#include <string.h>
#include <assert.h>
+/*
+** Characters that may appear in the second argument to matchinfo().
+*/
+#define FTS3_MATCHINFO_NPHRASE 'p' /* 1 value */
+#define FTS3_MATCHINFO_NCOL 'c' /* 1 value */
+#define FTS3_MATCHINFO_NDOC 'n' /* 1 value */
+#define FTS3_MATCHINFO_AVGLENGTH 'a' /* nCol values */
+#define FTS3_MATCHINFO_LENGTH 'l' /* nCol values */
+#define FTS3_MATCHINFO_LCS 's' /* nCol values */
+#define FTS3_MATCHINFO_HITS 'x' /* 3*nCol*nPhrase values */
+
+/*
+** The default value for the second argument to matchinfo().
+*/
+#define FTS3_MATCHINFO_DEFAULT "pcx"
+
/*
** Used as an fts3ExprIterate() context when loading phrase doclists to
struct MatchInfo {
Fts3Cursor *pCursor; /* FTS3 Cursor */
int nCol; /* Number of columns in table */
+ int nPhrase; /* Number of matchable phrases in query */
+ sqlite3_int64 nDoc; /* Number of docs in database */
u32 *aMatchinfo; /* Pre-allocated buffer */
};
/*
** fts3ExprIterate() callback used to collect the "global" matchinfo stats
-** for a single query. The "global" stats are those elements of the matchinfo
-** array that are constant for all rows returned by the current query.
+** for a single query.
+**
+** fts3ExprIterate() callback to load the 'global' elements of a
+** FTS3_MATCHINFO_HITS matchinfo array. The global stats are those elements
+** of the matchinfo array that are constant for all rows returned by the
+** current query.
+**
+** Argument pCtx is actually a pointer to a struct of type MatchInfo. This
+** function populates Matchinfo.aMatchinfo[] as follows:
+**
+** for(iCol=0; iCol<nCol; iCol++){
+** aMatchinfo[3*iPhrase*nCol + 3*iCol + 1] = X;
+** aMatchinfo[3*iPhrase*nCol + 3*iCol + 2] = Y;
+** }
+**
+** where X is the number of matches for phrase iPhrase is column iCol of all
+** rows of the table. Y is the number of rows for which column iCol contains
+** at least one instance of phrase iPhrase.
*/
-static int fts3ExprGlobalMatchinfoCb(
+static int fts3ExprGlobalHitsCb(
Fts3Expr *pExpr, /* Phrase expression node */
int iPhrase, /* Phrase number (numbered from zero) */
void *pCtx /* Pointer to MatchInfo structure */
char *pIter;
char *pEnd;
char *pFree = 0;
- const int iStart = 2 + (iPhrase * p->nCol * 3) + 1;
+ u32 *aOut = &p->aMatchinfo[3*iPhrase*p->nCol];
assert( pExpr->isLoaded );
assert( pExpr->eType==FTSQUERY_PHRASE );
pIter = pFree;
pEnd = &pFree[nFree];
}else{
- int nDoc = p->aMatchinfo[2 + 3*p->nCol*p->aMatchinfo[0]];
- for(ii=0; ii<p->nCol; ii++){
- p->aMatchinfo[iStart + ii*3] = nDoc;
- p->aMatchinfo[iStart + ii*3 + 1] = nDoc;
+ int iCol; /* Column index */
+ for(iCol=0; iCol<p->nCol; iCol++){
+ aOut[iCol*3 + 1] = (u32)p->nDoc;
+ aOut[iCol*3 + 2] = (u32)p->nDoc;
}
return SQLITE_OK;
}
/* Fill in the global hit count matrix row for this phrase. */
while( pIter<pEnd ){
while( *pIter++ & 0x80 ); /* Skip past docid. */
- fts3LoadColumnlistCounts(&pIter, &p->aMatchinfo[iStart], 1);
+ fts3LoadColumnlistCounts(&pIter, &aOut[1], 1);
}
sqlite3_free(pFree);
}
/*
-** fts3ExprIterate() callback used to collect the "local" matchinfo stats
-** for a single query. The "local" stats are those elements of the matchinfo
+** fts3ExprIterate() callback used to collect the "local" part of the
+** FTS3_MATCHINFO_HITS array. The local stats are those elements of the
** array that are different for each row returned by the query.
*/
-static int fts3ExprLocalMatchinfoCb(
+static int fts3ExprLocalHitsCb(
Fts3Expr *pExpr, /* Phrase expression node */
int iPhrase, /* Phrase number */
void *pCtx /* Pointer to MatchInfo structure */
if( pExpr->aDoclist ){
char *pCsr;
- int iStart = 2 + (iPhrase * p->nCol * 3);
+ int iStart = iPhrase * p->nCol * 3;
int i;
for(i=0; i<p->nCol; i++) p->aMatchinfo[iStart+i*3] = 0;
return SQLITE_OK;
}
+static int fts3MatchinfoCheck(
+ Fts3Table *pTab,
+ char cArg,
+ char **pzErr
+){
+ if( cArg==FTS3_MATCHINFO_NPHRASE
+ || cArg==FTS3_MATCHINFO_NCOL
+ || cArg==FTS3_MATCHINFO_NDOC && pTab->bHasStat
+ || cArg==FTS3_MATCHINFO_AVGLENGTH && pTab->bHasStat
+ || cArg==FTS3_MATCHINFO_LENGTH && pTab->bHasDocsize
+ || cArg==FTS3_MATCHINFO_LCS
+ || cArg==FTS3_MATCHINFO_HITS
+ ){
+ return SQLITE_OK;
+ }
+ *pzErr = sqlite3_mprintf("unrecognized matchinfo request: %c", cArg);
+ return SQLITE_ERROR;
+}
+
+static int fts3MatchinfoSize(MatchInfo *pInfo, char cArg){
+ int nVal; /* Number of integers output by cArg */
+
+ switch( cArg ){
+ case FTS3_MATCHINFO_NDOC:
+ case FTS3_MATCHINFO_NPHRASE:
+ case FTS3_MATCHINFO_NCOL:
+ nVal = 1;
+ break;
+
+ case FTS3_MATCHINFO_AVGLENGTH:
+ case FTS3_MATCHINFO_LENGTH:
+ case FTS3_MATCHINFO_LCS:
+ nVal = pInfo->nCol;
+ break;
+
+ case FTS3_MATCHINFO_HITS:
+ nVal = pInfo->nCol * pInfo->nPhrase * 3;
+ break;
+ }
+
+ return nVal;
+}
+
+static int fts3MatchinfoSelectDoctotal(
+ Fts3Table *pTab,
+ sqlite3_stmt **ppStmt,
+ sqlite3_int64 *pnDoc,
+ const char **paLen
+){
+ sqlite3_stmt *pStmt;
+ const char *a;
+ sqlite3_int64 nDoc;
+
+ if( !*ppStmt ){
+ int rc = sqlite3Fts3SelectDoctotal(pTab, ppStmt);
+ if( rc!=SQLITE_OK ) return rc;
+ }
+ pStmt = *ppStmt;
+
+ a = sqlite3_column_blob(pStmt, 0);
+ a += sqlite3Fts3GetVarint(a, &nDoc);
+ *pnDoc = (u32)nDoc;
+
+ if( paLen ) *paLen = a;
+ return SQLITE_OK;
+}
+
+
+static int fts3MatchinfoValues(
+ Fts3Cursor *pCsr, /* FTS3 cursor object */
+ int bGlobal, /* True to grab the global stats */
+ MatchInfo *pInfo, /* Matchinfo context object */
+ const char *zArg /* Matchinfo format string */
+){
+ int rc = SQLITE_OK;
+ int i;
+ Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
+
+ sqlite3_stmt *pSelect = 0;
+
+ for(i=0; zArg[i]; i++){
+
+ switch( zArg[i] ){
+ case FTS3_MATCHINFO_NPHRASE:
+ if( bGlobal ) pInfo->aMatchinfo[0] = pInfo->nPhrase;
+ break;
+
+ case FTS3_MATCHINFO_NCOL:
+ if( bGlobal ) pInfo->aMatchinfo[0] = pInfo->nCol;
+ break;
+
+ case FTS3_MATCHINFO_NDOC:
+ if( bGlobal ){
+ sqlite3_int64 nDoc;
+ rc = fts3MatchinfoSelectDoctotal(pTab, &pSelect, &nDoc, 0);
+ pInfo->aMatchinfo[0] = (u32)nDoc;
+ }
+ break;
+
+ case FTS3_MATCHINFO_AVGLENGTH:
+ if( bGlobal ){
+ sqlite3_int64 nDoc; /* Number of rows in table */
+ const char *a; /* Aggregate column length array */
+
+ rc = fts3MatchinfoSelectDoctotal(pTab, &pSelect, &nDoc, &a);
+ if( rc==SQLITE_OK ){
+ int iCol;
+ for(iCol=0; iCol<pInfo->nCol; iCol++){
+ sqlite3_int64 nToken;
+ a += sqlite3Fts3GetVarint(a, &nToken);
+ pInfo->aMatchinfo[iCol] = ((u32)(nToken&0xffffffff)+nDoc/2)/nDoc;
+ }
+ }
+ }
+ break;
+
+ case FTS3_MATCHINFO_LENGTH: {
+ sqlite3_stmt *pSelectDocsize = 0;
+ rc = sqlite3Fts3SelectDocsize(pTab, pCsr->iPrevId, &pSelectDocsize);
+ if( rc==SQLITE_OK ){
+ int iCol;
+ const char *a = sqlite3_column_blob(pSelectDocsize, 0);
+ for(iCol=0; iCol<pInfo->nCol; iCol++){
+ sqlite3_int64 nToken;
+ a += sqlite3Fts3GetVarint(a, &nToken);
+ pInfo->aMatchinfo[iCol] = (u32)nToken;
+ }
+ }
+ sqlite3_reset(pSelectDocsize);
+ break;
+ }
+
+ case FTS3_MATCHINFO_HITS: {
+ Fts3Expr *pExpr = pCsr->pExpr;
+ if( bGlobal ){
+ if( pCsr->pDeferred ){
+ rc = fts3MatchinfoSelectDoctotal(pTab, &pSelect, &pInfo->nDoc, 0);
+ }
+ (void)fts3ExprIterate(pExpr, fts3ExprGlobalHitsCb,(void*)pInfo);
+ }
+ (void)fts3ExprIterate(pExpr, fts3ExprLocalHitsCb,(void*)pInfo);
+ break;
+ }
+
+
+ default:
+ assert( zArg[i]==FTS3_MATCHINFO_LCS );
+ }
+
+ pInfo->aMatchinfo += fts3MatchinfoSize(pInfo, zArg[i]);
+ }
+
+ sqlite3_reset(pSelect);
+ return rc;
+}
+
+
/*
** Populate pCsr->aMatchinfo[] with data for the current row. The
** 'matchinfo' data is an array of 32-bit unsigned integers (C type u32).
*/
-static int fts3GetMatchinfo(Fts3Cursor *pCsr){
+static int fts3GetMatchinfo(
+ Fts3Cursor *pCsr, /* FTS3 Cursor object */
+ const char *zArg /* Second argument to matchinfo() function */
+){
MatchInfo sInfo;
Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int rc = SQLITE_OK;
+ int bGlobal = 0; /* Collect 'global' stats as well as local */
+ memset(&sInfo, 0, sizeof(MatchInfo));
sInfo.pCursor = pCsr;
sInfo.nCol = pTab->nColumn;
+ /* If there is cached matchinfo() data, but the format string for the
+ ** cache does not match the format string for this request, discard
+ ** the cached data. */
+ if( pCsr->zMatchinfo && strcmp(pCsr->zMatchinfo, zArg) ){
+ assert( pCsr->aMatchinfo );
+ sqlite3_free(pCsr->aMatchinfo);
+ pCsr->zMatchinfo = 0;
+ pCsr->aMatchinfo = 0;
+ }
+
+ /* If Fts3Cursor.aMatchinfo[] is NULL, then this is the first time the
+ ** matchinfo function has been called for this query. In this case
+ ** allocate the array used to accumulate the matchinfo data and
+ ** initialize those elements that are constant for every row.
+ */
if( pCsr->aMatchinfo==0 ){
- /* If Fts3Cursor.aMatchinfo[] is NULL, then this is the first time the
- ** matchinfo function has been called for this query. In this case
- ** allocate the array used to accumulate the matchinfo data and
- ** initialize those elements that are constant for every row.
- */
- int nPhrase; /* Number of phrases */
- int nMatchinfo; /* Number of u32 elements in match-info */
+ int nMatchinfo = 0; /* Number of u32 elements in match-info */
+ int nArg; /* Bytes in zArg */
+ int i; /* Used to iterate through zArg */
/* Load doclists for each phrase in the query. */
- rc = fts3ExprLoadDoclists(pCsr, &nPhrase, 0);
- if( rc!=SQLITE_OK ){
- return rc;
- }
- nMatchinfo = 2 + 3*sInfo.nCol*nPhrase;
- if( pTab->bHasDocsize ){
- nMatchinfo += 1 + 2*pTab->nColumn;
- }
+ rc = fts3ExprLoadDoclists(pCsr, &pCsr->nPhrase, 0);
+ if( rc!=SQLITE_OK ) return rc;
+ sInfo.nPhrase = pCsr->nPhrase;
- sInfo.aMatchinfo = (u32 *)sqlite3_malloc(sizeof(u32)*nMatchinfo);
- if( !sInfo.aMatchinfo ){
- return SQLITE_NOMEM;
+ for(i=0; zArg[i]; i++){
+ nMatchinfo += fts3MatchinfoSize(&sInfo, zArg[i]);
}
- memset(sInfo.aMatchinfo, 0, sizeof(u32)*nMatchinfo);
-
- /* First element of match-info is the number of phrases in the query */
- sInfo.aMatchinfo[0] = nPhrase;
- sInfo.aMatchinfo[1] = sInfo.nCol;
- if( pTab->bHasDocsize ){
- int ofst = 2 + 3*sInfo.aMatchinfo[0]*sInfo.aMatchinfo[1];
- rc = sqlite3Fts3MatchinfoDocsizeGlobal(pCsr, &sInfo.aMatchinfo[ofst]);
- }
- (void)fts3ExprIterate(pCsr->pExpr, fts3ExprGlobalMatchinfoCb,(void*)&sInfo);
- pCsr->aMatchinfo = sInfo.aMatchinfo;
+
+ /* Allocate space for Fts3Cursor.aMatchinfo[] and Fts3Cursor.zMatchinfo. */
+ nArg = strlen(zArg);
+ pCsr->aMatchinfo = (u32 *)sqlite3_malloc(sizeof(u32)*nMatchinfo + nArg + 1);
+ if( !pCsr->aMatchinfo ) return SQLITE_NOMEM;
+
+ pCsr->zMatchinfo = (char *)&pCsr->aMatchinfo[nMatchinfo];
+ pCsr->nMatchinfo = nMatchinfo;
+ memcpy(pCsr->zMatchinfo, zArg, nArg+1);
+ memset(pCsr->aMatchinfo, 0, sizeof(u32)*nMatchinfo);
pCsr->isMatchinfoNeeded = 1;
+ bGlobal = 1;
}
sInfo.aMatchinfo = pCsr->aMatchinfo;
+ sInfo.nPhrase = pCsr->nPhrase;
if( rc==SQLITE_OK && pCsr->isMatchinfoNeeded ){
- (void)fts3ExprIterate(pCsr->pExpr, fts3ExprLocalMatchinfoCb, (void*)&sInfo);
- if( pTab->bHasDocsize ){
- int ofst = 2 + 3*sInfo.aMatchinfo[0]*sInfo.aMatchinfo[1];
- rc = sqlite3Fts3MatchinfoDocsizeLocal(pCsr, &sInfo.aMatchinfo[ofst]);
- }
+ rc = fts3MatchinfoValues(pCsr, bGlobal, &sInfo, zArg);
pCsr->isMatchinfoNeeded = 0;
}
- return SQLITE_OK;
+ return rc;
}
/*
/*
** Implementation of matchinfo() function.
*/
-void sqlite3Fts3Matchinfo(sqlite3_context *pContext, Fts3Cursor *pCsr){
+void sqlite3Fts3Matchinfo(
+ sqlite3_context *pContext, /* Function call context */
+ Fts3Cursor *pCsr, /* FTS3 table cursor */
+ const char *zArg /* Second arg to matchinfo() function */
+){
+ Fts3Table *pTab = (Fts3Table *)pCsr->base.pVtab;
int rc;
+ int i;
+ const char *zFormat;
+
+ if( zArg ){
+ for(i=0; zArg[i]; i++){
+ char *zErr = 0;
+ if( fts3MatchinfoCheck(pTab, zArg[i], &zErr) ){
+ sqlite3_result_error(pContext, zErr, -1);
+ sqlite3_free(zErr);
+ return;
+ }
+ }
+ zFormat = zArg;
+ }else{
+ zFormat = FTS3_MATCHINFO_DEFAULT;
+ }
+
if( !pCsr->pExpr ){
sqlite3_result_blob(pContext, "", 0, SQLITE_STATIC);
return;
}
- rc = fts3GetMatchinfo(pCsr);
- sqlite3Fts3SegmentsClose((Fts3Table *)pCsr->base.pVtab );
+
+ /* Retrieve matchinfo() data. */
+ rc = fts3GetMatchinfo(pCsr, zFormat);
+ sqlite3Fts3SegmentsClose(pTab);
+
if( rc!=SQLITE_OK ){
sqlite3_result_error_code(pContext, rc);
}else{
- Fts3Table *pTab = (Fts3Table*)pCsr->base.pVtab;
- int n = sizeof(u32)*(2+pCsr->aMatchinfo[0]*pCsr->aMatchinfo[1]*3);
- if( pTab->bHasDocsize ){
- n += sizeof(u32)*(1 + 2*pTab->nColumn);
- }
+ int n = pCsr->nMatchinfo * sizeof(u32);
sqlite3_result_blob(pContext, pCsr->aMatchinfo, n, SQLITE_TRANSIENT);
}
}
-C Fix\sa\stypo\sin\sunixCurrentTimeInt64()\spreventing\scompilation\swith\sNO_GETTOD\sdefined.
-D 2010-11-22T17:26:07
+C Experimental\schanges\sto\sfts3\sfunction\smatchinfo().
+D 2010-11-23T19:16:48
F Makefile.arm-wince-mingw32ce-gcc d6df77f1f48d690bd73162294bbba7f59507c72f
F Makefile.in e7a59672eaeb04408d1fa8501618d7501a3c5e39
F Makefile.linux-gcc 91d710bdc4998cb015f39edf3cb314ec4f4d7e23
F ext/fts3/README.syntax a19711dc5458c20734b8e485e75fb1981ec2427a
F ext/fts3/README.tokenizers 998756696647400de63d5ba60e9655036cb966e9
F ext/fts3/README.txt 8c18f41574404623b76917b9da66fcb0ab38328d
-F ext/fts3/fts3.c 7e01fd82fa289decbc21bcebb924d44d307eb757
+F ext/fts3/fts3.c 591b6fdb563dd00cd05b57ea27272f388e8156c8
F ext/fts3/fts3.h 3a10a0af180d502cecc50df77b1b22df142817fe
-F ext/fts3/fts3Int.h bbcd6d671228c9ae2ed5514723fcc6217821d0a6
+F ext/fts3/fts3Int.h 0f7b9f189bcb0eed43a9489d2251b99acf06b833
F ext/fts3/fts3_expr.c ee48b9278b8b2432a05a03320fbcacba151dbaa5
F ext/fts3/fts3_hash.c 3c8f6387a4a7f5305588b203fa7c887d753e1f1c
F ext/fts3/fts3_hash.h 8331fb2206c609f9fc4c4735b9ab5ad6137c88ec
F ext/fts3/fts3_icu.c ac494aed69835008185299315403044664bda295
F ext/fts3/fts3_porter.c 8df6f6efcc4e9e31f8bf73a4007c2e9abca1dfba
-F ext/fts3/fts3_snippet.c 300c12b7f0a2a6ae0491bb2d00e2d5ff9c28f685
+F ext/fts3/fts3_snippet.c aa0a76088024a693ab6a51d0d383e06f97e4fd8c
F ext/fts3/fts3_tokenizer.c 1301b0ee3ef414caae3257a702215925cc48cd9c
F ext/fts3/fts3_tokenizer.h 13ffd9fcb397fec32a05ef5cd9e0fa659bf3dbd3
F ext/fts3/fts3_tokenizer1.c 6e5cbaa588924ac578263a598e4fb9f5c9bb179d
-F ext/fts3/fts3_write.c 3d12dad6cbe788ff7214ef227520fb6600ff5112
+F ext/fts3/fts3_write.c f5e7db79057ea46c1f3d15a4e2c8b658d306e37f
F ext/fts3/fts3speed.tcl b54caf6a18d38174f1a6e84219950d85e98bb1e9
F ext/fts3/mkfts3amal.tcl 252ecb7fe6467854f2aa237bf2c390b74e71f100
F ext/icu/README.txt bf8461d8cdc6b8f514c080e4e10dc3b2bbdfefa9
F test/fts3cov.test e0fb00d8b715ddae4a94c305992dfc3ef70353d7
F test/fts3d.test 95fb3c862cbc4297c93fceb9a635543744e9ef52
F test/fts3defer.test d6cb0db9b5997ecf863d96ff419f83f8f2c87f4f
-F test/fts3defer2.test d3c7db6584aab06a2781b8de58747c33b23cb19c
+F test/fts3defer2.test 548eb2ca7e6a1515a7bc151721e223be14c51f9d
F test/fts3e.test 1f6c6ac9cc8b772ca256e6b22aaeed50c9350851
F test/fts3expr.test 5e745b2b6348499d9ef8d59015de3182072c564c
F test/fts3expr2.test 18da930352e5693eaa163a3eacf96233b7290d1a
F test/fts3fault.test 81fd40ceb12f33f9d16c5637d0f8d95d4556c456
F test/fts3malloc.test 9c8cc3f885bb4dfc66d0460c52f68f45e4710d1b
-F test/fts3matchinfo.test 211b04434926bce9cc2b3cc5f58725affefe5165
+F test/fts3matchinfo.test 051bc3d09152998cedb7b9bdfd5b0730b5c0a939
F test/fts3near.test 2e318ee434d32babd27c167142e2b94ddbab4844
-F test/fts3query.test df864f4c9b14927daa7199242d82cff44599327b
+F test/fts3query.test ef79d31fdb355d094baec1c1b24b60439a1fb8a2
F test/fts3rnd.test 707533ce943f490443ce5e696236bb1675a37635
F test/fts3shared.test 8bb266521d7c5495c0ae522bb4d376ad5387d4a2
F test/fts3snippet.test a12f22a3ba4dd59751a57c79b031d07ab5f51ddd
F tool/speedtest8.c 2902c46588c40b55661e471d7a86e4dd71a18224
F tool/speedtest8inst1.c 293327bc76823f473684d589a8160bde1f52c14e
F tool/vdbe-compress.tcl d70ea6d8a19e3571d7ab8c9b75cba86d1173ff0f
-P e38c81cc18d2ceaa1644aaba64530ba5d1fbf2cf
-R d7dd6f7536f2f168157ffbcee03e168e
+P 3df3e79b56821201b4f5ecd23f94d485745c48c3
+R d2823d2eb839cb4c673b301711f75566
+T *branch * fts3-experimental
+T *sym-fts3-experimental *
+T -sym-trunk *
U dan
-Z 238874a9476b4d03764b69cbc465de50
+Z cb34d18554c4ac084411d81cdabfad29
projects / thirdparty / sqlite.git / commitdiff
