/ Check-in [c71f2359]
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Merge recent trunk changes with this branch.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | exp-window-functions
Files: files | file ages | folders
SHA3-256: c71f23590c25b4cecd27722e6c0fc8e3bf320d399c7d9398b7016dd5cf5b05eb
User & Date: dan 2018-06-09 18:09:44
Context
2018-06-10
07:42
Update Makefile.msc to include window.c in the build. check-in: 16db7384 user: dan tags: exp-window-functions
2018-06-09
18:09
Merge recent trunk changes with this branch. check-in: c71f2359 user: dan tags: exp-window-functions
17:58
Update the amalgamation build script to include window.c. check-in: 21d2f4a6 user: dan tags: exp-window-functions
16:49
Slightly smaller and faster code by encapsulating wal-index hash table location information in a separate WalHashLoc object rather than passing around the various elements as separate variables. check-in: 538a365b user: drh tags: trunk
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to src/btree.c.

5592
5593
5594
5595
5596
5597
5598
5599









5600
5601
5602
5603
5604
5605
5606
      }
      pCur->skipNext = 0;
    }
  }

  pPage = pCur->pPage;
  idx = ++pCur->ix;
  assert( pPage->isInit );










  /* If the database file is corrupt, it is possible for the value of idx 
  ** to be invalid here. This can only occur if a second cursor modifies
  ** the page while cursor pCur is holding a reference to it. Which can
  ** only happen if the database is corrupt in such a way as to link the
  ** page into more than one b-tree structure. */
  testcase( idx>pPage->nCell );







|
>
>
>
>
>
>
>
>
>







5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
      }
      pCur->skipNext = 0;
    }
  }

  pPage = pCur->pPage;
  idx = ++pCur->ix;
  if( !pPage->isInit ){
    /* The only known way for this to happen is for there to be a
    ** recursive SQL function that does a DELETE operation as part of a
    ** SELECT which deletes content out from under an active cursor
    ** in a corrupt database file where the table being DELETE-ed from
    ** has pages in common with the table being queried.  See TH3
    ** module cov1/btree78.test testcase 220 (2018-06-08) for an
    ** example. */
    return SQLITE_CORRUPT_BKPT;
  }

  /* If the database file is corrupt, it is possible for the value of idx 
  ** to be invalid here. This can only occur if a second cursor modifies
  ** the page while cursor pCur is holding a reference to it. Which can
  ** only happen if the database is corrupt in such a way as to link the
  ** page into more than one b-tree structure. */
  testcase( idx>pPage->nCell );

Changes to src/build.c.

1690
1691
1692
1693
1694
1695
1696

























1697
1698
1699
1700
1701
1702
1703
....
1839
1840
1841
1842
1843
1844
1845

1846
1847
1848
1849
1850
1851
1852
....
3272
3273
3274
3275
3276
3277
3278

3279
3280
3281
3282
3283
3284
3285

/* Return true if value x is found any of the first nCol entries of aiCol[]
*/
static int hasColumn(const i16 *aiCol, int nCol, int x){
  while( nCol-- > 0 ) if( x==*(aiCol++) ) return 1;
  return 0;
}


























/*
** This routine runs at the end of parsing a CREATE TABLE statement that
** has a WITHOUT ROWID clause.  The job of this routine is to convert both
** internal schema data structures and the generated VDBE code so that they
** are appropriate for a WITHOUT ROWID table instead of a rowid table.
** Changes include:
................................................................................
      }
    }
    assert( pPk->nColumn==j );
    assert( pTab->nCol==j );
  }else{
    pPk->nColumn = pTab->nCol;
  }

}

/*
** This routine is called to report the final ")" that terminates
** a CREATE TABLE statement.
**
** The table structure that other action routines have been building
................................................................................
  sqlite3DefaultRowEst(pIndex);
  if( pParse->pNewTable==0 ) estimateIndexWidth(pIndex);

  /* If this index contains every column of its table, then mark
  ** it as a covering index */
  assert( HasRowid(pTab) 
      || pTab->iPKey<0 || sqlite3ColumnOfIndex(pIndex, pTab->iPKey)>=0 );

  if( pTblName!=0 && pIndex->nColumn>=pTab->nCol ){
    pIndex->isCovering = 1;
    for(j=0; j<pTab->nCol; j++){
      if( j==pTab->iPKey ) continue;
      if( sqlite3ColumnOfIndex(pIndex,j)>=0 ) continue;
      pIndex->isCovering = 0;
      break;







>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>







 







>







 







>







1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
....
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
....
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312

/* Return true if value x is found any of the first nCol entries of aiCol[]
*/
static int hasColumn(const i16 *aiCol, int nCol, int x){
  while( nCol-- > 0 ) if( x==*(aiCol++) ) return 1;
  return 0;
}

/* Recompute the colNotIdxed field of the Index.
**
** colNotIdxed is a bitmask that has a 0 bit representing each indexed
** columns that are within the first 63 columns of the table.  The
** high-order bit of colNotIdxed is always 1.  All unindexed columns
** of the table have a 1.
**
** The colNotIdxed mask is AND-ed with the SrcList.a[].colUsed mask
** to determine if the index is covering index.
*/
static void recomputeColumnsNotIndexed(Index *pIdx){
  Bitmask m = 0;
  int j;
  for(j=pIdx->nColumn-1; j>=0; j--){
    int x = pIdx->aiColumn[j];
    if( x>=0 ){
      testcase( x==BMS-1 );
      testcase( x==BMS-2 );
      if( x<BMS-1 ) m |= MASKBIT(x);
    }
  }
  pIdx->colNotIdxed = ~m;
  assert( (pIdx->colNotIdxed>>63)==1 );
}

/*
** This routine runs at the end of parsing a CREATE TABLE statement that
** has a WITHOUT ROWID clause.  The job of this routine is to convert both
** internal schema data structures and the generated VDBE code so that they
** are appropriate for a WITHOUT ROWID table instead of a rowid table.
** Changes include:
................................................................................
      }
    }
    assert( pPk->nColumn==j );
    assert( pTab->nCol==j );
  }else{
    pPk->nColumn = pTab->nCol;
  }
  recomputeColumnsNotIndexed(pPk);
}

/*
** This routine is called to report the final ")" that terminates
** a CREATE TABLE statement.
**
** The table structure that other action routines have been building
................................................................................
  sqlite3DefaultRowEst(pIndex);
  if( pParse->pNewTable==0 ) estimateIndexWidth(pIndex);

  /* If this index contains every column of its table, then mark
  ** it as a covering index */
  assert( HasRowid(pTab) 
      || pTab->iPKey<0 || sqlite3ColumnOfIndex(pIndex, pTab->iPKey)>=0 );
  recomputeColumnsNotIndexed(pIndex);
  if( pTblName!=0 && pIndex->nColumn>=pTab->nCol ){
    pIndex->isCovering = 1;
    for(j=0; j<pTab->nCol; j++){
      if( j==pTab->iPKey ) continue;
      if( sqlite3ColumnOfIndex(pIndex,j)>=0 ) continue;
      pIndex->isCovering = 0;
      break;

Changes to src/sqliteInt.h.

1106
1107
1108
1109
1110
1111
1112


























1113
1114
1115
1116
1117
1118
1119
....
2209
2210
2211
2212
2213
2214
2215

2216
2217
2218
2219
2220
2221
2222
....
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
typedef struct VTable VTable;
typedef struct VtabCtx VtabCtx;
typedef struct Walker Walker;
typedef struct WhereInfo WhereInfo;
typedef struct Window Window;
typedef struct With With;



























/* A VList object records a mapping between parameters/variables/wildcards
** in the SQL statement (such as $abc, @pqr, or :xyz) and the integer
** variable number associated with that parameter.  See the format description
** on the sqlite3VListAdd() routine for more information.  A VList is really
** just an array of integers.
*/
typedef int VList;
................................................................................
  int nSample;             /* Number of elements in aSample[] */
  int nSampleCol;          /* Size of IndexSample.anEq[] and so on */
  tRowcnt *aAvgEq;         /* Average nEq values for keys not in aSample */
  IndexSample *aSample;    /* Samples of the left-most key */
  tRowcnt *aiRowEst;       /* Non-logarithmic stat1 data for this index */
  tRowcnt nRowEst0;        /* Non-logarithmic number of rows in the index */
#endif

};

/*
** Allowed values for Index.idxType
*/
#define SQLITE_IDXTYPE_APPDEF      0   /* Created using CREATE INDEX */
#define SQLITE_IDXTYPE_UNIQUE      1   /* Implements a UNIQUE constraint */
................................................................................
  struct IdList_item {
    char *zName;      /* Name of the identifier */
    int idx;          /* Index in some Table.aCol[] of a column named zName */
  } *a;
  int nId;         /* Number of identifiers on the list */
};

/*
** The bitmask datatype defined below is used for various optimizations.
**
** Changing this from a 64-bit to a 32-bit type limits the number of
** tables in a join to 32 instead of 64.  But it also reduces the size
** of the library by 738 bytes on ix86.
*/
#ifdef SQLITE_BITMASK_TYPE
  typedef SQLITE_BITMASK_TYPE Bitmask;
#else
  typedef u64 Bitmask;
#endif

/*
** The number of bits in a Bitmask.  "BMS" means "BitMask Size".
*/
#define BMS  ((int)(sizeof(Bitmask)*8))

/*
** A bit in a Bitmask
*/
#define MASKBIT(n)   (((Bitmask)1)<<(n))
#define MASKBIT32(n) (((unsigned int)1)<<(n))
#define ALLBITS      ((Bitmask)-1)

/*
** The following structure describes the FROM clause of a SELECT statement.
** Each table or subquery in the FROM clause is a separate element of
** the SrcList.a[] array.
**
** With the addition of multiple database support, the following structure
** can also be used to describe a particular table such as the table that







>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>







 







>







 







<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
....
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
....
2582
2583
2584
2585
2586
2587
2588

























2589
2590
2591
2592
2593
2594
2595
typedef struct VTable VTable;
typedef struct VtabCtx VtabCtx;
typedef struct Walker Walker;
typedef struct WhereInfo WhereInfo;
typedef struct Window Window;
typedef struct With With;


/*
** The bitmask datatype defined below is used for various optimizations.
**
** Changing this from a 64-bit to a 32-bit type limits the number of
** tables in a join to 32 instead of 64.  But it also reduces the size
** of the library by 738 bytes on ix86.
*/
#ifdef SQLITE_BITMASK_TYPE
  typedef SQLITE_BITMASK_TYPE Bitmask;
#else
  typedef u64 Bitmask;
#endif

/*
** The number of bits in a Bitmask.  "BMS" means "BitMask Size".
*/
#define BMS  ((int)(sizeof(Bitmask)*8))

/*
** A bit in a Bitmask
*/
#define MASKBIT(n)   (((Bitmask)1)<<(n))
#define MASKBIT32(n) (((unsigned int)1)<<(n))
#define ALLBITS      ((Bitmask)-1)

/* A VList object records a mapping between parameters/variables/wildcards
** in the SQL statement (such as $abc, @pqr, or :xyz) and the integer
** variable number associated with that parameter.  See the format description
** on the sqlite3VListAdd() routine for more information.  A VList is really
** just an array of integers.
*/
typedef int VList;
................................................................................
  int nSample;             /* Number of elements in aSample[] */
  int nSampleCol;          /* Size of IndexSample.anEq[] and so on */
  tRowcnt *aAvgEq;         /* Average nEq values for keys not in aSample */
  IndexSample *aSample;    /* Samples of the left-most key */
  tRowcnt *aiRowEst;       /* Non-logarithmic stat1 data for this index */
  tRowcnt nRowEst0;        /* Non-logarithmic number of rows in the index */
#endif
  Bitmask colNotIdxed;     /* 0 for unindexed columns in pTab */
};

/*
** Allowed values for Index.idxType
*/
#define SQLITE_IDXTYPE_APPDEF      0   /* Created using CREATE INDEX */
#define SQLITE_IDXTYPE_UNIQUE      1   /* Implements a UNIQUE constraint */
................................................................................
  struct IdList_item {
    char *zName;      /* Name of the identifier */
    int idx;          /* Index in some Table.aCol[] of a column named zName */
  } *a;
  int nId;         /* Number of identifiers on the list */
};


























/*
** The following structure describes the FROM clause of a SELECT statement.
** Each table or subquery in the FROM clause is a separate element of
** the SrcList.a[] array.
**
** With the addition of multiple database support, the following structure
** can also be used to describe a particular table such as the table that

Changes to src/wal.c.

875
876
877
878
879
880
881












882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924

925
926
927
928
929
930
931
...
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
...
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054

1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095


1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
....
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
....
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
....
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900

2901
2902
2903
2904
2905
2906
2907
  assert( iPage>0 );
  assert( (HASHTABLE_NSLOT & (HASHTABLE_NSLOT-1))==0 );
  return (iPage*HASHTABLE_HASH_1) & (HASHTABLE_NSLOT-1);
}
static int walNextHash(int iPriorHash){
  return (iPriorHash+1)&(HASHTABLE_NSLOT-1);
}













/* 
** Return pointers to the hash table and page number array stored on
** page iHash of the wal-index. The wal-index is broken into 32KB pages
** numbered starting from 0.
**
** Set output variable *paHash to point to the start of the hash table
** in the wal-index file. Set *piZero to one less than the frame 
** number of the first frame indexed by this hash table. If a
** slot in the hash table is set to N, it refers to frame number 
** (*piZero+N) in the log.
**
** Finally, set *paPgno so that *paPgno[1] is the page number of the
** first frame indexed by the hash table, frame (*piZero+1).
*/
static int walHashGet(
  Wal *pWal,                      /* WAL handle */
  int iHash,                      /* Find the iHash'th table */
  volatile ht_slot **paHash,      /* OUT: Pointer to hash index */
  volatile u32 **paPgno,          /* OUT: Pointer to page number array */
  u32 *piZero                     /* OUT: Frame associated with *paPgno[0] */
){
  int rc;                         /* Return code */
  volatile u32 *aPgno;

  rc = walIndexPage(pWal, iHash, &aPgno);
  assert( rc==SQLITE_OK || iHash>0 );

  if( rc==SQLITE_OK ){
    u32 iZero;
    volatile ht_slot *aHash;

    aHash = (volatile ht_slot *)&aPgno[HASHTABLE_NPAGE];
    if( iHash==0 ){
      aPgno = &aPgno[WALINDEX_HDR_SIZE/sizeof(u32)];
      iZero = 0;
    }else{
      iZero = HASHTABLE_NPAGE_ONE + (iHash-1)*HASHTABLE_NPAGE;
    }
  
    *paPgno = &aPgno[-1];
    *paHash = aHash;
    *piZero = iZero;

  }
  return rc;
}

/*
** Return the number of the wal-index page that contains the hash-table
** and page-number array that contain entries corresponding to WAL frame
................................................................................
**
** At most only the hash table containing pWal->hdr.mxFrame needs to be
** updated.  Any later hash tables will be automatically cleared when
** pWal->hdr.mxFrame advances to the point where those hash tables are
** actually needed.
*/
static void walCleanupHash(Wal *pWal){
  volatile ht_slot *aHash = 0;    /* Pointer to hash table to clear */
  volatile u32 *aPgno = 0;        /* Page number array for hash table */
  u32 iZero = 0;                  /* frame == (aHash[x]+iZero) */
  int iLimit = 0;                 /* Zero values greater than this */
  int nByte;                      /* Number of bytes to zero in aPgno[] */
  int i;                          /* Used to iterate through aHash[] */

  assert( pWal->writeLock );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE-1 );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE );
................................................................................

  /* Obtain pointers to the hash-table and page-number array containing 
  ** the entry that corresponds to frame pWal->hdr.mxFrame. It is guaranteed
  ** that the page said hash-table and array reside on is already mapped.
  */
  assert( pWal->nWiData>walFramePage(pWal->hdr.mxFrame) );
  assert( pWal->apWiData[walFramePage(pWal->hdr.mxFrame)] );
  walHashGet(pWal, walFramePage(pWal->hdr.mxFrame), &aHash, &aPgno, &iZero);

  /* Zero all hash-table entries that correspond to frame numbers greater
  ** than pWal->hdr.mxFrame.
  */
  iLimit = pWal->hdr.mxFrame - iZero;
  assert( iLimit>0 );
  for(i=0; i<HASHTABLE_NSLOT; i++){
    if( aHash[i]>iLimit ){
      aHash[i] = 0;
    }
  }
  
  /* Zero the entries in the aPgno array that correspond to frames with
  ** frame numbers greater than pWal->hdr.mxFrame. 
  */
  nByte = (int)((char *)aHash - (char *)&aPgno[iLimit+1]);
  memset((void *)&aPgno[iLimit+1], 0, nByte);

#ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT
  /* Verify that the every entry in the mapping region is still reachable
  ** via the hash table even after the cleanup.
  */
  if( iLimit ){
    int j;           /* Loop counter */
    int iKey;        /* Hash key */
    for(j=1; j<=iLimit; j++){
      for(iKey=walHash(aPgno[j]); aHash[iKey]; iKey=walNextHash(iKey)){
        if( aHash[iKey]==j ) break;
      }
      assert( aHash[iKey]==j );
    }
  }
#endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */
}


/*
** Set an entry in the wal-index that will map database page number
** pPage into WAL frame iFrame.
*/
static int walIndexAppend(Wal *pWal, u32 iFrame, u32 iPage){
  int rc;                         /* Return code */
  u32 iZero = 0;                  /* One less than frame number of aPgno[1] */
  volatile u32 *aPgno = 0;        /* Page number array */
  volatile ht_slot *aHash = 0;    /* Hash table */

  rc = walHashGet(pWal, walFramePage(iFrame), &aHash, &aPgno, &iZero);

  /* Assuming the wal-index file was successfully mapped, populate the
  ** page number array and hash table entry.
  */
  if( rc==SQLITE_OK ){
    int iKey;                     /* Hash table key */
    int idx;                      /* Value to write to hash-table slot */
    int nCollide;                 /* Number of hash collisions */

    idx = iFrame - iZero;
    assert( idx <= HASHTABLE_NSLOT/2 + 1 );
    
    /* If this is the first entry to be added to this hash-table, zero the
    ** entire hash table and aPgno[] array before proceeding. 
    */
    if( idx==1 ){
      int nByte = (int)((u8 *)&aHash[HASHTABLE_NSLOT] - (u8 *)&aPgno[1]);

      memset((void*)&aPgno[1], 0, nByte);
    }

    /* If the entry in aPgno[] is already set, then the previous writer
    ** must have exited unexpectedly in the middle of a transaction (after
    ** writing one or more dirty pages to the WAL to free up memory). 
    ** Remove the remnants of that writers uncommitted transaction from 
    ** the hash-table before writing any new entries.
    */
    if( aPgno[idx] ){
      walCleanupHash(pWal);
      assert( !aPgno[idx] );
    }

    /* Write the aPgno[] array entry and the hash-table slot. */
    nCollide = idx;
    for(iKey=walHash(iPage); aHash[iKey]; iKey=walNextHash(iKey)){
      if( (nCollide--)==0 ) return SQLITE_CORRUPT_BKPT;
    }
    aPgno[idx] = iPage;
    aHash[iKey] = (ht_slot)idx;

#ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT
    /* Verify that the number of entries in the hash table exactly equals
    ** the number of entries in the mapping region.
    */
    {
      int i;           /* Loop counter */
      int nEntry = 0;  /* Number of entries in the hash table */
      for(i=0; i<HASHTABLE_NSLOT; i++){ if( aHash[i] ) nEntry++; }
      assert( nEntry==idx );
    }

    /* Verify that the every entry in the mapping region is reachable
    ** via the hash table.  This turns out to be a really, really expensive
    ** thing to check, so only do this occasionally - not on every
    ** iteration.
    */
    if( (idx&0x3ff)==0 ){
      int i;           /* Loop counter */
      for(i=1; i<=idx; i++){


        for(iKey=walHash(aPgno[i]); aHash[iKey]; iKey=walNextHash(iKey)){
          if( aHash[iKey]==i ) break;
        }
        assert( aHash[iKey]==i );
      }
    }
#endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */
  }


  return rc;
................................................................................
      sizeof(ht_slot) * (iLast>HASHTABLE_NPAGE?HASHTABLE_NPAGE:iLast)
  );
  if( !aTmp ){
    rc = SQLITE_NOMEM_BKPT;
  }

  for(i=walFramePage(nBackfill+1); rc==SQLITE_OK && i<nSegment; i++){
    volatile ht_slot *aHash;
    u32 iZero;
    volatile u32 *aPgno;

    rc = walHashGet(pWal, i, &aHash, &aPgno, &iZero);
    if( rc==SQLITE_OK ){
      int j;                      /* Counter variable */
      int nEntry;                 /* Number of entries in this segment */
      ht_slot *aIndex;            /* Sorted index for this segment */

      aPgno++;
      if( (i+1)==nSegment ){
        nEntry = (int)(iLast - iZero);
      }else{
        nEntry = (int)((u32*)aHash - (u32*)aPgno);
      }
      aIndex = &((ht_slot *)&p->aSegment[p->nSegment])[iZero];
      iZero++;
  
      for(j=0; j<nEntry; j++){
        aIndex[j] = (ht_slot)j;
      }
      walMergesort((u32 *)aPgno, aTmp, aIndex, &nEntry);
      p->aSegment[i].iZero = iZero;
      p->aSegment[i].nEntry = nEntry;
      p->aSegment[i].aIndex = aIndex;
      p->aSegment[i].aPgno = (u32 *)aPgno;
    }
  }
  sqlite3_free(aTmp);

  if( rc!=SQLITE_OK ){
    walIteratorFree(p);
    p = 0;
................................................................................
      void *pBuf1 = sqlite3_malloc(szPage);
      void *pBuf2 = sqlite3_malloc(szPage);
      if( pBuf1==0 || pBuf2==0 ){
        rc = SQLITE_NOMEM;
      }else{
        u32 i = pInfo->nBackfillAttempted;
        for(i=pInfo->nBackfillAttempted; i>pInfo->nBackfill; i--){
          volatile ht_slot *dummy;
          volatile u32 *aPgno;      /* Array of page numbers */
          u32 iZero;                /* Frame corresponding to aPgno[0] */
          u32 pgno;                 /* Page number in db file */
          i64 iDbOff;               /* Offset of db file entry */
          i64 iWalOff;              /* Offset of wal file entry */

          rc = walHashGet(pWal, walFramePage(i), &dummy, &aPgno, &iZero);
          if( rc!=SQLITE_OK ) break;
          pgno = aPgno[i-iZero];
          iDbOff = (i64)(pgno-1) * szPage;

          if( iDbOff+szPage<=szDb ){
            iWalOff = walFrameOffset(i, szPage) + WAL_FRAME_HDRSIZE;
            rc = sqlite3OsRead(pWal->pWalFd, pBuf1, szPage, iWalOff);

            if( rc==SQLITE_OK ){
................................................................................
  **
  **   (iFrame<=iLast): 
  **     This condition filters out entries that were added to the hash
  **     table after the current read-transaction had started.
  */
  iMinHash = walFramePage(pWal->minFrame);
  for(iHash=walFramePage(iLast); iHash>=iMinHash; iHash--){
    volatile ht_slot *aHash;      /* Pointer to hash table */
    volatile u32 *aPgno;          /* Pointer to array of page numbers */
    u32 iZero;                    /* Frame number corresponding to aPgno[0] */
    int iKey;                     /* Hash slot index */
    int nCollide;                 /* Number of hash collisions remaining */
    int rc;                       /* Error code */

    rc = walHashGet(pWal, iHash, &aHash, &aPgno, &iZero);
    if( rc!=SQLITE_OK ){
      return rc;
    }
    nCollide = HASHTABLE_NSLOT;
    for(iKey=walHash(pgno); aHash[iKey]; iKey=walNextHash(iKey)){
      u32 iFrame = aHash[iKey] + iZero;
      if( iFrame<=iLast && iFrame>=pWal->minFrame && aPgno[aHash[iKey]]==pgno ){

        assert( iFrame>iRead || CORRUPT_DB );
        iRead = iFrame;
      }
      if( (nCollide--)==0 ){
        return SQLITE_CORRUPT_BKPT;
      }
    }







>
>
>
>
>
>
>
>
>
>
>
>






|
|


|

|
|




|
<
<


<

|



<
<
<
|

|
|

|

<
<
<
<
>







 







|
<
<







 







|




|


|
|






|
|









|
|

|












|
<
<

|









|






|
>
|








|

|




|


|
|








|











>
>
|
|

|







 







|
<
<

|





|

|

|

|
|




|
|


|







 







|
<
<




|

|







 







|
<
<




|




|
|
|
>







875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912


913
914

915
916
917
918
919



920
921
922
923
924
925
926




927
928
929
930
931
932
933
934
...
966
967
968
969
970
971
972
973


974
975
976
977
978
979
980
...
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034


1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
....
1632
1633
1634
1635
1636
1637
1638
1639


1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
....
2669
2670
2671
2672
2673
2674
2675
2676


2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
....
2877
2878
2879
2880
2881
2882
2883
2884


2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
  assert( iPage>0 );
  assert( (HASHTABLE_NSLOT & (HASHTABLE_NSLOT-1))==0 );
  return (iPage*HASHTABLE_HASH_1) & (HASHTABLE_NSLOT-1);
}
static int walNextHash(int iPriorHash){
  return (iPriorHash+1)&(HASHTABLE_NSLOT-1);
}

/*
** An instance of the WalHashLoc object is used to describe the location
** of a page hash table in the wal-index.  This becomes the return value
** from walHashGet().
*/
typedef struct WalHashLoc WalHashLoc;
struct WalHashLoc {
  volatile ht_slot *aHash;  /* Start of the wal-index hash table */
  volatile u32 *aPgno;      /* aPgno[1] is the page of first frame indexed */
  u32 iZero;                /* One less than the frame number of first indexed*/
};

/* 
** Return pointers to the hash table and page number array stored on
** page iHash of the wal-index. The wal-index is broken into 32KB pages
** numbered starting from 0.
**
** Set output variable pLoc->aHash to point to the start of the hash table
** in the wal-index file. Set pLoc->iZero to one less than the frame 
** number of the first frame indexed by this hash table. If a
** slot in the hash table is set to N, it refers to frame number 
** (pLoc->iZero+N) in the log.
**
** Finally, set pLoc->aPgno so that pLoc->aPgno[1] is the page number of the
** first frame indexed by the hash table, frame (pLoc->iZero+1).
*/
static int walHashGet(
  Wal *pWal,                      /* WAL handle */
  int iHash,                      /* Find the iHash'th table */
  WalHashLoc *pLoc                /* OUT: Hash table location */


){
  int rc;                         /* Return code */


  rc = walIndexPage(pWal, iHash, &pLoc->aPgno);
  assert( rc==SQLITE_OK || iHash>0 );

  if( rc==SQLITE_OK ){



    pLoc->aHash = (volatile ht_slot *)&pLoc->aPgno[HASHTABLE_NPAGE];
    if( iHash==0 ){
      pLoc->aPgno = &pLoc->aPgno[WALINDEX_HDR_SIZE/sizeof(u32)];
      pLoc->iZero = 0;
    }else{
      pLoc->iZero = HASHTABLE_NPAGE_ONE + (iHash-1)*HASHTABLE_NPAGE;
    }




    pLoc->aPgno = &pLoc->aPgno[-1];
  }
  return rc;
}

/*
** Return the number of the wal-index page that contains the hash-table
** and page-number array that contain entries corresponding to WAL frame
................................................................................
**
** At most only the hash table containing pWal->hdr.mxFrame needs to be
** updated.  Any later hash tables will be automatically cleared when
** pWal->hdr.mxFrame advances to the point where those hash tables are
** actually needed.
*/
static void walCleanupHash(Wal *pWal){
  WalHashLoc sLoc;                /* Hash table location */


  int iLimit = 0;                 /* Zero values greater than this */
  int nByte;                      /* Number of bytes to zero in aPgno[] */
  int i;                          /* Used to iterate through aHash[] */

  assert( pWal->writeLock );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE-1 );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE_ONE );
................................................................................

  /* Obtain pointers to the hash-table and page-number array containing 
  ** the entry that corresponds to frame pWal->hdr.mxFrame. It is guaranteed
  ** that the page said hash-table and array reside on is already mapped.
  */
  assert( pWal->nWiData>walFramePage(pWal->hdr.mxFrame) );
  assert( pWal->apWiData[walFramePage(pWal->hdr.mxFrame)] );
  walHashGet(pWal, walFramePage(pWal->hdr.mxFrame), &sLoc);

  /* Zero all hash-table entries that correspond to frame numbers greater
  ** than pWal->hdr.mxFrame.
  */
  iLimit = pWal->hdr.mxFrame - sLoc.iZero;
  assert( iLimit>0 );
  for(i=0; i<HASHTABLE_NSLOT; i++){
    if( sLoc.aHash[i]>iLimit ){
      sLoc.aHash[i] = 0;
    }
  }
  
  /* Zero the entries in the aPgno array that correspond to frames with
  ** frame numbers greater than pWal->hdr.mxFrame. 
  */
  nByte = (int)((char *)sLoc.aHash - (char *)&sLoc.aPgno[iLimit+1]);
  memset((void *)&sLoc.aPgno[iLimit+1], 0, nByte);

#ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT
  /* Verify that the every entry in the mapping region is still reachable
  ** via the hash table even after the cleanup.
  */
  if( iLimit ){
    int j;           /* Loop counter */
    int iKey;        /* Hash key */
    for(j=1; j<=iLimit; j++){
      for(iKey=walHash(sLoc.aPgno[j]);sLoc.aHash[iKey];iKey=walNextHash(iKey)){
        if( sLoc.aHash[iKey]==j ) break;
      }
      assert( sLoc.aHash[iKey]==j );
    }
  }
#endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */
}


/*
** Set an entry in the wal-index that will map database page number
** pPage into WAL frame iFrame.
*/
static int walIndexAppend(Wal *pWal, u32 iFrame, u32 iPage){
  int rc;                         /* Return code */
  WalHashLoc sLoc;                /* Wal-index hash table location */



  rc = walHashGet(pWal, walFramePage(iFrame), &sLoc);

  /* Assuming the wal-index file was successfully mapped, populate the
  ** page number array and hash table entry.
  */
  if( rc==SQLITE_OK ){
    int iKey;                     /* Hash table key */
    int idx;                      /* Value to write to hash-table slot */
    int nCollide;                 /* Number of hash collisions */

    idx = iFrame - sLoc.iZero;
    assert( idx <= HASHTABLE_NSLOT/2 + 1 );
    
    /* If this is the first entry to be added to this hash-table, zero the
    ** entire hash table and aPgno[] array before proceeding. 
    */
    if( idx==1 ){
      int nByte = (int)((u8 *)&sLoc.aHash[HASHTABLE_NSLOT]
                               - (u8 *)&sLoc.aPgno[1]);
      memset((void*)&sLoc.aPgno[1], 0, nByte);
    }

    /* If the entry in aPgno[] is already set, then the previous writer
    ** must have exited unexpectedly in the middle of a transaction (after
    ** writing one or more dirty pages to the WAL to free up memory). 
    ** Remove the remnants of that writers uncommitted transaction from 
    ** the hash-table before writing any new entries.
    */
    if( sLoc.aPgno[idx] ){
      walCleanupHash(pWal);
      assert( !sLoc.aPgno[idx] );
    }

    /* Write the aPgno[] array entry and the hash-table slot. */
    nCollide = idx;
    for(iKey=walHash(iPage); sLoc.aHash[iKey]; iKey=walNextHash(iKey)){
      if( (nCollide--)==0 ) return SQLITE_CORRUPT_BKPT;
    }
    sLoc.aPgno[idx] = iPage;
    sLoc.aHash[iKey] = (ht_slot)idx;

#ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT
    /* Verify that the number of entries in the hash table exactly equals
    ** the number of entries in the mapping region.
    */
    {
      int i;           /* Loop counter */
      int nEntry = 0;  /* Number of entries in the hash table */
      for(i=0; i<HASHTABLE_NSLOT; i++){ if( sLoc.aHash[i] ) nEntry++; }
      assert( nEntry==idx );
    }

    /* Verify that the every entry in the mapping region is reachable
    ** via the hash table.  This turns out to be a really, really expensive
    ** thing to check, so only do this occasionally - not on every
    ** iteration.
    */
    if( (idx&0x3ff)==0 ){
      int i;           /* Loop counter */
      for(i=1; i<=idx; i++){
        for(iKey=walHash(sLoc.aPgno[i]);
            sLoc.aHash[iKey];
            iKey=walNextHash(iKey)){
          if( sLoc.aHash[iKey]==i ) break;
        }
        assert( sLoc.aHash[iKey]==i );
      }
    }
#endif /* SQLITE_ENABLE_EXPENSIVE_ASSERT */
  }


  return rc;
................................................................................
      sizeof(ht_slot) * (iLast>HASHTABLE_NPAGE?HASHTABLE_NPAGE:iLast)
  );
  if( !aTmp ){
    rc = SQLITE_NOMEM_BKPT;
  }

  for(i=walFramePage(nBackfill+1); rc==SQLITE_OK && i<nSegment; i++){
    WalHashLoc sLoc;



    rc = walHashGet(pWal, i, &sLoc);
    if( rc==SQLITE_OK ){
      int j;                      /* Counter variable */
      int nEntry;                 /* Number of entries in this segment */
      ht_slot *aIndex;            /* Sorted index for this segment */

      sLoc.aPgno++;
      if( (i+1)==nSegment ){
        nEntry = (int)(iLast - sLoc.iZero);
      }else{
        nEntry = (int)((u32*)sLoc.aHash - (u32*)sLoc.aPgno);
      }
      aIndex = &((ht_slot *)&p->aSegment[p->nSegment])[sLoc.iZero];
      sLoc.iZero++;
  
      for(j=0; j<nEntry; j++){
        aIndex[j] = (ht_slot)j;
      }
      walMergesort((u32 *)sLoc.aPgno, aTmp, aIndex, &nEntry);
      p->aSegment[i].iZero = sLoc.iZero;
      p->aSegment[i].nEntry = nEntry;
      p->aSegment[i].aIndex = aIndex;
      p->aSegment[i].aPgno = (u32 *)sLoc.aPgno;
    }
  }
  sqlite3_free(aTmp);

  if( rc!=SQLITE_OK ){
    walIteratorFree(p);
    p = 0;
................................................................................
      void *pBuf1 = sqlite3_malloc(szPage);
      void *pBuf2 = sqlite3_malloc(szPage);
      if( pBuf1==0 || pBuf2==0 ){
        rc = SQLITE_NOMEM;
      }else{
        u32 i = pInfo->nBackfillAttempted;
        for(i=pInfo->nBackfillAttempted; i>pInfo->nBackfill; i--){
          WalHashLoc sLoc;          /* Hash table location */


          u32 pgno;                 /* Page number in db file */
          i64 iDbOff;               /* Offset of db file entry */
          i64 iWalOff;              /* Offset of wal file entry */

          rc = walHashGet(pWal, walFramePage(i), &sLoc);
          if( rc!=SQLITE_OK ) break;
          pgno = sLoc.aPgno[i-sLoc.iZero];
          iDbOff = (i64)(pgno-1) * szPage;

          if( iDbOff+szPage<=szDb ){
            iWalOff = walFrameOffset(i, szPage) + WAL_FRAME_HDRSIZE;
            rc = sqlite3OsRead(pWal->pWalFd, pBuf1, szPage, iWalOff);

            if( rc==SQLITE_OK ){
................................................................................
  **
  **   (iFrame<=iLast): 
  **     This condition filters out entries that were added to the hash
  **     table after the current read-transaction had started.
  */
  iMinHash = walFramePage(pWal->minFrame);
  for(iHash=walFramePage(iLast); iHash>=iMinHash; iHash--){
    WalHashLoc sLoc;              /* Hash table location */


    int iKey;                     /* Hash slot index */
    int nCollide;                 /* Number of hash collisions remaining */
    int rc;                       /* Error code */

    rc = walHashGet(pWal, iHash, &sLoc);
    if( rc!=SQLITE_OK ){
      return rc;
    }
    nCollide = HASHTABLE_NSLOT;
    for(iKey=walHash(pgno); sLoc.aHash[iKey]; iKey=walNextHash(iKey)){
      u32 iFrame = sLoc.aHash[iKey] + sLoc.iZero;
      if( iFrame<=iLast && iFrame>=pWal->minFrame
       && sLoc.aPgno[sLoc.aHash[iKey]]==pgno ){
        assert( iFrame>iRead || CORRUPT_DB );
        iRead = iFrame;
      }
      if( (nCollide--)==0 ){
        return SQLITE_CORRUPT_BKPT;
      }
    }

Changes to src/where.c.

2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
....
2467
2468
2469
2470
2471
2472
2473




































2474
2475
2476
2477
2478
2479
2480
....
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
....
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
....
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
....
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
        || (pNew->wsFlags & WHERE_COLUMN_NULL)!=0 
        || (pNew->wsFlags & WHERE_COLUMN_IN)!=0 
        || (pNew->wsFlags & WHERE_SKIPSCAN)!=0 
    );

    if( eOp & WO_IN ){
      Expr *pExpr = pTerm->pExpr;
      pNew->wsFlags |= WHERE_COLUMN_IN;
      if( ExprHasProperty(pExpr, EP_xIsSelect) ){
        /* "x IN (SELECT ...)":  TUNING: the SELECT returns 25 rows */
        int i;
        nIn = 46;  assert( 46==sqlite3LogEst(25) );

        /* The expression may actually be of the form (x, y) IN (SELECT...).
        ** In this case there is a separate term for each of (x) and (y).
................................................................................
        }
      }else if( ALWAYS(pExpr->x.pList && pExpr->x.pList->nExpr) ){
        /* "x IN (value, value, ...)" */
        nIn = sqlite3LogEst(pExpr->x.pList->nExpr);
        assert( nIn>0 );  /* RHS always has 2 or more terms...  The parser
                          ** changes "x IN (?)" into "x=?". */
      }




































    }else if( eOp & (WO_EQ|WO_IS) ){
      int iCol = pProbe->aiColumn[saved_nEq];
      pNew->wsFlags |= WHERE_COLUMN_EQ;
      assert( saved_nEq==pNew->u.btree.nEq );
      if( iCol==XN_ROWID 
       || (iCol>=0 && nInMul==0 && saved_nEq==pProbe->nKeyCol-1)
      ){
................................................................................
        }
      }
    }
  }
  return 0;
}

/*
** Return a bitmask where 1s indicate that the corresponding column of
** the table is used by an index.  Only the first 63 columns are considered.
*/
static Bitmask columnsInIndex(Index *pIdx){
  Bitmask m = 0;
  int j;
  for(j=pIdx->nColumn-1; j>=0; j--){
    int x = pIdx->aiColumn[j];
    if( x>=0 ){
      testcase( x==BMS-1 );
      testcase( x==BMS-2 );
      if( x<BMS-1 ) m |= MASKBIT(x);
    }
  }
  return m;
}

/* Check to see if a partial index with pPartIndexWhere can be used
** in the current query.  Return true if it can be and false if not.
*/
static int whereUsablePartialIndex(int iTab, WhereClause *pWC, Expr *pWhere){
  int i;
  WhereTerm *pTerm;
  Parse *pParse = pWC->pWInfo->pParse;
................................................................................
      if( rc ) break;
    }else{
      Bitmask m;
      if( pProbe->isCovering ){
        pNew->wsFlags = WHERE_IDX_ONLY | WHERE_INDEXED;
        m = 0;
      }else{
        m = pSrc->colUsed & ~columnsInIndex(pProbe);
        pNew->wsFlags = (m==0) ? (WHERE_IDX_ONLY|WHERE_INDEXED) : WHERE_INDEXED;
      }

      /* Full scan via index */
      if( b
       || !HasRowid(pTab)
       || pProbe->pPartIdxWhere!=0
................................................................................
      }
      rc = whereLoopAddVirtual(pBuilder, mPrereq, mUnusable);
    }else
#endif /* SQLITE_OMIT_VIRTUALTABLE */
    {
      rc = whereLoopAddBtree(pBuilder, mPrereq);
    }
    if( rc==SQLITE_OK ){
      rc = whereLoopAddOr(pBuilder, mPrereq, mUnusable);
    }
    mPrior |= pNew->maskSelf;
    if( rc || db->mallocFailed ) break;
  }

  whereLoopClear(db, pNew);
................................................................................
        pTerm = sqlite3WhereFindTerm(pWC, iCur, j, 0, opMask, pIdx);
        if( pTerm==0 ) break;
        testcase( pTerm->eOperator & WO_IS );
        pLoop->aLTerm[j] = pTerm;
      }
      if( j!=pIdx->nKeyCol ) continue;
      pLoop->wsFlags = WHERE_COLUMN_EQ|WHERE_ONEROW|WHERE_INDEXED;
      if( pIdx->isCovering || (pItem->colUsed & ~columnsInIndex(pIdx))==0 ){
        pLoop->wsFlags |= WHERE_IDX_ONLY;
      }
      pLoop->nLTerm = j;
      pLoop->u.btree.nEq = j;
      pLoop->u.btree.pIndex = pIdx;
      /* TUNING: Cost of a unique index lookup is 15 */
      pLoop->rRun = 39;  /* 39==sqlite3LogEst(15) */







<







 







>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>







 







<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<







 







|







 







|







 







|







2447
2448
2449
2450
2451
2452
2453

2454
2455
2456
2457
2458
2459
2460
....
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
....
2731
2732
2733
2734
2735
2736
2737


















2738
2739
2740
2741
2742
2743
2744
....
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
....
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
....
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
        || (pNew->wsFlags & WHERE_COLUMN_NULL)!=0 
        || (pNew->wsFlags & WHERE_COLUMN_IN)!=0 
        || (pNew->wsFlags & WHERE_SKIPSCAN)!=0 
    );

    if( eOp & WO_IN ){
      Expr *pExpr = pTerm->pExpr;

      if( ExprHasProperty(pExpr, EP_xIsSelect) ){
        /* "x IN (SELECT ...)":  TUNING: the SELECT returns 25 rows */
        int i;
        nIn = 46;  assert( 46==sqlite3LogEst(25) );

        /* The expression may actually be of the form (x, y) IN (SELECT...).
        ** In this case there is a separate term for each of (x) and (y).
................................................................................
        }
      }else if( ALWAYS(pExpr->x.pList && pExpr->x.pList->nExpr) ){
        /* "x IN (value, value, ...)" */
        nIn = sqlite3LogEst(pExpr->x.pList->nExpr);
        assert( nIn>0 );  /* RHS always has 2 or more terms...  The parser
                          ** changes "x IN (?)" into "x=?". */
      }
      if( pProbe->hasStat1 ){
        LogEst M, logK, safetyMargin;
        /* Let:
        **   N = the total number of rows in the table
        **   K = the number of entries on the RHS of the IN operator
        **   M = the number of rows in the table that match terms to the 
        **       to the left in the same index.  If the IN operator is on
        **       the left-most index column, M==N.
        **
        ** Given the definitions above, it is better to omit the IN operator
        ** from the index lookup and instead do a scan of the M elements,
        ** testing each scanned row against the IN operator separately, if:
        **
        **        M*log(K) < K*log(N)
        **
        ** Our estimates for M, K, and N might be inaccurate, so we build in
        ** a safety margin of 2 (LogEst: 10) that favors using the IN operator
        ** with the index, as using an index has better worst-case behavior.
        ** If we do not have real sqlite_stat1 data, always prefer to use
        ** the index.
        */
        M = pProbe->aiRowLogEst[saved_nEq];
        logK = estLog(nIn);
        safetyMargin = 10;  /* TUNING: extra weight for indexed IN */
        if( M + logK + safetyMargin < nIn + rLogSize ){
          WHERETRACE(0x40,
            ("Scan preferred over IN operator on column %d of \"%s\" (%d<%d)\n",
             saved_nEq, pProbe->zName, M+logK+10, nIn+rLogSize));
          continue;
        }else{
          WHERETRACE(0x40,
            ("IN operator preferred on column %d of \"%s\" (%d>=%d)\n",
             saved_nEq, pProbe->zName, M+logK+10, nIn+rLogSize));
        }
      }
      pNew->wsFlags |= WHERE_COLUMN_IN;
    }else if( eOp & (WO_EQ|WO_IS) ){
      int iCol = pProbe->aiColumn[saved_nEq];
      pNew->wsFlags |= WHERE_COLUMN_EQ;
      assert( saved_nEq==pNew->u.btree.nEq );
      if( iCol==XN_ROWID 
       || (iCol>=0 && nInMul==0 && saved_nEq==pProbe->nKeyCol-1)
      ){
................................................................................
        }
      }
    }
  }
  return 0;
}



















/* Check to see if a partial index with pPartIndexWhere can be used
** in the current query.  Return true if it can be and false if not.
*/
static int whereUsablePartialIndex(int iTab, WhereClause *pWC, Expr *pWhere){
  int i;
  WhereTerm *pTerm;
  Parse *pParse = pWC->pWInfo->pParse;
................................................................................
      if( rc ) break;
    }else{
      Bitmask m;
      if( pProbe->isCovering ){
        pNew->wsFlags = WHERE_IDX_ONLY | WHERE_INDEXED;
        m = 0;
      }else{
        m = pSrc->colUsed & pProbe->colNotIdxed;
        pNew->wsFlags = (m==0) ? (WHERE_IDX_ONLY|WHERE_INDEXED) : WHERE_INDEXED;
      }

      /* Full scan via index */
      if( b
       || !HasRowid(pTab)
       || pProbe->pPartIdxWhere!=0
................................................................................
      }
      rc = whereLoopAddVirtual(pBuilder, mPrereq, mUnusable);
    }else
#endif /* SQLITE_OMIT_VIRTUALTABLE */
    {
      rc = whereLoopAddBtree(pBuilder, mPrereq);
    }
    if( rc==SQLITE_OK && pBuilder->pWC->hasOr ){
      rc = whereLoopAddOr(pBuilder, mPrereq, mUnusable);
    }
    mPrior |= pNew->maskSelf;
    if( rc || db->mallocFailed ) break;
  }

  whereLoopClear(db, pNew);
................................................................................
        pTerm = sqlite3WhereFindTerm(pWC, iCur, j, 0, opMask, pIdx);
        if( pTerm==0 ) break;
        testcase( pTerm->eOperator & WO_IS );
        pLoop->aLTerm[j] = pTerm;
      }
      if( j!=pIdx->nKeyCol ) continue;
      pLoop->wsFlags = WHERE_COLUMN_EQ|WHERE_ONEROW|WHERE_INDEXED;
      if( pIdx->isCovering || (pItem->colUsed & pIdx->colNotIdxed)==0 ){
        pLoop->wsFlags |= WHERE_IDX_ONLY;
      }
      pLoop->nLTerm = j;
      pLoop->u.btree.nEq = j;
      pLoop->u.btree.pIndex = pIdx;
      /* TUNING: Cost of a unique index lookup is 15 */
      pLoop->rRun = 39;  /* 39==sqlite3LogEst(15) */

Changes to src/whereInt.h.

318
319
320
321
322
323
324

325
326
327
328
329
330
331
...
491
492
493
494
495
496
497

498
499
500
501
502
503
504
** the subclauses "(b AND c)" and "(d AND e)".  The pOuter field of the
** subclauses points to the WhereClause object for the whole clause.
*/
struct WhereClause {
  WhereInfo *pWInfo;       /* WHERE clause processing context */
  WhereClause *pOuter;     /* Outer conjunction */
  u8 op;                   /* Split operator.  TK_AND or TK_OR */

  int nTerm;               /* Number of terms */
  int nSlot;               /* Number of entries in a[] */
  WhereTerm *a;            /* Each a[] describes a term of the WHERE cluase */
#if defined(SQLITE_SMALL_STACK)
  WhereTerm aStatic[1];    /* Initial static space for a[] */
#else
  WhereTerm aStatic[8];    /* Initial static space for a[] */
................................................................................
);

/* whereexpr.c: */
void sqlite3WhereClauseInit(WhereClause*,WhereInfo*);
void sqlite3WhereClauseClear(WhereClause*);
void sqlite3WhereSplit(WhereClause*,Expr*,u8);
Bitmask sqlite3WhereExprUsage(WhereMaskSet*, Expr*);

Bitmask sqlite3WhereExprListUsage(WhereMaskSet*, ExprList*);
void sqlite3WhereExprAnalyze(SrcList*, WhereClause*);
void sqlite3WhereTabFuncArgs(Parse*, struct SrcList_item*, WhereClause*);











>







 







>







318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
...
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
** the subclauses "(b AND c)" and "(d AND e)".  The pOuter field of the
** subclauses points to the WhereClause object for the whole clause.
*/
struct WhereClause {
  WhereInfo *pWInfo;       /* WHERE clause processing context */
  WhereClause *pOuter;     /* Outer conjunction */
  u8 op;                   /* Split operator.  TK_AND or TK_OR */
  u8 hasOr;                /* True if any a[].eOperator is WO_OR */
  int nTerm;               /* Number of terms */
  int nSlot;               /* Number of entries in a[] */
  WhereTerm *a;            /* Each a[] describes a term of the WHERE cluase */
#if defined(SQLITE_SMALL_STACK)
  WhereTerm aStatic[1];    /* Initial static space for a[] */
#else
  WhereTerm aStatic[8];    /* Initial static space for a[] */
................................................................................
);

/* whereexpr.c: */
void sqlite3WhereClauseInit(WhereClause*,WhereInfo*);
void sqlite3WhereClauseClear(WhereClause*);
void sqlite3WhereSplit(WhereClause*,Expr*,u8);
Bitmask sqlite3WhereExprUsage(WhereMaskSet*, Expr*);
Bitmask sqlite3WhereExprUsageNN(WhereMaskSet*, Expr*);
Bitmask sqlite3WhereExprListUsage(WhereMaskSet*, ExprList*);
void sqlite3WhereExprAnalyze(SrcList*, WhereClause*);
void sqlite3WhereTabFuncArgs(Parse*, struct SrcList_item*, WhereClause*);




Changes to src/whereexpr.c.

668
669
670
671
672
673
674

675




676
677
678
679
680
681
682
....
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
....
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447



1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461



1462
1463
1464
1465
1466
1467
1468
  }

  /*
  ** Record the set of tables that satisfy case 3.  The set might be
  ** empty.
  */
  pOrInfo->indexable = indexable;

  pTerm->eOperator = indexable==0 ? 0 : WO_OR;





  /* For a two-way OR, attempt to implementation case 2.
  */
  if( indexable && pOrWc->nTerm==2 ){
    int iOne = 0;
    WhereTerm *pOne;
    while( (pOne = whereNthSubterm(&pOrWc->a[0],iOne++))!=0 ){
................................................................................
    }
  }else if( op==TK_ISNULL ){
    pTerm->prereqRight = 0;
  }else{
    pTerm->prereqRight = sqlite3WhereExprUsage(pMaskSet, pExpr->pRight);
  }
  pMaskSet->bVarSelect = 0;
  prereqAll = sqlite3WhereExprUsage(pMaskSet, pExpr);
  if( pMaskSet->bVarSelect ) pTerm->wtFlags |= TERM_VARSELECT;
  if( ExprHasProperty(pExpr, EP_FromJoin) ){
    Bitmask x = sqlite3WhereGetMask(pMaskSet, pExpr->iRightJoinTable);
    prereqAll |= x;
    extraRight = x-1;  /* ON clause terms may not be used with an index
                       ** on left table of a LEFT JOIN.  Ticket #3015 */
    if( (prereqAll>>1)>=x ){
................................................................................


/*
** These routines walk (recursively) an expression tree and generate
** a bitmask indicating which tables are used in that expression
** tree.
*/
Bitmask sqlite3WhereExprUsage(WhereMaskSet *pMaskSet, Expr *p){
  Bitmask mask;
  if( p==0 ) return 0;
  if( p->op==TK_COLUMN ){
    return sqlite3WhereGetMask(pMaskSet, p->iTable);



  }
  mask = (p->op==TK_IF_NULL_ROW) ? sqlite3WhereGetMask(pMaskSet, p->iTable) : 0;
  assert( !ExprHasProperty(p, EP_TokenOnly) );
  if( p->pLeft ) mask |= sqlite3WhereExprUsage(pMaskSet, p->pLeft);
  if( p->pRight ){
    mask |= sqlite3WhereExprUsage(pMaskSet, p->pRight);
    assert( p->x.pList==0 );
  }else if( ExprHasProperty(p, EP_xIsSelect) ){
    if( ExprHasProperty(p, EP_VarSelect) ) pMaskSet->bVarSelect = 1;
    mask |= exprSelectUsage(pMaskSet, p->x.pSelect);
  }else if( p->x.pList ){
    mask |= sqlite3WhereExprListUsage(pMaskSet, p->x.pList);
  }
  return mask;



}
Bitmask sqlite3WhereExprListUsage(WhereMaskSet *pMaskSet, ExprList *pList){
  int i;
  Bitmask mask = 0;
  if( pList ){
    for(i=0; i<pList->nExpr; i++){
      mask |= sqlite3WhereExprUsage(pMaskSet, pList->a[i].pExpr);







>
|
>
>
>
>







 







|







 







|

<


>
>
>


<
|

|








>
>
>







668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
....
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
....
1441
1442
1443
1444
1445
1446
1447
1448
1449

1450
1451
1452
1453
1454
1455
1456

1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
  }

  /*
  ** Record the set of tables that satisfy case 3.  The set might be
  ** empty.
  */
  pOrInfo->indexable = indexable;
  if( indexable ){
    pTerm->eOperator = WO_OR;
    pWC->hasOr = 1;
  }else{
    pTerm->eOperator = WO_OR;
  }

  /* For a two-way OR, attempt to implementation case 2.
  */
  if( indexable && pOrWc->nTerm==2 ){
    int iOne = 0;
    WhereTerm *pOne;
    while( (pOne = whereNthSubterm(&pOrWc->a[0],iOne++))!=0 ){
................................................................................
    }
  }else if( op==TK_ISNULL ){
    pTerm->prereqRight = 0;
  }else{
    pTerm->prereqRight = sqlite3WhereExprUsage(pMaskSet, pExpr->pRight);
  }
  pMaskSet->bVarSelect = 0;
  prereqAll = sqlite3WhereExprUsageNN(pMaskSet, pExpr);
  if( pMaskSet->bVarSelect ) pTerm->wtFlags |= TERM_VARSELECT;
  if( ExprHasProperty(pExpr, EP_FromJoin) ){
    Bitmask x = sqlite3WhereGetMask(pMaskSet, pExpr->iRightJoinTable);
    prereqAll |= x;
    extraRight = x-1;  /* ON clause terms may not be used with an index
                       ** on left table of a LEFT JOIN.  Ticket #3015 */
    if( (prereqAll>>1)>=x ){
................................................................................


/*
** These routines walk (recursively) an expression tree and generate
** a bitmask indicating which tables are used in that expression
** tree.
*/
Bitmask sqlite3WhereExprUsageNN(WhereMaskSet *pMaskSet, Expr *p){
  Bitmask mask;

  if( p->op==TK_COLUMN ){
    return sqlite3WhereGetMask(pMaskSet, p->iTable);
  }else if( ExprHasProperty(p, EP_TokenOnly|EP_Leaf) ){
    assert( p->op!=TK_IF_NULL_ROW );
    return 0;
  }
  mask = (p->op==TK_IF_NULL_ROW) ? sqlite3WhereGetMask(pMaskSet, p->iTable) : 0;

  if( p->pLeft ) mask |= sqlite3WhereExprUsageNN(pMaskSet, p->pLeft);
  if( p->pRight ){
    mask |= sqlite3WhereExprUsageNN(pMaskSet, p->pRight);
    assert( p->x.pList==0 );
  }else if( ExprHasProperty(p, EP_xIsSelect) ){
    if( ExprHasProperty(p, EP_VarSelect) ) pMaskSet->bVarSelect = 1;
    mask |= exprSelectUsage(pMaskSet, p->x.pSelect);
  }else if( p->x.pList ){
    mask |= sqlite3WhereExprListUsage(pMaskSet, p->x.pList);
  }
  return mask;
}
Bitmask sqlite3WhereExprUsage(WhereMaskSet *pMaskSet, Expr *p){
  return p ? sqlite3WhereExprUsageNN(pMaskSet,p) : 0;
}
Bitmask sqlite3WhereExprListUsage(WhereMaskSet *pMaskSet, ExprList *pList){
  int i;
  Bitmask mask = 0;
  if( pList ){
    for(i=0; i<pList->nExpr; i++){
      mask |= sqlite3WhereExprUsage(pMaskSet, pList->a[i].pExpr);

Changes to test/in6.test.

24
25
26
27
28
29
30



31
32
33
34
35
36
37
do_test in6-1.1 {
  db eval {
    CREATE TABLE t1(a,b,c,d);
    WITH RECURSIVE c(x) AS (VALUES(1) UNION ALL SELECT x+1 FROM c WHERE x<100)
      INSERT INTO t1(a,b,c,d)
        SELECT 100, 200+x/2, 300+x/5, x FROM c;
    CREATE INDEX t1abc ON t1(a,b,c);



  }
  set ::sqlite_search_count 0
  db eval {
    SELECT d FROM t1
     WHERE a=99
       AND b IN (200,205,201,204)
       AND c IN (304,302,309,308);







>
>
>







24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
do_test in6-1.1 {
  db eval {
    CREATE TABLE t1(a,b,c,d);
    WITH RECURSIVE c(x) AS (VALUES(1) UNION ALL SELECT x+1 FROM c WHERE x<100)
      INSERT INTO t1(a,b,c,d)
        SELECT 100, 200+x/2, 300+x/5, x FROM c;
    CREATE INDEX t1abc ON t1(a,b,c);
    ANALYZE;
    UPDATE sqlite_stat1 SET stat='1000000 500000 500 50';
    ANALYZE sqlite_master;
  }
  set ::sqlite_search_count 0
  db eval {
    SELECT d FROM t1
     WHERE a=99
       AND b IN (200,205,201,204)
       AND c IN (304,302,309,308);

Changes to test/rowvalue4.test.

220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
  CREATE TABLE d2(a, b, c);
  CREATE INDEX d2ab ON d2(a, b);
  CREATE INDEX d2c ON d2(c);

  WITH i(i) AS (
    VALUES(1) UNION ALL SELECT i+1 FROM i WHERE i<1000
  )
  INSERT INTO d2 SELECT i/3, i%3, i/3 FROM i;
  ANALYZE;
}

do_eqp_test 5.1 {
  SELECT * FROM d2 WHERE 
    (a, b) IN (SELECT x, y FROM d1) AND
    (c) IN (SELECT y FROM d1)







|







220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
  CREATE TABLE d2(a, b, c);
  CREATE INDEX d2ab ON d2(a, b);
  CREATE INDEX d2c ON d2(c);

  WITH i(i) AS (
    VALUES(1) UNION ALL SELECT i+1 FROM i WHERE i<1000
  )
  INSERT INTO d2 SELECT i/100, i%100, i/100 FROM i;
  ANALYZE;
}

do_eqp_test 5.1 {
  SELECT * FROM d2 WHERE 
    (a, b) IN (SELECT x, y FROM d1) AND
    (c) IN (SELECT y FROM d1)