From: dan <dan@noemail.net>
Date: Tue, 27 Jan 2015 18:43:02 +0000 (+0000)
Subject: Improve the performance of fts3/4 queries that use the OR operator and at least one... 
X-Git-Tag: version-3.8.9~152
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=18f6ff9eb7db02356102283c28053b0a602f55d7;p=thirdparty%2Fsqlite.git

Improve the performance of fts3/4 queries that use the OR operator and at least one auxiliary fts function.

FossilOrigin-Name: 245e8730451fbdc1c729beff7295c452df604009
---

diff --git a/ext/fts3/fts3.c b/ext/fts3/fts3.c
index fe28eb2cfe..ec9f0d3181 100644
--- a/ext/fts3/fts3.c
+++ b/ext/fts3/fts3.c
@@ -5020,6 +5020,22 @@ static void fts3EvalNextRow(
           }
           pExpr->iDocid = pLeft->iDocid;
           pExpr->bEof = (pLeft->bEof || pRight->bEof);
+          if( pExpr->eType==FTSQUERY_NEAR && pExpr->bEof ){
+            if( pRight->pPhrase && pRight->pPhrase->doclist.aAll ){
+              Fts3Doclist *pDl = &pRight->pPhrase->doclist;
+              while( *pRc==SQLITE_OK && pRight->bEof==0 ){
+                memset(pDl->pList, 0, pDl->nList);
+                fts3EvalNextRow(pCsr, pRight, pRc);
+              }
+            }
+            if( pLeft->pPhrase && pLeft->pPhrase->doclist.aAll ){
+              Fts3Doclist *pDl = &pLeft->pPhrase->doclist;
+              while( *pRc==SQLITE_OK && pLeft->bEof==0 ){
+                memset(pDl->pList, 0, pDl->nList);
+                fts3EvalNextRow(pCsr, pLeft, pRc);
+              }
+            }
+          }
         }
         break;
       }
@@ -5392,6 +5408,7 @@ static void fts3EvalRestart(
       }
       pPhrase->doclist.pNextDocid = 0;
       pPhrase->doclist.iDocid = 0;
+      pPhrase->pOrPoslist = 0;
     }
 
     pExpr->iDocid = 0;
@@ -5637,6 +5654,7 @@ int sqlite3Fts3EvalPhrasePoslist(
   iDocid = pExpr->iDocid;
   pIter = pPhrase->doclist.pList;
   if( iDocid!=pCsr->iPrevId || pExpr->bEof ){
+    int rc = SQLITE_OK;
     int bDescDoclist = pTab->bDescIdx;      /* For DOCID_CMP macro */
     int iMul;                     /* +1 if csr dir matches index dir, else -1 */
     int bOr = 0;
@@ -5662,72 +5680,43 @@ int sqlite3Fts3EvalPhrasePoslist(
     ** an incremental phrase. Load the entire doclist for the phrase
     ** into memory in this case.  */
     if( pPhrase->bIncr ){
-      int rc = SQLITE_OK;
-      int bEofSave = pExpr->bEof;
-      fts3EvalRestart(pCsr, pExpr, &rc);
-      while( rc==SQLITE_OK && !pExpr->bEof ){
-        fts3EvalNextRow(pCsr, pExpr, &rc);
-        if( bEofSave==0 && pExpr->iDocid==iDocid ) break;
+      int bEofSave = pNear->bEof;
+      fts3EvalRestart(pCsr, pNear, &rc);
+      while( rc==SQLITE_OK && !pNear->bEof ){
+        fts3EvalNextRow(pCsr, pNear, &rc);
+        if( bEofSave==0 && pNear->iDocid==iDocid ) break;
       }
-      pIter = pPhrase->doclist.pList;
       assert( rc!=SQLITE_OK || pPhrase->bIncr==0 );
-      if( rc!=SQLITE_OK ) return rc;
     }
-    
-    iMul = ((pCsr->bDesc==bDescDoclist) ? 1 : -1);
-    while( bTreeEof==1 
-        && pNear->bEof==0
-        && (DOCID_CMP(pNear->iDocid, pCsr->iPrevId) * iMul)<0
-    ){
-      int rc = SQLITE_OK;
-      fts3EvalNextRow(pCsr, pExpr, &rc);
-      if( rc!=SQLITE_OK ) return rc;
-      iDocid = pExpr->iDocid;
-      pIter = pPhrase->doclist.pList;
+    if( bTreeEof ){
+      while( rc==SQLITE_OK && !pNear->bEof ){
+        fts3EvalNextRow(pCsr, pNear, &rc);
+      }
     }
+    if( rc!=SQLITE_OK ) return rc;
 
-    bEof = (pPhrase->doclist.nAll==0);
-    assert( bDescDoclist==0 || bDescDoclist==1 );
-    assert( pCsr->bDesc==0 || pCsr->bDesc==1 );
-
-    if( bEof==0 ){
-      if( pCsr->bDesc==bDescDoclist ){
+    pIter = pPhrase->pOrPoslist;
+    iDocid = pPhrase->iOrDocid;
+    if( pCsr->bDesc==bDescDoclist ){
+      bEof = (pIter >= (pPhrase->doclist.aAll + pPhrase->doclist.nAll));
+      while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)<0 ) && bEof==0 ){
+        sqlite3Fts3DoclistNext(
+            bDescDoclist, pPhrase->doclist.aAll, pPhrase->doclist.nAll, 
+            &pIter, &iDocid, &bEof
+        );
+      }
+    }else{
+      bEof = !pPhrase->doclist.nAll || (pIter && pIter<=pPhrase->doclist.aAll);
+      while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)>0 ) && bEof==0 ){
         int dummy;
-        if( pNear->bEof ){
-          /* This expression is already at EOF. So position it to point to the
-          ** last entry in the doclist at pPhrase->doclist.aAll[]. Variable
-          ** iDocid is already set for this entry, so all that is required is
-          ** to set pIter to point to the first byte of the last position-list
-          ** in the doclist. 
-          **
-          ** It would also be correct to set pIter and iDocid to zero. In
-          ** this case, the first call to sqltie3Fts4DoclistPrev() below
-          ** would also move the iterator to point to the last entry in the 
-          ** doclist. However, this is expensive, as to do so it has to 
-          ** iterate through the entire doclist from start to finish (since
-          ** it does not know the docid for the last entry).  */
-          pIter = &pPhrase->doclist.aAll[pPhrase->doclist.nAll-1];
-          fts3ReversePoslist(pPhrase->doclist.aAll, &pIter);
-        }
-        while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)>0 ) && bEof==0 ){
-          sqlite3Fts3DoclistPrev(
-              bDescDoclist, pPhrase->doclist.aAll, pPhrase->doclist.nAll, 
-              &pIter, &iDocid, &dummy, &bEof
-          );
-        }
-      }else{
-        if( pNear->bEof ){
-          pIter = 0;
-          iDocid = 0;
-        }
-        while( (pIter==0 || DOCID_CMP(iDocid, pCsr->iPrevId)<0 ) && bEof==0 ){
-          sqlite3Fts3DoclistNext(
-              bDescDoclist, pPhrase->doclist.aAll, pPhrase->doclist.nAll, 
-              &pIter, &iDocid, &bEof
-          );
-        }
+        sqlite3Fts3DoclistPrev(
+            bDescDoclist, pPhrase->doclist.aAll, pPhrase->doclist.nAll, 
+            &pIter, &iDocid, &dummy, &bEof
+        );
       }
     }
+    pPhrase->pOrPoslist = pIter;
+    pPhrase->iOrDocid = iDocid;
 
     if( bEof || iDocid!=pCsr->iPrevId ) pIter = 0;
   }
@@ -5741,10 +5730,13 @@ int sqlite3Fts3EvalPhrasePoslist(
   }
   while( iThis<iCol ){
     fts3ColumnlistCopy(0, &pIter);
-    if( *pIter==0x00 ) return 0;
+    if( *pIter==0x00 ) return SQLITE_OK;
     pIter++;
     pIter += fts3GetVarint32(pIter, &iThis);
   }
+  if( *pIter==0x00 ){
+    pIter = 0;
+  }
 
   *ppOut = ((iCol==iThis)?pIter:0);
   return SQLITE_OK;
diff --git a/ext/fts3/fts3Int.h b/ext/fts3/fts3Int.h
index b2827b7352..1b8b7bd97e 100644
--- a/ext/fts3/fts3Int.h
+++ b/ext/fts3/fts3Int.h
@@ -375,6 +375,11 @@ struct Fts3Phrase {
   int bIncr;                 /* True if doclist is loaded incrementally */
   int iDoclistToken;
 
+  /* Used by sqlite3Fts3EvalPhrasePoslist() if this is a descendent of an
+  ** OR condition.  */
+  char *pOrPoslist;
+  i64 iOrDocid;
+
   /* Variables below this point are populated by fts3_expr.c when parsing 
   ** a MATCH expression. Everything above is part of the evaluation phase. 
   */
diff --git a/ext/fts3/fts3_snippet.c b/ext/fts3/fts3_snippet.c
index aa8779fa61..93ee86f6af 100644
--- a/ext/fts3/fts3_snippet.c
+++ b/ext/fts3/fts3_snippet.c
@@ -442,37 +442,39 @@ static int fts3BestSnippet(
   sIter.nSnippet = nSnippet;
   sIter.nPhrase = nList;
   sIter.iCurrent = -1;
-  (void)fts3ExprIterate(pCsr->pExpr, fts3SnippetFindPositions, (void *)&sIter);
+  rc = fts3ExprIterate(pCsr->pExpr, fts3SnippetFindPositions, (void *)&sIter);
+  if( rc==SQLITE_OK ){
 
-  /* Set the *pmSeen output variable. */
-  for(i=0; i<nList; i++){
-    if( sIter.aPhrase[i].pHead ){
-      *pmSeen |= (u64)1 << i;
+    /* Set the *pmSeen output variable. */
+    for(i=0; i<nList; i++){
+      if( sIter.aPhrase[i].pHead ){
+        *pmSeen |= (u64)1 << i;
+      }
     }
-  }
 
-  /* Loop through all candidate snippets. Store the best snippet in 
-  ** *pFragment. Store its associated 'score' in iBestScore.
-  */
-  pFragment->iCol = iCol;
-  while( !fts3SnippetNextCandidate(&sIter) ){
-    int iPos;
-    int iScore;
-    u64 mCover;
-    u64 mHighlight;
-    fts3SnippetDetails(&sIter, mCovered, &iPos, &iScore, &mCover, &mHighlight);
-    assert( iScore>=0 );
-    if( iScore>iBestScore ){
-      pFragment->iPos = iPos;
-      pFragment->hlmask = mHighlight;
-      pFragment->covered = mCover;
-      iBestScore = iScore;
+    /* Loop through all candidate snippets. Store the best snippet in 
+     ** *pFragment. Store its associated 'score' in iBestScore.
+     */
+    pFragment->iCol = iCol;
+    while( !fts3SnippetNextCandidate(&sIter) ){
+      int iPos;
+      int iScore;
+      u64 mCover;
+      u64 mHighlite;
+      fts3SnippetDetails(&sIter, mCovered, &iPos, &iScore, &mCover,&mHighlite);
+      assert( iScore>=0 );
+      if( iScore>iBestScore ){
+        pFragment->iPos = iPos;
+        pFragment->hlmask = mHighlite;
+        pFragment->covered = mCover;
+        iBestScore = iScore;
+      }
     }
-  }
 
+    *piScore = iBestScore;
+  }
   sqlite3_free(sIter.aPhrase);
-  *piScore = iBestScore;
-  return SQLITE_OK;
+  return rc;
 }
 
 
diff --git a/manifest b/manifest
index 26c720e3c2..37143105d6 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fix\sa\s(almost\salways\sharmless)\sread\spast\sthe\send\sof\sa\smemory\sallocation\nthat\scomes\sabout\sbecause\sthe\sExpr.pTab\sfield\sis\schecked\son\san\nEXPR_REDUCEDSIZE\sExpr\sobject\sbefore\schecking\sthe\sExpr.op\sfield\sto\nknow\sthat\sthe\sExpr.pTab\sfield\sis\smeaningless.
-D 2015-01-27T13:17:05.225
+C Improve\sthe\sperformance\sof\sfts3/4\squeries\sthat\suse\sthe\sOR\soperator\sand\sat\sleast\sone\sauxiliary\sfts\sfunction.
+D 2015-01-27T18:43:02.971
 F Makefile.arm-wince-mingw32ce-gcc d6df77f1f48d690bd73162294bbba7f59507c72f
 F Makefile.in 5407a688f4d77a05c18a8142be8ae5a2829dd610
 F Makefile.linux-gcc 91d710bdc4998cb015f39edf3cb314ec4f4d7e23
@@ -78,16 +78,16 @@ F ext/fts3/README.content fdc666a70d5257a64fee209f97cf89e0e6e32b51
 F ext/fts3/README.syntax a19711dc5458c20734b8e485e75fb1981ec2427a
 F ext/fts3/README.tokenizers e0a8b81383ea60d0334d274fadf305ea14a8c314
 F ext/fts3/README.txt 8c18f41574404623b76917b9da66fcb0ab38328d
-F ext/fts3/fts3.c 5c464816508e40feb3c61f1f5566551764698fc8
+F ext/fts3/fts3.c 845f20440dacac6e09f5c7735205609b9a86536b
 F ext/fts3/fts3.h 3a10a0af180d502cecc50df77b1b22df142817fe
-F ext/fts3/fts3Int.h 53d4eca1fb23eab00681fb028fb82eb5705c1e21
+F ext/fts3/fts3Int.h 394858c12a17740f7a1f6bd372c4606d4425a8d1
 F ext/fts3/fts3_aux.c 5c211e17a64885faeb16b9ba7772f9d5445c2365
 F ext/fts3/fts3_expr.c 40123785eaa3ebd4c45c9b23407cc44ac0c49905
 F ext/fts3/fts3_hash.c 29b986e43f4e9dd40110eafa377dc0d63c422c60
 F ext/fts3/fts3_hash.h 39cf6874dc239d6b4e30479b1975fe5b22a3caaf
 F ext/fts3/fts3_icu.c e319e108661147bcca8dd511cd562f33a1ba81b5
 F ext/fts3/fts3_porter.c 3565faf04b626cddf85f03825e86056a4562c009
-F ext/fts3/fts3_snippet.c 51beb5c1498176fd9caccaf1c75b55cb803a985a
+F ext/fts3/fts3_snippet.c 55c126e07158b2a705f52dee2cdc57208d3e999a
 F ext/fts3/fts3_term.c a521f75132f9a495bdca1bdd45949b3191c52763
 F ext/fts3/fts3_test.c 8a3a78c4458b2d7c631fcf4b152a5cd656fa7038
 F ext/fts3/fts3_tokenize_vtab.c becc661223db7898b213f9e8a23d75bac02408c9
@@ -1237,7 +1237,7 @@ F tool/vdbe_profile.tcl 67746953071a9f8f2f668b73fe899074e2c6d8c1
 F tool/warnings-clang.sh f6aa929dc20ef1f856af04a730772f59283631d4
 F tool/warnings.sh 0abfd78ceb09b7f7c27c688c8e3fe93268a13b32
 F tool/win/sqlite.vsix deb315d026cc8400325c5863eef847784a219a2f
-P 1964e656b4b420e8d6a4ba12d270ed02db292b88
-R 5d4aecd212970d14e41b3c7464003655
-U drh
-Z 469718f07e1956a0a1c83ab2938852ec
+P e098de691002a78270540430b0df1e120582b53f
+R 097e5fd012edfb2d64381b5476250a91
+U dan
+Z 72e3dfc5e16e519968e47157e86381d8
diff --git a/manifest.uuid b/manifest.uuid
index 488022b3f9..771d94784d 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-e098de691002a78270540430b0df1e120582b53f
\ No newline at end of file
+245e8730451fbdc1c729beff7295c452df604009
\ No newline at end of file