X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Ftrunc.c;h=909999ac8f5d4cf0290d5cc2de8f898b051031b2;hp=ba3f18f50bc6631ddc3f1e557b6eeef3f0e5cce8;hb=6c9fcd3b5d3108702fa1ffc92dab4ab6060f9a19;hpb=47ea1fc957c7b97bb30a26698f072109cae275e4 diff --git a/index/trunc.c b/index/trunc.c index ba3f18f..909999a 100644 --- a/index/trunc.c +++ b/index/trunc.c @@ -1,27 +1,31 @@ -/* - * Copyright (C) 1994-2002, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss, Heikki Levanto - * - * $Id: trunc.c,v 1.26 2002-07-25 13:06:43 adam Exp $ - * - */ +/* $Id: trunc.c,v 1.50 2005-01-15 19:38:28 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 #include -#define NEW_TRUNC 1 - #include "index.h" -#include -#include -#include -#include -#include -#include -#include -#if NEW_TRUNC -#include -#endif +#include struct trunc_info { int *ptr; @@ -122,24 +126,29 @@ static void heap_close (struct trunc_info *ti) 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) + 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"); - parms.rset_term = rset_term_create (term, length, flags, term_type); result = rset_create (rset_kind_temp, &parms); + */ + 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) { RSFD *rsfd; RSET *rset; - int term_index; int i, i_add = (to-from)/merge_chunk + 1; struct trunc_info *ti; int rscur = 0; @@ -154,23 +163,25 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, rset[rscur] = rset_trunc_r (zi, term, length, flags, isam_p, i, i+i_add, merge_chunk, preserve_position, - term_type); + term_type, rset_nmem, + kctrl, scope,termid); else rset[rscur] = rset_trunc_r (zi, term, length, flags, isam_p, i, to, merge_chunk, preserve_position, - term_type); + term_type, rset_nmem, + kctrl, scope,termid); rscur++; } ti = heap_init (rscur, sizeof(struct it_key), key_compare_it); for (i = rscur; --i >= 0; ) { rsfd[i] = rset_open (rset[i], RSETF_READ); - if (rset_read (rset[i], rsfd[i], ti->tmpbuf, &term_index)) + 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]); } } @@ -178,14 +189,15 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, { 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, &term_index)) + 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; } @@ -201,61 +213,6 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, xfree (rsfd); heap_close (ti); } - else if (zi->reg->isam) - { - ISPT *ispt; - int i; - struct trunc_info *ti; - - ispt = (ISPT *) xmalloc (sizeof(*ispt) * (to-from)); - - ti = heap_init (to-from, sizeof(struct it_key), - key_compare_it); - for (i = to-from; --i >= 0; ) - { - ispt[i] = is_position (zi->reg->isam, isam_p[from+i]); - if (is_readkey (ispt[i], ti->tmpbuf)) - heap_insert (ti, ti->tmpbuf, i); - else - is_pt_free (ispt[i]); - } - while (ti->heapnum) - { - int n = ti->indx[ti->ptr[1]]; - - rset_write (result, result_rsfd, ti->heap[ti->ptr[1]]); - if (preserve_position) - { -/* section that preserve all keys */ - heap_delete (ti); - if (is_readkey (ispt[n], ti->tmpbuf)) - heap_insert (ti, ti->tmpbuf, 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) - { - heap_delete (ti); - heap_insert (ti, ti->tmpbuf, n); - break; - } - } - } - } - heap_close (ti); - xfree (ispt); - } else if (zi->reg->isamc) { ISAMC_PP *ispt; @@ -278,7 +235,8 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, { 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++; if (preserve_position) { heap_delete (ti); @@ -309,66 +267,12 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, heap_close (ti); xfree (ispt); } - - else if (zi->reg->isamd) - { - ISAMD_PP *ispt; - int i; - struct trunc_info *ti; - - ispt = (ISAMD_PP *) xmalloc (sizeof(*ispt) * (to-from)); - - ti = heap_init (to-from, sizeof(struct it_key), - key_compare_it); - for (i = to-from; --i >= 0; ) - { - logf(LOG_FATAL, "isam_d does not (currently) support truncs"); - abort(); - /*ispt[i] = isamd_pp_open (zi->reg->isamd, isam_p[from+i]); */ - if (isamd_pp_read (ispt[i], ti->tmpbuf)) - heap_insert (ti, ti->tmpbuf, i); - else - isamd_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 (isamd_pp_read (ispt[n], ti->tmpbuf)) - heap_insert (ti, ti->tmpbuf, n); - else - isamd_pp_close (ispt[n]); -#else -/* section that preserve all keys with unique sysnos */ - while (1) - { - if (!isamd_pp_read (ispt[n], ti->tmpbuf)) - { - heap_delete (ti); - isamd_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); - } else if (zi->reg->isams) { ISAMS_PP *ispt; int i; struct trunc_info *ti; + int nn = 0; ispt = (ISAMS_PP *) xmalloc (sizeof(*ispt) * (to-from)); @@ -386,7 +290,8 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, { 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 (!isams_pp_read (ispt[n], ti->tmpbuf)) @@ -418,17 +323,20 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, key_compare_it); for (i = to-from; --i >= 0; ) { - ispt[i] = isamb_pp_open (zi->reg->isamb, isam_p[from+i]); - if (isamb_pp_read (ispt[i], ti->tmpbuf)) - heap_insert (ti, ti->tmpbuf, i); - else - isamb_pp_close (ispt[i]); + 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, result_rsfd, ti->heap[ti->ptr[1]]); + rset_write (result_rsfd, ti->heap[ti->ptr[1]]); + nn++; if (preserve_position) { @@ -461,9 +369,9 @@ static RSET rset_trunc_r (ZebraHandle zi, const char *term, int length, xfree (ispt); } else - logf (LOG_WARN, "Unknown isam set in rset_trunc_r"); + yaz_log (YLOG_WARN, "Unknown isam set in rset_trunc_r"); - rset_close (result, result_rsfd); + rset_close (result_rsfd); return result; } @@ -472,169 +380,79 @@ static int isams_trunc_cmp (const void *p1, const void *p2) ISAMS_P i1 = *(ISAMS_P*) p1; ISAMS_P i2 = *(ISAMS_P*) p2; - return i1 - i2; -} - -static int isam_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); + 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; - - d = isc_type (i1) - isc_type (i2); - if (d) - return d; - return isc_block (i1) - isc_block (i2); -} -static int isamd_trunc_cmp (const void *p1, const void *p2) -{ - ISAMD_P i1 = *(ISAMD_P*) p1; - ISAMD_P i2 = *(ISAMD_P*) p2; - int d; - - d = isamd_type (i1) - isamd_type (i2); - if (d) - return d; - return isamd_block (i1) - isamd_block (i2); + zint d; + + 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 (ZebraHandle zi, ISAMS_P *isam_p, int no, const char *term, int length, const char *flags, - int preserve_position, int term_type) + int preserve_position, int term_type, NMEM rset_nmem, + const struct key_control *kctrl, int scope) { - logf (LOG_DEBUG, "rset_trunc no=%d", no); + TERMID termid; + yaz_log (YLOG_DEBUG, "rset_trunc no=%d", no); if (no < 1) - { - rset_null_parms parms; - parms.rset_term = rset_term_create (term, length, flags, term_type); - return rset_create (rset_kind_null, &parms); - } + return rsnull_create (rset_nmem,kctrl); + termid=rset_term_create(term, length, flags, term_type,rset_nmem); if (zi->reg->isams) { if (no == 1) - { - rset_isams_parms parms; - - parms.pos = *isam_p; - parms.is = zi->reg->isams; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_isams, &parms); - } + 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->isam) - { - if (no == 1) - { - rset_isam_parms parms; - - parms.pos = *isam_p; - parms.is = zi->reg->isam; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_isam, &parms); - } - qsort (isam_p, no, sizeof(*isam_p), isam_trunc_cmp); - } else if (zi->reg->isamc) { if (no == 1) - { - rset_isamc_parms parms; - - parms.key_size = sizeof(struct it_key); - parms.cmp = key_compare_it; - parms.pos = *isam_p; - parms.is = zi->reg->isamc; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_isamc, &parms); - } -#if NEW_TRUNC - else if (no < 10000) - { - rset_m_or_parms parms; - - parms.key_size = sizeof(struct it_key); - parms.cmp = key_compare_it; - parms.isc = zi->reg->isamc; - parms.isam_positions = isam_p; - parms.no_isam_positions = no; - parms.no_save_positions = 100000; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_m_or, &parms); - } -#endif + 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->reg->isamd) - { - if (no == 1) - { - rset_isamd_parms parms; - - logf(LOG_FATAL, "isam_d does not (currently) support truncs"); - abort(); - /* parms.pos = *isam_p; */ - parms.is = zi->reg->isamd; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_isamd, &parms); - } -#if NEW_TRUNC_NOT_DONE_FOR_ISAM_D - else if (no < 10000) - { - rset_m_or_parms parms; - - parms.key_size = sizeof(struct it_key); - parms.cmp = key_compare_it; - parms.isc = 0; - parms.isamd=zi->reg->isamd; - parms.isam_positions = isam_p; - parms.no_isam_positions = no; - parms.no_save_positions = 100000; - parms.rset_term = rset_term_create (term, length, flags); - return rset_create (rset_kind_m_or, &parms); - } -#endif - qsort (isam_p, no, sizeof(*isam_p), isamd_trunc_cmp); - } else if (zi->reg->isamb) { if (no == 1) + return rsisamb_create(rset_nmem,kctrl, scope, + zi->reg->isamb, *isam_p, termid); + else if (no <10000 ) /* FIXME - hardcoded number */ { - rset_isamb_parms parms; - - parms.key_size = sizeof(struct it_key); - parms.cmp = key_compare_it; - parms.pos = *isam_p; - parms.is = zi->reg->isamb; - parms.rset_term = rset_term_create (term, length, flags, - term_type); - return rset_create (rset_kind_isamb, &parms); - } - qsort (isam_p, no, sizeof(*isam_p), isamd_trunc_cmp); + RSET r; + RSET *rsets=xmalloc(no*sizeof(RSET)); /* use nmem! */ + int i; + for (i=0;ireg->isamb, isam_p[i], termid); + r=rsmultior_create( rset_nmem, kctrl, scope, no, rsets); + xfree(rsets); + return r; + } + qsort (isam_p, no, sizeof(*isam_p), isamc_trunc_cmp); } else { - logf (LOG_WARN, "Unknown isam set in rset_trunc"); - return rset_create (rset_kind_null, NULL); + 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); + preserve_position, term_type, rset_nmem,kctrl,scope, + termid); }