfd, ti->heap[ti->ptr[1]]);
-#if 0
-/* section that preserve all keys */
- heap_delete (ti);
- if (is_readkey (ispt[n], ti->tmpbuf))
- heap_insert (ti, ti->tmpbuf, n);
- else
- isc_pp_close (ispt[n]);
-#else
-/* section that preserve all keys with unique sysnos */
+ rset_write (result_rsfd, ti->heap[ti->ptr[1]]);
+ nn++;
while (1)
{
- if (!isc_pp_read (ispt[n], ti->tmpbuf))
+ if (!isams_pp_read (ispt[n], ti->tmpbuf))
{
heap_delete (ti);
- isc_pp_close (ispt[n]);
+ isams_pp_close (ispt[n]);
break;
}
if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
break;
}
}
-#endif
}
heap_close (ti);
- xfree (ispt);
+ xfree(ispt);
+ }
+ else if (zi->reg->isamb)
+ {
+ ISAMB_PP *ispt;
+ int i;
+ struct trunc_info *ti;
+
+ ispt = (ISAMB_PP *) xmalloc(sizeof(*ispt) * (to-from));
+
+ ti = heap_init (to-from, sizeof(struct it_key),
+ key_compare_it);
+ for (i = to-from; --i >= 0; )
+ {
+ if (isam_p[from+i]) {
+ ispt[i] = isamb_pp_open (zi->reg->isamb, isam_p[from+i], scope);
+ if (isamb_pp_read (ispt[i], ti->tmpbuf))
+ heap_insert (ti, ti->tmpbuf, i);
+ else
+ isamb_pp_close (ispt[i]);
+ }
+ }
+ while (ti->heapnum)
+ {
+ int n = ti->indx[ti->ptr[1]];
+
+ rset_write (result_rsfd, ti->heap[ti->ptr[1]]);
+ nn++;
+
+ if (preserve_position)
+ {
+ heap_delete (ti);
+ if (isamb_pp_read (ispt[n], ti->tmpbuf))
+ heap_insert (ti, ti->tmpbuf, n);
+ else
+ isamb_pp_close (ispt[n]);
+ }
+ else
+ {
+ while (1)
+ {
+ if (!isamb_pp_read (ispt[n], ti->tmpbuf))
+ {
+ heap_delete (ti);
+ isamb_pp_close (ispt[n]);
+ break;
+ }
+ if ((*ti->cmp)(ti->tmpbuf, ti->heap[ti->ptr[1]]) > 1)
+ {
+ heap_delete (ti);
+ heap_insert (ti, ti->tmpbuf, n);
+ break;
+ }
+ }
+ }
+ }
+ heap_close (ti);
+ xfree(ispt);
}
- rset_close (result, result_rsfd);
+ else
+ yaz_log (YLOG_WARN, "Unknown isam set in rset_trunc_r");
+
+ rset_close (result_rsfd);
return result;
}
-static int isam_trunc_cmp (const void *p1, const void *p2)
+static int isams_trunc_cmp (const void *p1, const void *p2)
{
- ISAM_P i1 = *(ISAM_P*) p1;
- ISAM_P i2 = *(ISAM_P*) p2;
- int d;
-
- d = is_type (i1) - is_type (i2);
- if (d)
- return d;
- return is_block (i1) - is_block (i2);
+ ISAMS_P i1 = *(ISAMS_P*) p1;
+ ISAMS_P i2 = *(ISAMS_P*) p2;
+
+ if (i1 > i2)
+ return 1;
+ else if (i1 < i2)
+ return -1;
+ return 0;
}
static int isamc_trunc_cmp (const void *p1, const void *p2)
{
ISAMC_P i1 = *(ISAMC_P*) p1;
ISAMC_P i2 = *(ISAMC_P*) p2;
- int d;
+ zint d;
- d = isc_type (i1) - isc_type (i2);
- if (d)
- return d;
- return isc_block (i1) - isc_block (i2);
+ d = (isc_type (i1) - isc_type (i2));
+ if (d == 0)
+ d = isc_block (i1) - isc_block (i2);
+ if (d > 0)
+ return 1;
+ else if (d < 0)
+ return -1;
+ return 0;
}
-RSET rset_trunc (ZServerInfo *zi, ISAM_P *isam_p, int no)
+RSET rset_trunc (ZebraHandle zi, ISAMS_P *isam_p, int no,
+ const char *term, int length, const char *flags,
+ int preserve_position, int term_type, NMEM rset_nmem,
+ const struct key_control *kctrl, int scope)
{
- if (zi->isam)
+ TERMID termid;
+ yaz_log (YLOG_DEBUG, "rset_trunc no=%d", no);
+ if (no < 1)
+ return rsnull_create (rset_nmem,kctrl);
+ termid = rset_term_create(term, length, flags, term_type,rset_nmem);
+ if (zi->reg->isams)
{
- if (no < 1)
- return rset_create (rset_kind_null, NULL);
- else if (no == 1)
- {
- rset_isam_parms parms;
-
- parms.pos = *isam_p;
- parms.is = zi->isam;
- return rset_create (rset_kind_isam, &parms);
- }
- qsort (isam_p, no, sizeof(*isam_p), isam_trunc_cmp);
+ if (no == 1)
+ return rsisams_create(rset_nmem, kctrl, scope,
+ zi->reg->isams, *isam_p, termid);
+ qsort (isam_p, no, sizeof(*isam_p), isams_trunc_cmp);
+ }
+ else if (zi->reg->isamc)
+ {
+ if (no == 1)
+ return rsisamc_create(rset_nmem, kctrl, scope,
+ zi->reg->isamc, *isam_p, termid);
+ qsort (isam_p, no, sizeof(*isam_p), isamc_trunc_cmp);
}
- else if (zi->isamc)
+ else if (zi->reg->isamb)
{
- if (no < 1)
- return rset_create (rset_kind_null, NULL);
- else if (no == 1)
+ if (no == 1)
+ return rsisamb_create(rset_nmem,kctrl, scope,
+ zi->reg->isamb, *isam_p, termid);
+ else if (no <10000 ) /* FIXME - hardcoded number */
{
- rset_isamc_parms parms;
-
- parms.pos = *isam_p;
- parms.is = zi->isamc;
- return rset_create (rset_kind_isamc, &parms);
- }
+ RSET r;
+ RSET *rsets = xmalloc(no*sizeof(RSET)); /* use nmem! */
+ int i;
+ for (i = 0; i<no; i++)
+ rsets[i]=rsisamb_create(rset_nmem, kctrl, scope,
+ zi->reg->isamb, isam_p[i], termid);
+ r = rsmulti_or_create( rset_nmem, kctrl, scope, no, rsets);
+ xfree(rsets);
+ return r;
+ }
qsort (isam_p, no, sizeof(*isam_p), isamc_trunc_cmp);
}
else
- logf (LOG_FATAL, "Neither isam nor isamc set in rset_trunc");
- return rset_trunc_r (zi, isam_p, 0, no, 100);
+ {
+ yaz_log (YLOG_WARN, "Unknown isam set in rset_trunc");
+ return rsnull_create (rset_nmem, kctrl);
+ }
+ return rset_trunc_r (zi, term, length, flags, isam_p, 0, no, 100,
+ preserve_position, term_type, rset_nmem,kctrl,scope,
+ termid);
}