Use a Bloom filter to improve performance of IN operators when the RHS of

the IN operator is a subquery.

FossilOrigin-Name: 1933496539c19cbf429a39d6b0b1c6b1b2af50733a3c4aea4920990ced652f6a

3 files changed, 39 insertions, 2 deletions

diff --git a/src/expr.c b/src/expr.c
index 6d2b6d133..6d67d77f6 100644
--- a/src/expr.c
+++ b/src/expr.c
@@ -3536,19 +3536,34 @@ void sqlite3CodeRhsOfIN(
       SelectDest dest;
       int i;
       int rc;
+      int addrBloom = 0;
       sqlite3SelectDestInit(&dest, SRT_Set, iTab);
       dest.zAffSdst = exprINAffinity(pParse, pExpr);
       pSelect->iLimit = 0;
+      if( addrOnce && OptimizationEnabled(pParse->db, SQLITE_BloomFilter) ){
+        int regBloom = ++pParse->nMem;
+        addrBloom = sqlite3VdbeAddOp2(v, OP_Blob, 10000, regBloom);
+        VdbeComment((v, "Bloom filter"));
+        dest.iSDParm2 = regBloom;
+      }
       testcase( pSelect->selFlags & SF_Distinct );
       testcase( pKeyInfo==0 ); /* Caused by OOM in sqlite3KeyInfoAlloc() */
       pCopy = sqlite3SelectDup(pParse->db, pSelect, 0);
       rc = pParse->db->mallocFailed ? 1 :sqlite3Select(pParse, pCopy, &dest);
       sqlite3SelectDelete(pParse->db, pCopy);
       sqlite3DbFree(pParse->db, dest.zAffSdst);
+      if( addrBloom ){
+        sqlite3VdbeGetOp(v, addrOnce)->p3 = dest.iSDParm2;
+        if( dest.iSDParm2==0 ){
+          sqlite3VdbeChangeToNoop(v, addrBloom);
+        }else{
+          sqlite3VdbeGetOp(v, addrOnce)->p3 = dest.iSDParm2;
+        }
+      }
       if( rc ){
         sqlite3KeyInfoUnref(pKeyInfo);
         return;
-      }     
+      }
       assert( pKeyInfo!=0 ); /* OOM will cause exit after sqlite3Select() */
       assert( pEList!=0 );
       assert( pEList->nExpr>0 );
@@ -3987,6 +4002,15 @@ static void sqlite3ExprCodeIN(
     sqlite3VdbeAddOp4(v, OP_Affinity, rLhs, nVector, 0, zAff, nVector);
     if( destIfFalse==destIfNull ){
       /* Combine Step 3 and Step 5 into a single opcode */
+      if( ExprHasProperty(pExpr, EP_Subrtn) ){
+        const VdbeOp *pOp = sqlite3VdbeGetOp(v, pExpr->y.sub.iAddr);
+        assert( pOp->opcode==OP_Once || pParse->nErr );
+        if( pOp->opcode==OP_Once && pOp->p3>0 ){
+          assert( OptimizationEnabled(pParse->db, SQLITE_BloomFilter) );
+          sqlite3VdbeAddOp4Int(v, OP_Filter, pOp->p3, destIfFalse,
+                               rLhs, nVector); VdbeCoverage(v);
+        }
+      }
       sqlite3VdbeAddOp4Int(v, OP_NotFound, iTab, destIfFalse,
                            rLhs, nVector); VdbeCoverage(v);
       goto sqlite3ExprCodeIN_finished;
diff --git a/src/select.c b/src/select.c
index b43861cef..a1da3f51b 100644
--- a/src/select.c
+++ b/src/select.c
@@ -1377,12 +1377,17 @@ static void selectInnerLoop(
         ** case the order does matter */
         pushOntoSorter(
             pParse, pSort, p, regResult, regOrig, nResultCol, nPrefixReg);
+        pDest->iSDParm2 = 0; /* Signal that any Bloom filter is unpopulated */
       }else{
         int r1 = sqlite3GetTempReg(pParse);
         assert( sqlite3Strlen30(pDest->zAffSdst)==nResultCol );
         sqlite3VdbeAddOp4(v, OP_MakeRecord, regResult, nResultCol,
             r1, pDest->zAffSdst, nResultCol);
         sqlite3VdbeAddOp4Int(v, OP_IdxInsert, iParm, r1, regResult, nResultCol);
+        if( pDest->iSDParm2 ){
+          sqlite3VdbeAddOp4Int(v, OP_FilterAdd, pDest->iSDParm2, 0,
+                               regResult, nResultCol);
+        }
         sqlite3ReleaseTempReg(pParse, r1);
       }
       break;
@@ -3316,6 +3321,10 @@ static int generateOutputSubroutine(
           r1, pDest->zAffSdst, pIn->nSdst);
       sqlite3VdbeAddOp4Int(v, OP_IdxInsert, pDest->iSDParm, r1,
                            pIn->iSdst, pIn->nSdst);
+      if( pDest->iSDParm2>0 ){
+        sqlite3VdbeAddOp4Int(v, OP_FilterAdd, pDest->iSDParm2, 0,
+                             pIn->iSdst, pIn->nSdst);
+      }
       sqlite3ReleaseTempReg(pParse, r1);
       break;
     }
diff --git a/src/sqliteInt.h b/src/sqliteInt.h
index fd9faf6b4..df6774aef 100644
--- a/src/sqliteInt.h
+++ b/src/sqliteInt.h
@@ -3622,7 +3622,11 @@ struct Select {
 **     SRT_Set         The result must be a single column.  Store each
 **                     row of result as the key in table pDest->iSDParm.
 **                     Apply the affinity pDest->affSdst before storing
-**                     results.  Used to implement "IN (SELECT ...)".
+**                     results.  if pDest->iSDParm2 is positive, then it is
+**                     a regsiter holding a Bloom filter for the IN operator
+**                     that should be populated in addition to the 
+**                     pDest->iSDParm table.  This SRT is used to
+**                     implement "IN (SELECT ...)".
 **
 **     SRT_EphemTab    Create an temporary table pDest->iSDParm and store
 **                     the result there. The cursor is left open after