X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fzrpn.c;h=a51d937479aba67cc95940b5a454efaf5fcef1b9;hb=85df66537199c30a492ad54be4fbe25fa77e18c8;hp=698dae1c971b8849bbbb59414e7f06606cce758b;hpb=9fe3bf8d659adc5c0623185779272d933fc51d15;p=idzebra-moved-to-github.git diff --git a/index/zrpn.c b/index/zrpn.c index 698dae1..a51d937 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -4,7 +4,35 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.31 1995-10-17 18:02:10 adam + * Revision 1.38 1995-12-11 09:12:55 adam + * The rec_get function returns NULL if record doesn't exist - will + * happen in the server if the result set records have been deleted since + * the creation of the set (i.e. the search). + * The server saves a result temporarily if it is 'volatile', i.e. the + * set is register dependent. + * + * Revision 1.37 1995/12/06 15:05:28 adam + * More verbose in count_set. + * + * Revision 1.36 1995/12/06 12:41:27 adam + * New command 'stat' for the index program. + * Filenames can be read from stdin by specifying '-'. + * Bug fix/enhancement of the transformation from terms to regular + * expressons in the search engine. + * + * Revision 1.35 1995/11/27 09:29:00 adam + * Bug fixes regarding conversion to regular expressions. + * + * Revision 1.34 1995/11/16 17:00:56 adam + * Better logging of rpn query. + * + * Revision 1.33 1995/11/01 13:58:28 quinn + * Moving data1 to yaz/retrieval + * + * Revision 1.32 1995/10/27 14:00:11 adam + * Implemented detection of database availability. + * + * Revision 1.31 1995/10/17 18:02:10 adam * New feature: databases. Implemented as prefix to words in dictionary. * * Revision 1.30 1995/10/16 09:32:38 adam @@ -107,9 +135,10 @@ #include #include #include +#include #include "zserver.h" -#include +#include "attribute.h" #include #include @@ -118,7 +147,7 @@ #include int index_word_prefix_map (char *string, oid_value attrSet, int attrUse, - int num_bases, char **basenames) + char *basename) { attent *attp; @@ -128,47 +157,7 @@ int index_word_prefix_map (char *string, oid_value attrSet, int attrUse, return -1; logf (LOG_DEBUG, "ord=%d", attp->attset_ordinal); return index_word_prefix (string, attp->attset_ordinal, - attp->local_attribute, - num_bases, basenames); -} - -/* - * attr_print: log attributes - */ -static void attr_print (Z_AttributesPlusTerm *t) -{ - int of, i; - for (of = 0; of < t->num_attributes; of++) - { - Z_AttributeElement *element; - element = t->attributeList[of]; - - switch (element->which) - { - case Z_AttributeValue_numeric: - logf (LOG_DEBUG, "attributeType=%d value=%d", - *element->attributeType, - *element->value.numeric); - break; - case Z_AttributeValue_complex: - logf (LOG_DEBUG, "attributeType=%d complex", - *element->attributeType); - for (i = 0; ivalue.complex->num_list; i++) - { - if (element->value.complex->list[i]->which == - Z_StringOrNumeric_string) - logf (LOG_DEBUG, " string: '%s'", - element->value.complex->list[i]->u.string); - else if (element->value.complex->list[i]->which == - Z_StringOrNumeric_numeric) - logf (LOG_DEBUG, " numeric: '%d'", - *element->value.complex->list[i]->u.numeric); - } - break; - default: - assert (0); - } - } + attp->local_attribute, basename); } typedef struct { @@ -595,7 +584,8 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, const char *term_sub, char *term_dict, oid_value attributeSet, - struct grep_info *grep_info) + struct grep_info *grep_info, + int *max_pos) { AttrType relation; int relation_value; @@ -636,7 +626,7 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, return 0; } logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, + r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, max_pos, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); @@ -644,13 +634,20 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, return 1; } +static void verbatim_char (int ch, int *indx, char *dst) +{ + if (!isalnum (ch)) + dst[(*indx)++] = '\\'; + dst[(*indx)++] = ch; +} + static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, const char *term_sub, oid_value attributeSet, struct grep_info *grep_info, int num_bases, char **basenames) { char term_dict[2*IT_MAX_WORD+2]; - int i, j, r; + int i, j, r, base_no; AttrType truncation; int truncation_value; AttrType use; @@ -666,63 +663,82 @@ static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, if (use_value == -1) use_value = 1016; - i = index_word_prefix_map (term_dict, curAttributeSet, use_value, - num_bases, basenames); - if (i < 0) - { - zi->errCode = 114; - return -1; - } - if (relational_term (zi, zapt, term_sub, term_dict, - attributeSet, grep_info)) - return 0; - switch (truncation_value) + for (base_no = 0; base_no < num_bases; base_no++) { - case -1: /* not specified */ - case 100: /* do not truncate */ - strcat (term_dict, "("); - strcat (term_dict, term_sub); - strcat (term_dict, ")"); - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep fail, trunc=none: %d", r); - break; - case 1: /* right truncation */ - strcat (term_dict, term_sub); - strcat (term_dict, ".*"); - dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, grep_handle); - break; - case 2: /* left truncation */ - case 3: /* left&right truncation */ - zi->errCode = 120; - return -1; - case 101: /* process # in term */ - for (j = strlen(term_dict), i = 0; term_sub[i] && i < 2; i++) - term_dict[j++] = term_sub[i]; - for (; term_sub[i]; i++) - if (term_sub[i] == '#') + int max_pos; + int prefix_len = index_word_prefix_map (term_dict, curAttributeSet, + use_value, + basenames[base_no]); + if (prefix_len < 0) + { + zi->errCode = 114; + return -1; + } + if (!relational_term (zi, zapt, term_sub, term_dict, + attributeSet, grep_info, &max_pos)) + { + switch (truncation_value) { - term_dict[j++] = '.'; - term_dict[j++] = '*'; + case -1: /* not specified */ + case 100: /* do not truncate */ + j = strlen(term_dict); + term_dict[j++] = '('; + for (i = 0; term_sub[i]; i++) + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ")"); + r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, + &max_pos, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r); + break; + case 1: /* right truncation */ + j = strlen(term_dict); + term_dict[j++] = '('; + for (i = 0; term_sub[i]; i++) + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ".*)"); + dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, + &max_pos, grep_handle); + break; + case 2: /* left truncation */ + case 3: /* left&right truncation */ + zi->errCode = 120; + return -1; + case 101: /* process # in term */ + j = strlen(term_dict); + term_dict[j++] = '('; + for (i=0; term_sub[i]; i++) + if (term_sub[i] == '#' && i > 2) + { + term_dict[j++] = '.'; + term_dict[j++] = '*'; + } + else + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ")"); + r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, + &max_pos, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", + r); + break; + case 102: /* regular expression */ + sprintf (term_dict + strlen(term_dict), "(%s)", term_sub); + r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, + &max_pos, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", + r); + break; } - else - term_dict[j++] = term_sub[i]; - term_dict[j] = '\0'; - dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, grep_handle); - break; - case 102: /* regular expression */ - strcat (term_dict, "("); - strcat (term_dict, term_sub); - strcat (term_dict, ")"); - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep fail, trunc=regular: %d", r); - break; + } + if (max_pos <= strlen(basenames[base_no])) + { + zi->errCode = 109; /* Database unavailable */ + zi->errString = basenames[base_no]; + return -1; + } } logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 0; @@ -1107,12 +1123,15 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, if (zs->which == Z_RPNStructure_complex) { rset_bool_parms bool_parms; + int soft = 0; bool_parms.rset_l = rpn_search_structure (zi, zs->u.complex->s1, attributeSet, num_bases, basenames); if (bool_parms.rset_l == NULL) return NULL; + if (rset_is_ranked(bool_parms.rset_l)) + soft = 1; bool_parms.rset_r = rpn_search_structure (zi, zs->u.complex->s2, attributeSet, num_bases, basenames); @@ -1121,19 +1140,21 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, rset_delete (bool_parms.rset_l); return NULL; } + if (rset_is_ranked(bool_parms.rset_r)) + soft = 1; bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare; switch (zs->u.complex->operator->which) { case Z_Operator_and: - r = rset_create (rset_kind_and, &bool_parms); + r = rset_create (soft ? rset_kind_sand:rset_kind_and, &bool_parms); break; case Z_Operator_or: - r = rset_create (rset_kind_or, &bool_parms); + r = rset_create (soft ? rset_kind_sor:rset_kind_or, &bool_parms); break; case Z_Operator_and_not: - r = rset_create (rset_kind_not, &bool_parms); + r = rset_create (soft ? rset_kind_snot:rset_kind_not, &bool_parms); break; default: assert (0); @@ -1164,13 +1185,46 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, return r; } +void count_set_save (RSET *r, int *count) +{ + int psysno = 0; + int kno = 0; + struct it_key key; + RSFD rfd, wfd; + RSET w; + rset_temp_parms parms; + + logf (LOG_DEBUG, "count_set_save"); + *count = 0; + parms.key_size = sizeof(struct it_key); + w = rset_create (rset_kind_temp, &parms); + wfd = rset_open (w, RSETF_WRITE|RSETF_SORT_SYSNO); + rfd = rset_open (*r, RSETF_READ|RSETF_SORT_SYSNO); + while (rset_read (*r, rfd, &key)) + { + if (key.sysno != psysno) + { + rset_write (w, wfd, &key); + psysno = key.sysno; + (*count)++; + } + kno++; + } + rset_close (*r, rfd); + rset_delete (*r); + rset_close (w, wfd); + *r = w; + logf (LOG_DEBUG, "%d keys, %d distinct sysnos", kno, *count); +} + static void count_set (RSET r, int *count) { int psysno = 0; + int kno = 0; struct it_key key; RSFD rfd; - logf (LOG_DEBUG, "rpn_save_set"); + logf (LOG_DEBUG, "count_set"); *count = 0; rfd = rset_open (r, RSETF_READ|RSETF_SORT_SYSNO); while (rset_read (r, rfd, &key)) @@ -1180,9 +1234,10 @@ static void count_set (RSET r, int *count) psysno = key.sysno; (*count)++; } + kno++; } rset_close (r, rfd); - logf (LOG_DEBUG, "%d distinct sysnos", *count); + logf (LOG_DEBUG, "%d keys, %d distinct sysnos", kno, *count); } int rpn_search (ZServerInfo *zi, @@ -1193,17 +1248,21 @@ int rpn_search (ZServerInfo *zi, oident *attrset; oid_value attributeSet; + zlog_rpn (rpn); + zi->errCode = 0; zi->errString = NULL; - + attrset = oid_getentbyoid (rpn->attributeSetId); attributeSet = attrset->value; - rset = rpn_search_structure (zi, rpn->RPNStructure, attributeSet, num_bases, basenames); if (!rset) return zi->errCode; - count_set (rset, hits); + if (rset_is_volatile(rset)) + count_set_save(&rset,hits); + else + count_set (rset, hits); resultSetAdd (zi, setname, 1, rset); if (zi->errCode) logf (LOG_DEBUG, "search error: %d", zi->errCode); @@ -1253,12 +1312,18 @@ static int scan_handle (Dict_char *name, const char *info, int pos, return 0; } -int rpn_scan (ZServerInfo *zi, ODR odr, Z_AttributesPlusTerm *zapt, + +static int dummy_handle (Dict_char *name, const char *info, void *p) +{ + return 0; +} + +int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, int num_bases, char **basenames, int *position, int *num_entries, struct scan_entry **list, int *status) { - int i, j, sizez; + int i, j, sizez, max_pos; int pos = *position; int num = *num_entries; int before; @@ -1275,12 +1340,13 @@ int rpn_scan (ZServerInfo *zi, ODR odr, Z_AttributesPlusTerm *zapt, return 111; scan_info.before = before = pos-1; scan_info.after = after = 1+num-pos; - scan_info.odr = odr; + scan_info.odr = zi->odr; logf (LOG_DEBUG, "scan, before = %d, after = %d", before, after); scan_info.isam = zi->wordIsam; - scan_info.list = odr_malloc (odr, (before+after)*sizeof(*scan_info.list)); + scan_info.list = odr_malloc (zi->odr, (before+after)* + sizeof(*scan_info.list)); for (j = 0; jwordDict, termz, 0, NULL, &max_pos, + dummy_handle); + if (max_pos <= strlen(*basenames)) + { + zi->errString = *basenames; + return zi->errCode = 109; /* Database unavailable */ + } strcpy (scan_info.prefix, termz); sizez = term->u.general->len; if (sizez > IT_MAX_WORD)