X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Finvstat.c;h=f4b548c59f451b65a66548313c1c99db28ee7e09;hb=3e72747f4b28856e5f684dbe8984f8fb28680e62;hp=a43cfa3885b7891277901d9e67db196c696558e0;hpb=1a6ff0680d514e270257e7e599d34b6f8c9317a1;p=idzebra-moved-to-github.git diff --git a/index/invstat.c b/index/invstat.c index a43cfa3..f4b548c 100644 --- a/index/invstat.c +++ b/index/invstat.c @@ -10,17 +10,11 @@ #include #include "index.h" -#include "recindex.h" -#include "../isamc/isamh-p.h" #include "../isamc/isamd-p.h" struct inv_stat_info { - ISAM isam; - ISAMC isamc; - ISAMS isams; - ISAMH isamh; - ISAMD isamd; - int no_isam_entries[8]; + ZebraHandle zh; + int no_isam_entries[9]; int no_dict_entries; int no_dict_bytes; int isam_bounds[20]; @@ -28,112 +22,105 @@ struct inv_stat_info { char tmp[128]; }; +#define SINGLETON_TYPE 8 /* the type to use for singletons that */ + /* have no block and no block type */ + static int inv_stat_handle (char *name, const char *info, int pos, void *client) { int occur = 0; int i = 0; struct inv_stat_info *stat_info = (struct inv_stat_info*) client; - ISAM_P isam_p; + ISAMS_P isam_p; stat_info->no_dict_entries++; stat_info->no_dict_bytes += strlen(name); - assert (*info == sizeof(ISAM_P)); - memcpy (&isam_p, info+1, sizeof(ISAM_P)); + assert (*info == sizeof(ISAMS_P)); + memcpy (&isam_p, info+1, sizeof(ISAMS_P)); - //printf ("---\n"); - if (stat_info->isam) + if (stat_info->zh->reg->isams) + { + ISAMS_PP pp; + int occurx = 0; + struct it_key key; + + pp = isams_pp_open (stat_info->zh->reg->isams, isam_p); + occur = isams_pp_num (pp); + while (isams_pp_read(pp, &key)) + { + //printf ("sysno=%d seqno=%d\n", key.sysno, key.seqno); + occurx++; + } + assert (occurx == occur); + stat_info->no_isam_entries[0] += occur; + isams_pp_close (pp); + } + if (stat_info->zh->reg->isam) { ISPT ispt; - ispt = is_position (stat_info->isam, isam_p); + ispt = is_position (stat_info->zh->reg->isam, isam_p); occur = is_numkeys (ispt); is_pt_free (ispt); } - if (stat_info->isamc) + if (stat_info->zh->reg->isamc) { ISAMC_PP pp; int occurx = 0; struct it_key key; - pp = isc_pp_open (stat_info->isamc, isam_p); + pp = isc_pp_open (stat_info->zh->reg->isamc, isam_p); occur = isc_pp_num (pp); while (isc_pp_read(pp, &key)) { - printf ("sysno=%d seqno=%d\n", key.sysno, key.seqno); + //printf ("sysno=%d seqno=%d\n", key.sysno, key.seqno); occurx++; } assert (occurx == occur); stat_info->no_isam_entries[isc_type(isam_p)] += occur; isc_pp_close (pp); } - if (stat_info->isamh) - { - ISAMH_PP pp; - int occurx = 0; - struct it_key key; - - pp = isamh_pp_open (stat_info->isamh, isam_p); - - occur = isamh_pp_num (pp); - // printf (" opening item %d=%d:%d \n", - // isam_p, isamh_type(isam_p),isamh_block(isam_p)); - while (isamh_pp_read(pp, &key)) - { - occurx++; - logf (LOG_LOG,"sysno=%d seqno=%d (%x/%x) oc=%d/%d ofs=%d ", - key.sysno, key.seqno, - key.sysno, key.seqno, - occur,occurx, pp->offset); - } - if (occurx != occur) - logf(LOG_LOG,"Count error!!! read %d, counted %d", occur, occurx); - assert (occurx == occur); - stat_info->no_isam_entries[isamh_type(isam_p)] += occur; - isamh_pp_close (pp); - } - if (stat_info->isamd) + if (stat_info->zh->reg->isamd) { ISAMD_PP pp; int occurx = 0; struct it_key key; - pp = isamd_pp_open (stat_info->isamd, isam_p); + pp = isamd_pp_open (stat_info->zh->reg->isamd, isam_p); occur = isamd_pp_num (pp); - // printf (" opening item %d=%d:%d \n", - // isam_p, isamh_type(isam_p),isamh_block(isam_p)); while (isamd_pp_read(pp, &key)) { occurx++; - logf (LOG_LOG,"sysno=%d seqno=%d (%x/%x) oc=%d/%d ofs=%d ", + if ( pp->is->method->debug >8 ) + logf (LOG_LOG,"sysno=%d seqno=%d (%x/%x) oc=%d/%d ofs=%d ", key.sysno, key.seqno, key.sysno, key.seqno, occur,occurx, pp->offset); } + if ( pp->is->method->debug >7 ) + logf(LOG_LOG,"item %d=%d:%d says %d keys, counted %d", + isam_p, isamd_type(isam_p), isamd_block(isam_p), + occur, occurx); if (occurx != occur) logf(LOG_LOG,"Count error!!! read %d, counted %d", occur, occurx); assert (occurx == occur); - stat_info->no_isam_entries[isamd_type(isam_p)] += occur; + if ( is_singleton(isam_p) ) + stat_info->no_isam_entries[SINGLETON_TYPE] += occur; + else + stat_info->no_isam_entries[isamd_type(isam_p)] += occur; isamd_pp_close (pp); } - if (stat_info->isams) + if (stat_info->zh->reg->isamb) { - ISAMS_PP pp; - int occurx = 0; - struct it_key key; + ISAMB_PP pp; + struct it_key key; - pp = isams_pp_open (stat_info->isams, isam_p); - occur = isams_pp_num (pp); - while (isams_pp_read(pp, &key)) - { - //printf ("sysno=%d seqno=%d\n", key.sysno, key.seqno); - occurx++; - } - assert (occurx == occur); - stat_info->no_isam_entries[isc_type(isam_p)] += occur; - isams_pp_close (pp); + pp = isamb_pp_open(stat_info->zh->reg->isamb, isam_p); + while (isamb_pp_read(pp, &key)) + occur++; + isamb_pp_close (pp); } while (occur > stat_info->isam_bounds[i] && stat_info->isam_bounds[i]) @@ -142,93 +129,29 @@ static int inv_stat_handle (char *name, const char *info, int pos, return 0; } -void inv_prstat (BFiles bfs) +void zebra_register_statistics (ZebraHandle zh) { - Dict dict; - ISAM isam = NULL; - ISAMC isamc = NULL; - ISAMS isams = NULL; - ISAMH isamh = NULL; - ISAMD isamd = NULL; - Records records; + int blocks; + int size; + int count; int i, prev; int before = 0; int after = 1000000000; struct inv_stat_info stat_info; char term_dict[2*IT_MAX_WORD+2]; + if (zebra_begin_read (zh)) + return; + + stat_info.zh = zh; + term_dict[0] = 1; term_dict[1] = 0; - dict = dict_open (bfs, FNAME_DICT, 100, 0, 0); - if (!dict) - { - logf (LOG_FATAL, "dict_open fail"); - exit (1); - } - if (res_get_match (common_resource, "isam", "i", NULL)) - { - isam = is_open (bfs, FNAME_ISAM, key_compare, 0, - sizeof(struct it_key), common_resource); - if (!isam) - { - logf (LOG_FATAL, "is_open fail"); - exit (1); - } - } - else if (res_get_match (common_resource, "isam", "s", NULL)) - { - struct ISAMS_M_s isams_m; - isams = isams_open (bfs, FNAME_ISAMS, 0, - key_isams_m(common_resource, &isams_m)); - if (!isams) - { - logf (LOG_FATAL, "isams_open fail"); - exit (1); - } - } - else if (res_get_match (common_resource, "isam", "h", NULL)) - { - isamh = isamh_open (bfs, FNAME_ISAMH, 0, key_isamh_m(common_resource)); - if (!isamh) - { - logf (LOG_FATAL, "isamh_open fail"); - exit (1); - } - } - else if (res_get_match (common_resource, "isam", "d", NULL)) - { - struct ISAMD_M_s isamd_m; - isamd = isamd_open (bfs, FNAME_ISAMD, 0, - key_isamd_m(common_resource,&isamd_m)); - if (!isamd) - { - logf (LOG_FATAL, "isamd_open fail"); - exit (1); - } - } - else - { - struct ISAMC_M_s isamc_m; - isamc = isc_open (bfs, FNAME_ISAMC, 0, - key_isamc_m (common_resource, &isamc_m)); - if (!isamc) - { - logf (LOG_FATAL, "isc_open fail"); - exit (1); - } - } - records = rec_open (bfs, 0, 0); - - for (i = 0; i<8; i++) + for (i = 0; i<=SINGLETON_TYPE; i++) stat_info.no_isam_entries[i] = 0; stat_info.no_dict_entries = 0; stat_info.no_dict_bytes = 0; - stat_info.isam = isam; - stat_info.isamc = isamc; - stat_info.isams = isams; - stat_info.isamh = isamh; - stat_info.isamd = isamd; stat_info.isam_bounds[0] = 1; stat_info.isam_bounds[1] = 2; stat_info.isam_bounds[2] = 3; @@ -252,28 +175,60 @@ void inv_prstat (BFiles bfs) for (i = 0; i<20; i++) stat_info.isam_occurrences[i] = 0; - dict_scan (dict, term_dict, &before, &after, &stat_info, inv_stat_handle); + dict_scan (zh->reg->dict, term_dict, &before, &after, &stat_info, + inv_stat_handle); - if (isamc) + if (zh->reg->isamc) { fprintf (stderr, " Blocks Occur Size KB Bytes/Entry\n"); - for (i = 0; isc_block_used (isamc, i) >= 0; i++) + for (i = 0; isc_block_used (zh->reg->isamc, i) >= 0; i++) { - fprintf (stderr, " %8d %8d", isc_block_used (isamc, i), + fprintf (stderr, " %8d %8d", isc_block_used (zh->reg->isamc, i), stat_info.no_isam_entries[i]); if (stat_info.no_isam_entries[i]) fprintf (stderr, " %8d %f", - (int) ((1023.0 + (double) isc_block_used(isamc, i) * - isc_block_size(isamc,i))/1024), - ((double) isc_block_used(isamc, i) * - isc_block_size(isamc,i))/ + (int) ((1023.0 + (double) + isc_block_used(zh->reg->isamc, i) * + isc_block_size(zh->reg->isamc,i))/1024), + ((double) isc_block_used(zh->reg->isamc, i) * + isc_block_size(zh->reg->isamc,i))/ stat_info.no_isam_entries[i]); fprintf (stderr, "\n"); } } - - fprintf (stderr, "\n%d words using %d bytes\n", + if (zh->reg->isamd) + { + fprintf (stderr, " Blocks Occur KB Bytes/Entry\n"); + if (zh->reg->isamd->method->debug >0) + logf(LOG_LOG," Blocks Occur KB Bytes/Entry"); + for (i = 0; i<=SINGLETON_TYPE; i++) + { + blocks= isamd_block_used(zh->reg->isamd,i); + size= isamd_block_size(zh->reg->isamd,i); + count=stat_info.no_isam_entries[i]; + if (i==SINGLETON_TYPE) + blocks=size=0; + if (stat_info.no_isam_entries[i]) + { + fprintf (stderr, "%c %7d %7d %7d %5.2f\n", + (i==SINGLETON_TYPE)?('z'):('A'+i), + blocks, + count, + (int) ((1023.0 + (double) blocks * size)/1024), + ((double) blocks * size)/count); + if (zh->reg->isamd->method->debug >0) + logf(LOG_LOG, "%c %7d %7d %7d %5.2f", + (i==SINGLETON_TYPE)?('z'):('A'+i), + blocks, + count, + (int) ((1023.0 + (double) blocks * size)/1024), + ((double) blocks * size)/count); + } /* entries */ + } /* for */ + } /* isamd */ + if ( (zh->reg->isamd) && (zh->reg->isamd->method->debug>0)) + fprintf (stderr, "\n%d words using %d bytes\n", stat_info.no_dict_entries, stat_info.no_dict_bytes); fprintf (stderr, " Occurrences Words\n"); prev = 1; @@ -286,28 +241,45 @@ void inv_prstat (BFiles bfs) } fprintf (stderr, "%7d- %7d\n", prev, stat_info.isam_occurrences[i]); - - rec_close (&records); - dict_close (dict); - - if (isam) - is_close (isam); - if (isamc) - isc_close (isamc); - if (isams) - isams_close (isams); - if (isamh) - isamh_close (isamh); - if (isamd) - isamd_close (isamd); - + xmalloc_trav("unfreed"); /*! while hunting memory leaks */ + zebra_end_read (zh); } /* * * $Log: invstat.c,v $ - * Revision 1.15 1999-08-18 08:34:53 heikki + * Revision 1.25 2002-04-26 08:44:47 adam + * Index statistics working again + * + * Revision 1.24 2002/04/05 08:46:26 adam + * Zebra with full functionality + * + * Revision 1.23 2002/04/04 14:14:13 adam + * Multiple registers (alpha early) + * + * Revision 1.22 2002/02/20 17:30:01 adam + * Work on new API. Locking system re-implemented + * + * Revision 1.21 2000/07/13 10:14:20 heikki + * Removed compiler warnings when making zebra + * + * Revision 1.20 1999/12/01 13:30:30 adam + * Updated configure for Zmbol/Zebra dependent settings. + * + * Revision 1.19 1999/11/30 13:48:03 adam + * Improved installation. Updated for inclusion of YAZ header files. + * + * Revision 1.18 1999/10/06 11:46:36 heikki + * mproved statistics on isam-d + * + * Revision 1.17 1999/08/20 08:28:37 heikki + * Log levels + * + * Revision 1.16 1999/08/18 08:38:22 heikki + * Memory leak hunting + * + * Revision 1.15 1999/08/18 08:34:53 heikki * isamd * * Revision 1.14 1999/07/14 10:59:26 adam