From: dan <dan@noemail.net>
Date: Mon, 27 Jun 2011 11:15:53 +0000 (+0000)
Subject: Changes to improve the selection of deferred tokens within phrases.
X-Git-Tag: version-3.7.8~38^2~35^2~4
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=5b217c308ffe74c1f9d12cd3664cccc0d5f4286b;p=thirdparty%2Fsqlite.git

Changes to improve the selection of deferred tokens within phrases.

FossilOrigin-Name: 2c4bbd90e2fca593c186bf412b608aff8c9f9061
---

diff --git a/ext/fts3/fts3.c b/ext/fts3/fts3.c
index 40a0297add..16067aacc7 100644
--- a/ext/fts3/fts3.c
+++ b/ext/fts3/fts3.c
@@ -2553,24 +2553,15 @@ static int fts3TermSelect(
 ** that the doclist is simply a list of docids stored as delta encoded 
 ** varints.
 */
-static int fts3DoclistCountDocids(int isPoslist, char *aList, int nList){
+static int fts3DoclistCountDocids(char *aList, int nList){
   int nDoc = 0;                   /* Return value */
   if( aList ){
     char *aEnd = &aList[nList];   /* Pointer to one byte after EOF */
     char *p = aList;              /* Cursor */
-    if( !isPoslist ){
-      /* The number of docids in the list is the same as the number of 
-      ** varints. In FTS3 a varint consists of a single byte with the 0x80 
-      ** bit cleared and zero or more bytes with the 0x80 bit set. So to
-      ** count the varints in the buffer, just count the number of bytes
-      ** with the 0x80 bit clear.  */
-      while( p<aEnd ) nDoc += (((*p++)&0x80)==0);
-    }else{
-      while( p<aEnd ){
-        nDoc++;
-        while( (*p++)&0x80 );     /* Skip docid varint */
-        fts3PoslistCopy(0, &p);   /* Skip over position list */
-      }
+    while( p<aEnd ){
+      nDoc++;
+      while( (*p++)&0x80 );     /* Skip docid varint */
+      fts3PoslistCopy(0, &p);   /* Skip over position list */
     }
   }
 
@@ -3914,27 +3905,35 @@ static int fts3EvalSelectDeferred(
     ** this (the logic that selects the tokens to be deferred) is probably
     ** the bit that needs to change.
     */
-    if( ii==0 ){
-      if( pTC->nOvfl ){
-        nDocEst = (pTC->nOvfl * pTab->nPgsz + pTab->nPgsz) / 10;
-      }else{
-        Fts3PhraseToken *pToken = pTC->pToken;
-        int nList = 0;
-        char *pList = 0;
-        rc = fts3TermSelect(pTab, pToken, pTC->iCol, &nList, &pList);
-        assert( rc==SQLITE_OK || pList==0 );
-        if( rc==SQLITE_OK ){
-          nDocEst = fts3DoclistCountDocids(1, pList, nList);
-          fts3EvalPhraseMergeToken(pTab, pTC->pPhrase, pTC->iToken,pList,nList);
-        }
-      }
-    }else{
-      if( pTC->nOvfl>=(nDocEst*nDocSize) ){
-        Fts3PhraseToken *pToken = pTC->pToken;
-        rc = sqlite3Fts3DeferToken(pCsr, pToken, pTC->iCol);
-        fts3SegReaderCursorFree(pToken->pSegcsr);
-        pToken->pSegcsr = 0;
+
+    if( ii && pTC->nOvfl>=(nDocEst*nDocSize) ){
+      /* The number of overflow pages to load for this (and therefore all
+      ** subsequent) tokens is greater than the estimated number of pages 
+      ** that will be loaded if all subsequent tokens are deferred.
+      */
+      Fts3PhraseToken *pToken = pTC->pToken;
+      rc = sqlite3Fts3DeferToken(pCsr, pToken, pTC->iCol);
+      fts3SegReaderCursorFree(pToken->pSegcsr);
+      pToken->pSegcsr = 0;
+    }else if( ii==0 || pTC->pPhrase->nToken>1 ){
+      /* Either this is the cheapest token in the entire query, or it is
+      ** part of a multi-token phrase. Either way, the entire doclist will
+      ** (eventually) be loaded into memory. It may as well be now. */
+      Fts3PhraseToken *pToken = pTC->pToken;
+      int nList = 0;
+      char *pList = 0;
+      rc = fts3TermSelect(pTab, pToken, pTC->iCol, &nList, &pList);
+      assert( rc==SQLITE_OK || pList==0 );
+      if( rc==SQLITE_OK ){
+        fts3EvalPhraseMergeToken(pTab, pTC->pPhrase, pTC->iToken,pList,nList);
+        nDocEst = fts3DoclistCountDocids(
+            pTC->pPhrase->doclist.aAll, pTC->pPhrase->doclist.nAll
+        );
       }
+    }else {
+      /* This token will not be deferred. And it will not be loaded into
+      ** memory at this point either. So assume that it filters out 75% of
+      ** the currently estimated number of documents. */
       nDocEst = 1 + (nDocEst/4);
     }
     pTC->pToken = 0;
diff --git a/manifest b/manifest
index 5f55a7ba07..c54d53e491 100644
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Fix\ssome\sof\sthe\scode\sissues\s(missing\scomments\setc.)\sin\sthe\snew\sFTS\scode.
-D 2011-06-23T17:09:51.936
+C Changes\sto\simprove\sthe\sselection\sof\sdeferred\stokens\swithin\sphrases.
+D 2011-06-27T11:15:53.752
 F Makefile.arm-wince-mingw32ce-gcc d6df77f1f48d690bd73162294bbba7f59507c72f
 F Makefile.in c1d7a7f4fd8da6b1815032efca950e3d5125407e
 F Makefile.linux-gcc 91d710bdc4998cb015f39edf3cb314ec4f4d7e23
@@ -62,7 +62,7 @@ F ext/fts2/mkfts2amal.tcl 974d5d438cb3f7c4a652639262f82418c1e4cff0
 F ext/fts3/README.syntax a19711dc5458c20734b8e485e75fb1981ec2427a
 F ext/fts3/README.tokenizers 998756696647400de63d5ba60e9655036cb966e9
 F ext/fts3/README.txt 8c18f41574404623b76917b9da66fcb0ab38328d
-F ext/fts3/fts3.c c25723c7e96763fc88652a8716564e4b22c8a327
+F ext/fts3/fts3.c 70127bf57c54c45545a4ddd73161caf2d9bd1abe
 F ext/fts3/fts3.h 3a10a0af180d502cecc50df77b1b22df142817fe
 F ext/fts3/fts3Int.h 7b452eeb592134c7adf87720c9e56d6060d1ed5a
 F ext/fts3/fts3_aux.c 0ebfa7b86cf8ff6a0861605fcc63b83ec1b70691
@@ -949,10 +949,7 @@ F tool/symbols.sh bc2a3709940d47c8ac8e0a1fdf17ec801f015a00
 F tool/tostr.awk 11760e1b94a5d3dcd42378f3cc18544c06cfa576
 F tool/vdbe-compress.tcl d70ea6d8a19e3571d7ab8c9b75cba86d1173ff0f
 F tool/warnings.sh 347d974d143cf132f953b565fbc03026f19fcb4d
-P c271f7e88fc081a460dd3f4afb24aa9fb7fa2917
-R b9f14b373682048157818c9e1116d00f
-T *branch * fts3-changes
-T *sym-fts3-changes *
-T -sym-trunk *
+P 8230d83120e0f4d217fde56e22c6f05aa5adee09
+R a3718a7f2e746e02b745366a3c6c0fd0
 U dan
-Z a7e048a2a0bdbbc293d8f131a4bc43fc
+Z deeb565ce480cc0dfb49de44342bb4c6
diff --git a/manifest.uuid b/manifest.uuid
index f18b078c6a..b01b761298 100644
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-8230d83120e0f4d217fde56e22c6f05aa5adee09
\ No newline at end of file
+2c4bbd90e2fca593c186bf412b608aff8c9f9061
\ No newline at end of file