X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fextract.c;h=a7ef24f8bb64168ac770cb718e7a4178f8dd1440;hb=ea597d9e75a2106887526480b81b883d908e0fb4;hp=b1269a0ad93a9e82cd436ecee5e8081522023706;hpb=5e6b6a90f396b08a6c8f8bd626e8961a08009ba1;p=idzebra-moved-to-github.git diff --git a/index/extract.c b/index/extract.c index b1269a0..a7ef24f 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,4 +1,4 @@ -/* $Id: extract.c,v 1.129 2002-10-24 21:54:29 adam Exp $ +/* $Id: extract.c,v 1.132 2002-11-26 22:18:34 adam Exp $ Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002 Index Data Aps @@ -361,8 +361,9 @@ static char *fileMatchStr (ZebraHandle zh, 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 @@ -890,8 +891,6 @@ int fileExtract (ZebraHandle zh, SYSNO *sysno, const char *fname, 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, @@ -899,6 +898,44 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, 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, + sysno, + match_criteria, + "")); +} +/* + 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, + int *sysno, + const char *match_criteria, + const char *fname) + +{ RecordAttr *recordAttr; struct recExtractCtrl extractCtrl; int i, r; @@ -906,7 +943,7 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, RecType recType; char subType[1024]; void *clientData; - const char *fname = ""; + SYSNO sysnotmp; Record rec; long recordOffset = 0; struct zebra_fetch_control fc; @@ -926,24 +963,31 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, extractCtrl.fh = &fc; /* announce database */ - if (zebraExplain_curDatabase (zh->reg->zei, databaseName)) + if (zebraExplain_curDatabase (zh->reg->zei, rGroup->databaseName)) { - if (zebraExplain_newDatabase (zh->reg->zei, databaseName, 0)) + if (zebraExplain_newDatabase (zh->reg->zei, rGroup->databaseName, 0)) return 0; } + + if (!(rGroup->recordType)) { + logf (LOG_WARN, "No such record type defined"); + return 0; + } + if (!(recType = - recType_byName (zh->reg->recTypes, recordType, subType, + recType_byName (zh->reg->recTypes, rGroup->recordType, subType, &clientData))) { - logf (LOG_WARN, "No such record type: %s", recordType); + logf (LOG_WARN, "No such record type: %s", rGroup->recordType); return 0; } - zh->reg->keys.buf_used = 0; zh->reg->keys.prevAttrUse = -1; 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; extractCtrl.subType = subType; extractCtrl.init = extract_init; @@ -988,18 +1032,40 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, 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) { /* new record */ if (delete_flag) { - logf (LOG_LOG, "delete %s %s %ld", recordType, + logf (LOG_LOG, "delete %s %s %ld", rGroup->recordType, fname, (long) recordOffset); logf (LOG_WARN, "cannot delete record above (seems new)"); return 1; } - logf (LOG_LOG, "add %s %s %ld", recordType, fname, + logf (LOG_LOG, "add %s %s %ld", rGroup->recordType, fname, (long) recordOffset); rec = rec_new (zh->reg->records); @@ -1014,11 +1080,14 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, } 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); @@ -1028,35 +1097,41 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, if (recordAttr->runNumber == zebraExplain_runNumberIncrement (zh->reg->zei, 0)) { - logf (LOG_LOG, "skipped %s %s %ld", recordType, + logf (LOG_LOG, "skipped %s %s %ld", rGroup->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) { /* record going to be deleted */ if (!delkeys.buf_used) { - logf (LOG_LOG, "delete %s %s %ld", recordType, + logf (LOG_LOG, "delete %s %s %ld", rGroup->recordType, fname, (long) recordOffset); logf (LOG_WARN, "cannot delete file above, storeKeys false"); } else { - logf (LOG_LOG, "delete %s %s %ld", recordType, + logf (LOG_LOG, "delete %s %s %ld", rGroup->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 @@ -1064,23 +1139,24 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, /* record going to be updated */ if (!delkeys.buf_used) { - logf (LOG_LOG, "update %s %s %ld", recordType, + logf (LOG_LOG, "update %s %s %ld", rGroup->recordType, fname, (long) recordOffset); logf (LOG_WARN, "cannot update file above, storeKeys false"); } else { - logf (LOG_LOG, "update %s %s %ld", recordType, + logf (LOG_LOG, "update %s %s %ld", rGroup->recordType, fname, (long) recordOffset); extract_flushSortKeys (zh, *sysno, 1, &zh->reg->sortKeys); extract_flushRecordKeys (zh, *sysno, 1, &zh->reg->keys); + zh->records_updated++; } } } /* update file type */ xfree (rec->info[recInfo_fileType]); rec->info[recInfo_fileType] = - rec_strdup (recordType, &rec->size[recInfo_fileType]); + rec_strdup (rGroup->recordType, &rec->size[recInfo_fileType]); /* update filename */ xfree (rec->info[recInfo_filename]); @@ -1089,7 +1165,7 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, /* 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; @@ -1102,6 +1178,14 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, 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); @@ -1121,7 +1205,7 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, /* 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 *) @@ -1152,17 +1236,19 @@ int extract_rec_in_mem (ZebraHandle zh, const char *recordType, /* 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;