Modify incremental merge code to merge nMin segments at a time.

FossilOrigin-Name: cd34bc1af4ba608ea3b52bab55bcfe0086711900

diff --git a/ext/fts3/fts3_write.c b/ext/fts3/fts3_write.c
index 08f07c10b2f58d4e5b1e3e95f5d1d047b0acaaa5..a6ee1f6919f0ac213c9ef681c4e7718e9153de0c 100644 (file)
--- a/ext/fts3/fts3_write.c
+++ b/ext/fts3/fts3_write.c
@@ -306,10 +306,10 @@ static int fts3SqlStmt(
          "  ORDER BY (level %% 1024) DESC LIMIT 1",
 
 /* Estimate the upper limit on the number of leaf nodes in a new segment
-** created by merging the two segments with idx=0 and idx=1 from absolute
-** level ?. See function fts3Incrmerge() for details.  */
+** created by merging the oldest :2 segments from absolute level :1. See 
+** function fts3Incrmerge() for details.  */
 /* 29 */ "SELECT 2 * total(1 + leaves_end_block - start_block) "
-         "  FROM %Q.'%q_segdir' WHERE level = ? AND idx BETWEEN 0 AND 1",
+         "  FROM %Q.'%q_segdir' WHERE level = ? AND idx < ?",
 
 /* SQL_DELETE_SEGDIR_ENTRY
 **   Delete the %_segdir entry on absolute level :1 with index :2.  */
@@ -1077,6 +1077,9 @@ static int fts3AllocateSegdirIdx(
     ** if iNext is less than FTS3_MERGE_COUNT, allocate index iNext.
     */
     if( iNext>=FTS3_MERGE_COUNT ){
+sqlite3_log(SQLITE_OK, 
+    "16-way merge at level=%d langid=%d index=%d", iLevel, iLangid, iIndex
+);
       rc = fts3SegmentMerge(p, iLangid, iIndex, iLevel);
       *piIdx = 0;
     }else{
@@ -3232,25 +3235,30 @@ static int fts3DoRebuild(Fts3Table *p){
 static int fts3IncrmergeCsr(
   Fts3Table *p,                   /* FTS3 table handle */
   sqlite3_int64 iAbsLevel,        /* Absolute level to open */
+  int nSeg,                       /* Number of segments to merge */
   Fts3MultiSegReader *pCsr        /* Cursor object to populate */
 ){
   int rc;                         /* Return Code */
-  sqlite3_stmt *pStmt = 0;
+  sqlite3_stmt *pStmt = 0;        /* Statement used to read %_segdir entry */  
+  int nByte;                      /* Bytes allocated at pCsr->apSegment[] */
 
+  assert( nSeg>=2 );
   memset(pCsr, 0, sizeof(*pCsr));
-  pCsr->apSegment = (Fts3SegReader **)sqlite3_malloc(sizeof(Fts3SegReader *)*2);
+  nByte = sizeof(Fts3SegReader *) * nSeg;
+
+  pCsr->apSegment = (Fts3SegReader **)sqlite3_malloc(nByte);
   if( pCsr->apSegment==0 ){
     rc = SQLITE_NOMEM;
   }else{
-    memset(pCsr->apSegment, 0, sizeof(Fts3SegReader *)*2);
-    pCsr->nSegment = 2;
+    memset(pCsr->apSegment, 0, nByte);
+    pCsr->nSegment = nSeg;
     rc = fts3SqlStmt(p, SQL_SELECT_LEVEL, &pStmt, 0);
   }
   if( rc==SQLITE_OK ){
     int i;
     int rc2;
     sqlite3_bind_int64(pStmt, 1, iAbsLevel);
-    for(i=0; rc==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW && i<2; i++){
+    for(i=0; rc==SQLITE_OK && sqlite3_step(pStmt)==SQLITE_ROW && i<nSeg; i++){
       rc = sqlite3Fts3SegReaderNew(i, 0,
           sqlite3_column_int64(pStmt, 1),        /* segdir.start_block */
           sqlite3_column_int64(pStmt, 2),        /* segdir.leaves_end_block */
@@ -3818,8 +3826,7 @@ static int fts3IncrmergeLoad(
 static int fts3IncrmergeWriter( 
   Fts3Table *p,                   /* Fts3 table handle */
   sqlite3_int64 iAbsLevel,        /* Absolute level of input segments */
-  const char *zKey,               /* First key to write */
-  int nKey,                       /* Number of bytes in nKey */
+  Fts3MultiSegReader *pCsr,       /* Cursor that data will be read from */
   IncrmergeWriter *pWriter        /* Populate this object */
 ){
   int rc;                         /* Return Code */
@@ -3829,6 +3836,8 @@ static int fts3IncrmergeWriter(
   sqlite3_stmt *pLeafEst = 0;     /* SQL used to determine nLeafEst */
   sqlite3_stmt *pFirstBlock = 0;  /* SQL used to determine first block */
   sqlite3_stmt *pOutputIdx = 0;   /* SQL used to find output index */
+  const char *zKey = pCsr->zTerm;
+  int nKey = pCsr->nTerm;
 
   rc = fts3SqlStmt(p, SQL_NEXT_SEGMENT_INDEX, &pOutputIdx, 0);
   if( rc==SQLITE_OK ){
@@ -3849,6 +3858,7 @@ static int fts3IncrmergeWriter(
   rc = fts3SqlStmt(p, SQL_MAX_LEAF_NODE_ESTIMATE, &pLeafEst, 0);
   if( rc==SQLITE_OK ){
     sqlite3_bind_int64(pLeafEst, 1, iAbsLevel);
+    sqlite3_bind_int64(pLeafEst, 2, pCsr->nSegment);
     if( SQLITE_ROW==sqlite3_step(pLeafEst) ){
       nLeafEst = sqlite3_column_int(pLeafEst, 0);
     }
@@ -4120,15 +4130,17 @@ static int fts3IncrmergeChomp(
   int i;
   int rc = SQLITE_OK;
 
-  assert( pCsr->nSegment==2 );
-  assert( (pCsr->apSegment[0]->iIdx==0 && pCsr->apSegment[1]->iIdx==1)
-       || (pCsr->apSegment[1]->iIdx==0 && pCsr->apSegment[0]->iIdx==1) 
-  );
+  for(i=pCsr->nSegment-1; i>=0 && rc==SQLITE_OK; i--){
+    Fts3SegReader *pSeg = 0;
+    int j;
 
-  for(i=1; i>=0 && rc==SQLITE_OK; i--){
-    Fts3SegReader *pSeg = pCsr->apSegment[0];
-    if( pSeg->iIdx!=i ) pSeg = pCsr->apSegment[1];
-    assert( pSeg->iIdx==i );
+    /* Find the Fts3SegReader object with Fts3SegReader.iIdx==i. It is hiding
+    ** somewhere in the pCsr->apSegment[] array.  */
+    for(j=0; ALWAYS(j<pCsr->nSegment); j++){
+      pSeg = pCsr->apSegment[j];
+      if( pSeg->iIdx==i ) break;
+    }
+    assert( j<pCsr->nSegment && pSeg->iIdx==i );
 
     if( pSeg->aNode==0 ){
       /* Seg-reader is at EOF. Remove the entire input segment. */
@@ -4178,6 +4190,8 @@ static int fts3Incrmerge(Fts3Table *p, int nMerge, int nMin){
     rc = fts3SqlStmt(p, SQL_FIND_MERGE_LEVEL, &pFindLevel, 0);
     sqlite3_bind_int(pFindLevel, 1, nMin);
     if( sqlite3_step(pFindLevel)!=SQLITE_ROW ){
+      /* There are no levels with nMin or more segments. Or an error has
+      ** occurred. Either way, exit early.  */
       return sqlite3_reset(pFindLevel);
     }
     iAbsLevel = sqlite3_column_int64(pFindLevel, 0);
@@ -4193,14 +4207,16 @@ static int fts3Incrmerge(Fts3Table *p, int nMerge, int nMin){
     /* Open a cursor to iterate through the contents of indexes 0 and 1 of
     ** the selected absolute level. */
     pFilter->flags = FTS3_SEGMENT_REQUIRE_POS;
-    rc = fts3IncrmergeCsr(p, iAbsLevel, pCsr);
-
+    rc = fts3IncrmergeCsr(p, iAbsLevel, nMin, pCsr);
+sqlite3_log(SQLITE_OK, "%d-way merge from level=%d to level=%d", 
+    nMin, (int)iAbsLevel, (int)iAbsLevel+1
+);
     if( rc==SQLITE_OK ){
       rc = sqlite3Fts3SegReaderStart(p, pCsr, pFilter);
     }
     if( rc==SQLITE_OK ){
       if( SQLITE_ROW==(rc = sqlite3Fts3SegReaderStep(p, pCsr)) ){
-        rc = fts3IncrmergeWriter(p, iAbsLevel, pCsr->zTerm,pCsr->nTerm,pWriter);
+        rc = fts3IncrmergeWriter(p, iAbsLevel, pCsr, pWriter);
         if( rc==SQLITE_OK ){
           do {
             rc = fts3IncrmergeAppend(p, pWriter, pCsr);

diff --git a/manifest b/manifest
index 06d7ae13cee7c829100cac68ed631df55a7e5809..88b357ba300270173be89af09afacd857bc4ca5d 100644 (file)
--- a/manifest
+++ b/manifest
@@ -1,5 +1,5 @@
-C Add\stests\sfor\sincremental\smerge\scode.
-D 2012-03-14T20:01:52.250
+C Modify\sincremental\smerge\scode\sto\smerge\snMin\ssegments\sat\sa\stime.
+D 2012-03-15T17:45:50.857
 F Makefile.arm-wince-mingw32ce-gcc d6df77f1f48d690bd73162294bbba7f59507c72f
 F Makefile.in 3f79a373e57c3b92dabf76f40b065e719d31ac34
 F Makefile.linux-gcc 91d710bdc4998cb015f39edf3cb314ec4f4d7e23
@@ -78,7 +78,7 @@ F ext/fts3/fts3_test.c 6b7cc68aef4efb084e1449f7d20c4b20d3bdf6b4
 F ext/fts3/fts3_tokenizer.c 3da7254a9881f7e270ab28e2004e0d22b3212bce
 F ext/fts3/fts3_tokenizer.h 66dec98e365854b6cd2d54f1a96bb6d428fc5a68
 F ext/fts3/fts3_tokenizer1.c 0dde8f307b8045565cf63797ba9acfaff1c50c68
-F ext/fts3/fts3_write.c 355121666034ece7d2472caa3f322d0cb61b0f0d
+F ext/fts3/fts3_write.c be60aaf4a887a68aaf0d7541bc78413e7ba66291
 F ext/fts3/fts3speed.tcl b54caf6a18d38174f1a6e84219950d85e98bb1e9
 F ext/fts3/mkfts3amal.tcl 252ecb7fe6467854f2aa237bf2c390b74e71f100
 F ext/icu/README.txt bf8461d8cdc6b8f514c080e4e10dc3b2bbdfefa9
@@ -448,7 +448,7 @@ F test/fts2q.test b2fbbe038b7a31a52a6079b215e71226d8c6a682
 F test/fts2r.test b154c30b63061d8725e320fba1a39e2201cadd5e
 F test/fts2token.test d8070b241a15ff13592a9ae4a8b7c171af6f445a
 F test/fts3.test 672a040ea57036fb4b6fdc09027c18d7d24ab654
-F test/fts3_common.tcl 1b2e522476236a018c1a4b8ae4f7599ce29b7be0
+F test/fts3_common.tcl d65de7e21c2b933bf17152830eb924356f197c8f
 F test/fts3aa.test 909d5f530d30a8e36b9328d67285eae6537c79c0
 F test/fts3ab.test 09aeaa162aee6513d9ff336b6932211008b9d1f9
 F test/fts3ac.test 636ed7486043055d4f126a0e385f2d5a82ebbf63
@@ -497,7 +497,7 @@ F test/fts3sort.test 95be0b19d7e41c44b29014f13ea8bddd495fd659
 F test/fts4aa.test 6e7f90420b837b2c685f3bcbe84c868492d40a68
 F test/fts4content.test 17b2360f7d1a9a7e5aa8022783f5c5731b6dfd4f
 F test/fts4langid.test fabdd5a8db0fa00292e0704809f566e3fb6dba3a
-F test/fts4merge.test d841f283bf9557d549afde513fe38d0b48adc4f7
+F test/fts4merge.test 1c3389a3be18498934baf76afe67663d8b4b5e42
 F test/fts4merge2.test 5faa558d1b672f82b847d2a337465fa745e46891
 F test/func.test 6c5ce11e3a0021ca3c0649234e2d4454c89110ca
 F test/func2.test 772d66227e4e6684b86053302e2d74a2500e1e0f
@@ -994,7 +994,7 @@ F tool/tostr.awk e75472c2f98dd76e06b8c9c1367f4ab07e122d06
 F tool/vdbe-compress.tcl d70ea6d8a19e3571d7ab8c9b75cba86d1173ff0f
 F tool/warnings-clang.sh 9f406d66e750e8ac031c63a9ef3248aaa347ef2a
 F tool/warnings.sh fbc018d67fd7395f440c28f33ef0f94420226381
-P 36ae510de45be44efd34cff242d02fb21b7419ac
-R 326f29ae3378aab5424c0bd2deaa1142
+P 570473729d6561d81e6e5f8884fd18487008636e
+R 8deb5f38fb61fdd986c036580c00c7d8
 U dan
-Z 9fe983241614f7794efd59c0386308c4
+Z c18f3e75d81727125e658ae77334fefb

diff --git a/manifest.uuid b/manifest.uuid
index 9cec3849f5321cbd99170febe0c6df1a1e564f71..08619b20e53c70478e145bc977954dd70de15fa1 100644 (file)
--- a/manifest.uuid
+++ b/manifest.uuid
@@ -1 +1 @@
-570473729d6561d81e6e5f8884fd18487008636e
\ No newline at end of file
+cd34bc1af4ba608ea3b52bab55bcfe0086711900
\ No newline at end of file

diff --git a/test/fts3_common.tcl b/test/fts3_common.tcl
index 6c9c2fcec9452bbe360ddbc51eaadc3b8805125b..f031f927d65c7d17b090f77668439cc26e3aa518 100644 (file)
--- a/test/fts3_common.tcl
+++ b/test/fts3_common.tcl
@@ -66,10 +66,10 @@ proc fts3_build_db_2 {n} {
 
   for {set i 0} {$i < $n} {incr i} {
     set word ""
-    set n [llength $chars]
-    append word [lindex $chars [expr {($i / 1)   % $n}]]
-    append word [lindex $chars [expr {($i / $n)  % $n}]]
-    append word [lindex $chars [expr {($i / ($n*$n)) % $n}]]
+    set nChar [llength $chars]
+    append word [lindex $chars [expr {($i / 1)   % $nChar}]]
+    append word [lindex $chars [expr {($i / $nChar)  % $nChar}]]
+    append word [lindex $chars [expr {($i / ($nChar*$nChar)) % $nChar}]]
 
     db eval { INSERT INTO t2(docid, content) VALUES($i, $word) }
   }

diff --git a/test/fts4merge.test b/test/fts4merge.test
index a5995921f34133309843d8942177544deb90cee1..9f7e05b0ed5fb206945aa3fcb7d95e8f5d66e8d0 100644 (file)
--- a/test/fts4merge.test
+++ b/test/fts4merge.test
@@ -47,9 +47,9 @@ for {set i 0} {$i<20} {incr i} {
 do_execsql_test 1.3 { 
   SELECT level, group_concat(idx, ' ') FROM t1_segdir GROUP BY level 
 } {
-  0 {0 1 2 3 4 5 6 7} 
-  1 {0 1 2 3 4 5 6 7}
-  2 {0 1 2 3 4 5 6}
+  0 {0 1 2 3} 
+  1 {0 1 2 3 4 5 6} 
+  2 {0 1 2 3}
 }
 
 for {set i 0} {$i<100} {incr i} {
@@ -64,9 +64,9 @@ do_execsql_test 1.5 {
   SELECT level, group_concat(idx, ' ') FROM t1_segdir GROUP BY level 
 } {
   0 {0 1 2 3} 
-  1 {0 1 2 3} 
-  2 {0 1 2 3} 
-  3 {0 1 2}
+  1 {0 1 2} 
+  2 0
+  3 0
 }
 
 #-------------------------------------------------------------------------
@@ -103,7 +103,7 @@ do_test 3.1 { fts3_integrity_check t2 } {ok}
 do_execsql_test 3.2 { 
   SELECT level, group_concat(idx, ' ') FROM t2_segdir GROUP BY level 
 } {
-  0 {0 1 2 3 4 5 6 7} 
+  0 {0 1 2 3 4 5 6} 
   1 {0 1 2 3 4} 
   2 {0 1 2 3 4} 
   3 {0 1 2 3 4 5 6}
@@ -113,10 +113,10 @@ do_execsql_test 3.3 {
   INSERT INTO t2(t2) VALUES('merge=1000000,2');
   SELECT level, group_concat(idx, ' ') FROM t2_segdir GROUP BY level 
 } {
-  0 {0 1} 
+  0 0 
   1 {0 1} 
   2 0 
-  3 {0 1}  
+  3 {0 1} 
   4 {0 1} 
   5 0
 }