-/* $Id: extract.c,v 1.127 2002-10-23 14:28:20 adam Exp $
+/* $Id: extract.c,v 1.137 2003-02-26 12:30:54 pop Exp $
Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
Index Data Aps
spec_src = rGroup->groupName;
else if (!strcmp (special, "database"))
spec_src = rGroup->databaseName;
- else if (!strcmp (special, "filename"))
+ else if (!strcmp (special, "filename")) {
spec_src = fname;
+ }
else if (!strcmp (special, "type"))
spec_src = rGroup->recordType;
else
close (fd);
return r;
}
-
-
int extract_rec_in_mem (ZebraHandle zh, const char *recordType,
const char *buf, size_t buf_size,
const char *databaseName, int delete_flag,
int store_keys, int store_data,
const char *match_criteria)
{
+ struct recordGroup rGroup;
+ rGroup.groupName = NULL;
+ rGroup.databaseName = (char *)databaseName;
+ rGroup.path = NULL;
+ rGroup.recordId = NULL;
+ rGroup.recordType = (char *)recordType;
+ rGroup.flagStoreData = store_data;
+ rGroup.flagStoreKeys = store_keys;
+ rGroup.flagRw = 1;
+ rGroup.databaseNamePath = 0;
+ rGroup.explainDatabase = 0;
+ rGroup.fileVerboseLimit = 100000;
+ rGroup.followLinks = -1;
+ return (bufferExtractRecord (zh,
+ buf, buf_size,
+ &rGroup,
+ delete_flag,
+ test_mode,
+ recordType,
+ sysno,
+ match_criteria,
+ "<no file>"));
+}
+/*
+ If sysno is provided, then it's used to identify the reocord.
+ If not, and match_criteria is provided, then sysno is guessed
+ If not, and a record is provided, then sysno is got from there
+
+ */
+int bufferExtractRecord (ZebraHandle zh,
+ const char *buf, size_t buf_size,
+ struct recordGroup *rGroup,
+ int delete_flag,
+ int test_mode,
+ const char *recordType,
+ int *sysno,
+ const char *match_criteria,
+ const char *fname)
+
+{
RecordAttr *recordAttr;
struct recExtractCtrl extractCtrl;
int i, r;
char *matchStr = 0;
- RecType recType;
+ RecType recType = NULL;
char subType[1024];
void *clientData;
- const char *fname = "<no file>";
Record rec;
long recordOffset = 0;
struct zebra_fetch_control fc;
extractCtrl.fh = &fc;
/* announce database */
- if (zebraExplain_curDatabase (zh->reg->zei, databaseName))
- {
- if (zebraExplain_newDatabase (zh->reg->zei, databaseName, 0))
- return 0;
- }
- if (!(recType =
- recType_byName (zh->reg->recTypes, recordType, subType,
- &clientData)))
+ if (zebraExplain_curDatabase (zh->reg->zei, rGroup->databaseName))
{
- logf (LOG_WARN, "No such record type: %s", recordType);
- return 0;
+ if (zebraExplain_newDatabase (zh->reg->zei, rGroup->databaseName, 0))
+ return 0;
}
zh->reg->keys.buf_used = 0;
zh->reg->keys.prevAttrSet = -1;
zh->reg->keys.prevSeqNo = 0;
zh->reg->sortKeys.buf_used = 0;
+ zh->reg->sortKeys.buf_max = 0;
+ zh->reg->sortKeys.buf = 0;
+
+ if (*recordType) {
+ logf (LOG_DEBUG, "Record type explicitly specified: %s", recordType);
+ recType = recType_byName (zh->reg->recTypes, recordType, subType,
+ &clientData);
+ } else {
+ if (!(rGroup->recordType)) {
+ logf (LOG_WARN, "No such record type defined");
+ return 0;
+ }
+ logf (LOG_DEBUG, "Get record type from rgroup: %s",rGroup->recordType);
+ recType = recType_byName (zh->reg->recTypes, rGroup->recordType, subType,
+ &clientData);
+ recordType = rGroup->recordType;
+ }
+
+ if (!recType) {
+ logf (LOG_WARN, "No such record type: %s", rGroup->recordType);
+ return 0;
+ }
extractCtrl.subType = subType;
extractCtrl.init = extract_init;
return 1;
}
/* match criteria */
+ matchStr = NULL;
+
+ if (! *sysno) {
+ char *rinfo;
+ if (strlen(match_criteria) > 0) {
+ matchStr = (char *)match_criteria;
+ } else {
+ if (rGroup->recordId && *rGroup->recordId) {
+ matchStr = fileMatchStr (zh, &zh->reg->keys, rGroup, fname,
+ rGroup->recordId);
+ }
+ }
+ if (matchStr) {
+ rinfo = dict_lookup (zh->reg->matchDict, matchStr);
+ if (rinfo)
+ memcpy (sysno, rinfo+1, sizeof(*sysno));
+ } else {
+ logf (LOG_WARN, "Bad match criteria (recordID)");
+ return 0;
+ }
+
+ }
if (! *sysno)
{
}
extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys);
extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys);
+
+ zh->records_inserted++;
}
else
{
/* record already exists */
struct recKeys delkeys;
+ struct sortKeys sortKeys;
rec = rec_get (zh->reg->records, *sysno);
assert (rec);
{
logf (LOG_LOG, "skipped %s %s %ld", recordType,
fname, (long) recordOffset);
+ extract_flushSortKeys (zh, *sysno, -1, &zh->reg->sortKeys);
rec_rm (&rec);
+ logRecord(zh);
return 1;
}
delkeys.buf_used = rec->size[recInfo_delKeys];
delkeys.buf = rec->info[recInfo_delKeys];
- extract_flushSortKeys (zh, *sysno, 0, &zh->reg->sortKeys);
+
+ sortKeys.buf_used = rec->size[recInfo_sortKeys];
+ sortKeys.buf = rec->info[recInfo_sortKeys];
+
+ extract_flushSortKeys (zh, *sysno, 0, &sortKeys);
extract_flushRecordKeys (zh, *sysno, 0, &delkeys);
if (delete_flag)
{
{
logf (LOG_LOG, "delete %s %s %ld", recordType,
fname, (long) recordOffset);
-#if 0
+ zh->records_deleted++;
if (matchStr)
- dict_delete (matchDict, matchStr);
-#endif
+ dict_delete (zh->reg->matchDict, matchStr);
rec_del (zh->reg->records, &rec);
}
rec_rm (&rec);
+ logRecord(zh);
return 1;
}
else
fname, (long) recordOffset);
extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys);
extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys);
+ zh->records_updated++;
}
}
}
/* update delete keys */
xfree (rec->info[recInfo_delKeys]);
- if (zh->reg->keys.buf_used > 0 && store_keys == 1)
+ if (zh->reg->keys.buf_used > 0 && rGroup->flagStoreKeys == 1)
{
rec->size[recInfo_delKeys] = zh->reg->keys.buf_used;
rec->info[recInfo_delKeys] = zh->reg->keys.buf;
rec->size[recInfo_delKeys] = 0;
}
+ /* update sort keys */
+ xfree (rec->info[recInfo_sortKeys]);
+
+ rec->size[recInfo_sortKeys] = zh->reg->sortKeys.buf_used;
+ rec->info[recInfo_sortKeys] = zh->reg->sortKeys.buf;
+ zh->reg->sortKeys.buf = NULL;
+ zh->reg->sortKeys.buf_max = 0;
+
/* save file size of original record */
zebraExplain_recordBytesIncrement (zh->reg->zei,
- recordAttr->recordSize);
/* update store data */
xfree (rec->info[recInfo_storeData]);
- if (store_data == 1)
+ if (rGroup->flagStoreData == 1)
{
rec->size[recInfo_storeData] = recordAttr->recordSize;
rec->info[recInfo_storeData] = (char *)
/* update database name */
xfree (rec->info[recInfo_databaseName]);
rec->info[recInfo_databaseName] =
- rec_strdup (databaseName, &rec->size[recInfo_databaseName]);
+ rec_strdup (rGroup->databaseName, &rec->size[recInfo_databaseName]);
/* update offset */
recordAttr->recordOffset = recordOffset;
/* commit this record */
rec_put (zh->reg->records, &rec);
-
+ logRecord(zh);
return 0;
}
+
+
int explain_extract (void *handle, Record rec, data1_node *n)
{
ZebraHandle zh = (ZebraHandle) handle;
static void extract_add_sort_string (RecWord *p, const char *string,
int length)
{
-#if 1
ZebraHandle zh = p->extractCtrl->handle;
struct sortKeys *sk = &zh->reg->sortKeys;
- size_t off = 0;
- int slen;
+ int off = 0;
while (off < sk->buf_used)
{
- int set, use, l;
+ int set, use, slen;
- l = key_SU_decode(&set, sk->buf + off);
- off += l;
- l = key_SU_decode(&use, sk->buf + off);
- off += l;
- l = key_SU_decode(&slen, sk->buf + off);
- off += l + slen;
+ off += key_SU_decode(&set, sk->buf + off);
+ off += key_SU_decode(&use, sk->buf + off);
+ off += key_SU_decode(&slen, sk->buf + off);
+ off += slen;
if (p->attrSet == set && p->attrUse == use)
return;
}
}
off += key_SU_encode(p->attrSet, sk->buf + off);
off += key_SU_encode(p->attrUse, sk->buf + off);
- slen = strlen(string);
- off += key_SU_encode(slen, sk->buf + off);
- memcpy (sk->buf + off, string, slen);
- sk->buf_used = off + slen;
-#else
- struct sortKey *sk;
- ZebraHandle zh = p->extractCtrl->handle;
-
-
- for (sk = zh->reg->sortKeys; sk; sk = sk->next)
- if (sk->attrSet == p->attrSet && sk->attrUse == p->attrUse)
- return;
-
- sk = (struct sortKey *) xmalloc (sizeof(*sk));
- sk->next = zh->reg->sortKeys;
- zh->reg->sortKeys = sk;
-
- sk->string = (char *) xmalloc (length);
- sk->length = length;
- memcpy (sk->string, string, length);
-
- sk->attrSet = p->attrSet;
- sk->attrUse = p->attrUse;
-#endif
+ off += key_SU_encode(length, sk->buf + off);
+ memcpy (sk->buf + off, string, length);
+ sk->buf_used = off + length;
}
void extract_add_string (RecWord *p, const char *string, int length)
void extract_token_add (RecWord *p)
{
WRBUF wrbuf;
+
+#if 0
+ yaz_log (LOG_LOG, "reg_type=%c attrSet=%d attrUse=%d seqno=%d s=%.*s",
+ p->reg_type, p->attrSet, p->attrUse, p->seqno, p->length,
+ p->string);
+#endif
if ((wrbuf = zebra_replace(p->zebra_maps, p->reg_type, 0,
p->string, p->length)))
{
int cmd, struct sortKeys *sk)
{
SortIdx sortIdx = zh->reg->sortIdx;
- size_t off = 0;
+ int off = 0;
sortIdx_sysno (sortIdx, sysno);
while (off < sk->buf_used)
{
- int set, use, slen, l;
+ int set, use, slen;
off += key_SU_decode(&set, sk->buf + off);
off += key_SU_decode(&use, sk->buf + off);
while ((*bp++ = *k++))
;
i->keylen= bp - i->buf -1;
+ assert(i->keylen+1+sizeof(struct it_key) < ENCODE_BUFLEN);
}
else
{
i->prevseq=key.seqno;
i->prevcmd=*k;
}
-#ifdef SKIPTHIS_OLDCODE
- bp = encode_key_int ( (key.sysno - i->sysno) * 2 + *k, bp);
- if (i->sysno != key.sysno)
- {
- i->sysno = key.sysno;
- i->seqno = 0;
- }
- else if (!i->seqno && !key.seqno && i->cmd == *k)
- return;
- bp = encode_key_int (key.seqno - i->seqno, bp);
- i->seqno = key.seqno;
- i->cmd = *k;
- if (fwrite (i->buf, bp - i->buf, 1, outf) != 1)
- {
- logf (LOG_FATAL|LOG_ERRNO, "fwrite");
- exit (1);
- }
- i->keylen=0; /* ok, it's written, forget it */
-#endif
}
void encode_key_flush (struct encode_info *i, FILE *outf)
projects / idzebra-moved-to-github.git / blobdiff