SQLite

/*
** A convenience macro that returns the number of elements in
** an array.
*/
#define ArraySize(X)    ((int)(sizeof(X)/sizeof(X[0])))

/*
** Determine if the argument is a power of two
*/
#define IsPowerOfTwo(X) (((X)&((X)-1))==0)

/*
** The following value as a destructor means to use sqlite3DbFree().
** The sqlite3DbFree() routine requires two parameters instead of the 
** one parameter that destructors normally want.  So we have to introduce 
** this magic value that the code knows to handle differently.  Any 
** pointer will work here as long as it is distinct from SQLITE_STATIC
** and SQLITE_TRANSIENT.

typedef struct WhereTerm WhereTerm;
struct WhereTerm {
  Expr *pExpr;            /* Pointer to the subexpression that is this term */
  int iParent;            /* Disable pWC->a[iParent] when this term disabled */
  int leftCursor;         /* Cursor number of X in "X <op> <expr>" */
  union {
    int leftColumn;         /* Column number of X in "X <op> <expr>" */
    WhereOrInfo *pOrInfo;   /* Extra information if (eOperator & WO_OR)!=0 */
    WhereAndInfo *pAndInfo; /* Extra information if (eOperator& WO_AND)!=0 */
  } u;
  u16 eOperator;          /* A WO_xx value describing <op> */
  u8 wtFlags;             /* TERM_xxx bit flags.  See below */
  u8 nChild;              /* Number of children that must disable us */
  WhereClause *pWC;       /* The clause this term is part of */
  Bitmask prereqRight;    /* Bitmask of tables used by pExpr->pRight */
  Bitmask prereqAll;      /* Bitmask of tables referenced by pExpr */

#define WO_LE     (WO_EQ<<(TK_LE-TK_EQ))
#define WO_GT     (WO_EQ<<(TK_GT-TK_EQ))
#define WO_GE     (WO_EQ<<(TK_GE-TK_EQ))
#define WO_MATCH  0x040
#define WO_ISNULL 0x080
#define WO_OR     0x100       /* Two or more OR-connected terms */
#define WO_AND    0x200       /* Two or more AND-connected terms */
#define WO_EQUIV  0x400       /* Of the form A==B, both columns */
#define WO_NOOP   0x800       /* This term does not restrict search space */

#define WO_ALL    0xfff       /* Mask of all possible WO_* values */
#define WO_SINGLE 0x0ff       /* Mask of all non-compound WO_* values */

/*
** Value for wsFlags returned by bestIndex() and stored in

  int iCur,             /* Cursor number of LHS */
  int iColumn,          /* Column number of LHS */
  Bitmask notReady,     /* RHS must not overlap with this mask */
  u32 op,               /* Mask of WO_xx values describing operator */
  Index *pIdx           /* Must be compatible with this index, if not NULL */
){
  WhereTerm *pTerm;
  WhereTerm *pResult = 0;
  WhereClause *pWCOrig = pWC;
  int j, k;
  int aEquiv[8];
  int nEquiv = 2;
  int iEquiv = 2;
  Expr *pX;
  Parse *pParse;

  assert( iCur>=0 );
  aEquiv[0] = iCur;
  aEquiv[1] = iColumn;
  for(;;){
    for(pWC=pWCOrig; pWC; pWC=pWC->pOuter){
      for(pTerm=pWC->a, k=pWC->nTerm; k; k--, pTerm++){
        if( pTerm->leftCursor==iCur

          && pTerm->u.leftColumn==iColumn
          && (pTerm->eOperator & op & WO_ALL)!=0
        ){
          if( (pTerm->prereqRight & notReady)==0 ){
            if( iColumn>=0 && pIdx && (pTerm->eOperator & WO_ISNULL)==0 ){

              CollSeq *pColl;
              char idxaff;
      
              pX = pTerm->pExpr;
              pParse = pWC->pParse;

              idxaff = pIdx->pTable->aCol[iColumn].affinity;
              if( !sqlite3IndexAffinityOk(pX, idxaff) ) continue;
      
              /* Figure out the collation sequence required from an index for
              ** it to be useful for optimising expression pX. Store this
              ** value in variable pColl.
              */
              assert(pX->pLeft);
              pColl = sqlite3BinaryCompareCollSeq(pParse,pX->pLeft,pX->pRight);
              if( pColl==0 ) pColl = pParse->db->pDfltColl;
      
              for(j=0; pIdx->aiColumn[j]!=iColumn; j++){
                if( NEVER(j>=pIdx->nColumn) ) return 0;
              }
              if( sqlite3StrICmp(pColl->zName, pIdx->azColl[j]) ) continue;
            }
            pResult = pTerm;
            if( pTerm->prereqRight==0 ) goto findTerm_success;
          }
          if( (op&WO_EQ)!=0
           && (pTerm->eOperator & WO_EQUIV)!=0
           && nEquiv<ArraySize(aEquiv)
          ){
            pX = sqlite3ExprSkipCollate(pTerm->pExpr->pRight);
            assert( pX->op==TK_COLUMN );
            for(j=0; j<nEquiv; j+=2){
              if( aEquiv[j]==pX->iTable && aEquiv[j+1]==pX->iColumn ) break;
            }
            if( j==nEquiv ){
              aEquiv[j] = pX->iTable;
              aEquiv[j+1] = pX->iColumn;
              nEquiv += 2;
            }
          }
        }
      }
    }
    if( iEquiv>=nEquiv ) break;
    iCur = aEquiv[iEquiv++];
    iColumn = aEquiv[iEquiv++];
    op &= WO_EQ;
  }
findTerm_success:
  return pResult;
}

/* Forward reference */
static void exprAnalyze(SrcList*, WhereClause*, int);

/*
** Call exprAnalyze on all terms in a WHERE clause.  

      Bitmask b;
      b = getMask(pMaskSet, pOrTerm->leftCursor);
      if( pOrTerm->wtFlags & TERM_VIRTUAL ){
        WhereTerm *pOther = &pOrWc->a[pOrTerm->iParent];
        b |= getMask(pMaskSet, pOther->leftCursor);
      }
      indexable &= b;
      if( (pOrTerm->eOperator & WO_EQ)==0 ){
        chngToIN = 0;
      }else{
        chngToIN &= b;
      }
    }
  }

    ** will be recorded in iCursor and iColumn.  There might not be any
    ** such table and column.  Set okToChngToIN if an appropriate table
    ** and column is found but leave okToChngToIN false if not found.
    */
    for(j=0; j<2 && !okToChngToIN; j++){
      pOrTerm = pOrWc->a;
      for(i=pOrWc->nTerm-1; i>=0; i--, pOrTerm++){
        assert( pOrTerm->eOperator & WO_EQ );
        pOrTerm->wtFlags &= ~TERM_OR_OK;
        if( pOrTerm->leftCursor==iCursor ){
          /* This is the 2-bit case and we are on the second iteration and
          ** current term is from the first iteration.  So skip this term. */
          assert( j==1 );
          continue;
        }

        iCursor = pOrTerm->leftCursor;
        break;
      }
      if( i<0 ){
        /* No candidate table+column was found.  This can only occur
        ** on the second iteration */
        assert( j==1 );
        assert( IsPowerOfTwo(chngToIN) );
        assert( chngToIN==getMask(pMaskSet, iCursor) );
        break;
      }
      testcase( j==1 );

      /* We have found a candidate table and column.  Check to see if that
      ** table and column is common to every term in the OR clause */
      okToChngToIN = 1;
      for(; i>=0 && okToChngToIN; i--, pOrTerm++){
        assert( pOrTerm->eOperator & WO_EQ );
        if( pOrTerm->leftCursor!=iCursor ){
          pOrTerm->wtFlags &= ~TERM_OR_OK;
        }else if( pOrTerm->u.leftColumn!=iColumn ){
          okToChngToIN = 0;
        }else{
          int affLeft, affRight;
          /* If the right-hand side is also a column, then the affinities

      Expr *pDup;            /* A transient duplicate expression */
      ExprList *pList = 0;   /* The RHS of the IN operator */
      Expr *pLeft = 0;       /* The LHS of the IN operator */
      Expr *pNew;            /* The complete IN operator */

      for(i=pOrWc->nTerm-1, pOrTerm=pOrWc->a; i>=0; i--, pOrTerm++){
        if( (pOrTerm->wtFlags & TERM_OR_OK)==0 ) continue;
        assert( pOrTerm->eOperator & WO_EQ );
        assert( pOrTerm->leftCursor==iCursor );
        assert( pOrTerm->u.leftColumn==iColumn );
        pDup = sqlite3ExprDup(db, pOrTerm->pExpr->pRight, 0);
        pList = sqlite3ExprListAppend(pWC->pParse, pList, pDup);
        pLeft = pOrTerm->pExpr->pLeft;
      }
      assert( pLeft!=0 );

      }
      pTerm->eOperator = WO_NOOP;  /* case 1 trumps case 2 */
    }
  }
}
#endif /* !SQLITE_OMIT_OR_OPTIMIZATION && !SQLITE_OMIT_SUBQUERY */

/*
** Check to see if pExpr is an expression of the form A==B where both
** A and B are columns with the same affinity and collating sequence.
** If A and B are equivalent, return true.
*/
static int isEquivalenceExpr(Parse *pParse, Expr *pExpr){
  const CollSeq *pCLeft, *pCRight;
  if( pExpr->op!=TK_EQ ) return 0;
  if( ExprHasProperty(pExpr, EP_FromJoin) ) return 0;
  assert( sqlite3ExprSkipCollate(pExpr->pRight)->op==TK_COLUMN );
  assert( sqlite3ExprSkipCollate(pExpr->pLeft)->op==TK_COLUMN );
  if( sqlite3ExprAffinity(pExpr->pLeft)!=sqlite3ExprAffinity(pExpr->pRight) ){
    return 0;
  }
  pCLeft = sqlite3ExprCollSeq(pParse, pExpr->pLeft);
  if( pCLeft ){
    pCRight = sqlite3ExprCollSeq(pParse, pExpr->pRight);
    if( pCRight && pCRight!=pCLeft ) return 0;
  }
  return 1;
}

/*
** The input to this routine is an WhereTerm structure with only the
** "pExpr" field filled in.  The job of this routine is to analyze the
** subexpression and populate all the other fields of the WhereTerm
** structure.
**

      pTerm->leftCursor = pLeft->iTable;
      pTerm->u.leftColumn = pLeft->iColumn;
      pTerm->eOperator = operatorMask(op);
    }
    if( pRight && pRight->op==TK_COLUMN ){
      WhereTerm *pNew;
      Expr *pDup;
      u16 eExtraOp = 0;        /* Extra bits for pNew->eOperator */
      if( pTerm->leftCursor>=0 ){
        int idxNew;
        pDup = sqlite3ExprDup(db, pExpr, 0);
        if( db->mallocFailed ){
          sqlite3ExprDelete(db, pDup);
          return;
        }
        idxNew = whereClauseInsert(pWC, pDup, TERM_VIRTUAL|TERM_DYNAMIC);
        if( idxNew==0 ) return;
        pNew = &pWC->a[idxNew];
        pNew->iParent = idxTerm;
        pTerm = &pWC->a[idxTerm];
        pTerm->nChild = 1;
        pTerm->wtFlags |= TERM_COPIED;
        if( isEquivalenceExpr(pParse, pExpr) ){
          pTerm->eOperator |= WO_EQUIV;
          eExtraOp = WO_EQUIV;
        }
      }else{
        pDup = pExpr;
        pNew = pTerm;
      }
      exprCommute(pParse, pDup);
      pLeft = sqlite3ExprSkipCollate(pDup->pLeft);
      pNew->leftCursor = pLeft->iTable;
      pNew->u.leftColumn = pLeft->iColumn;
      testcase( (prereqLeft | extraRight) != prereqLeft );
      pNew->prereqRight = prereqLeft | extraRight;
      pNew->prereqAll = prereqAll;
      pNew->eOperator = operatorMask(pDup->op) + eExtraOp;
    }
  }

#ifndef SQLITE_OMIT_BETWEEN_OPTIMIZATION
  /* If a term is the BETWEEN operator, create two new virtual terms
  ** that define the range that the BETWEEN implements.  For example:
  **

  }
  if( pWC->wctrlFlags & WHERE_AND_ONLY ){
    return;
  }

  /* Search the WHERE clause terms for a usable WO_OR term. */
  for(pTerm=pWC->a; pTerm<pWCEnd; pTerm++){
    if( (pTerm->eOperator & WO_OR)!=0
     && ((pTerm->prereqAll & ~maskSrc) & p->notReady)==0
     && (pTerm->u.pOrInfo->indexable & maskSrc)!=0 
    ){
      WhereClause * const pOrWC = &pTerm->u.pOrInfo->wc;
      WhereTerm * const pOrWCEnd = &pOrWC->a[pOrWC->nTerm];
      WhereTerm *pOrTerm;
      int flags = WHERE_MULTI_OR;
      double rTotal = 0;
      double nRow = 0;
      Bitmask used = 0;
      WhereBestIdx sBOI;

      sBOI = *p;
      sBOI.pOrderBy = 0;
      sBOI.pDistinct = 0;
      sBOI.ppIdxInfo = 0;
      for(pOrTerm=pOrWC->a; pOrTerm<pOrWCEnd; pOrTerm++){
        WHERETRACE(("... Multi-index OR testing for term %d of %d....\n", 
          (pOrTerm - pOrWC->a), (pTerm - pWC->a)
        ));
        if( (pOrTerm->eOperator& WO_AND)!=0 ){
          sBOI.pWC = &pOrTerm->u.pAndInfo->wc;
          bestIndex(&sBOI);
        }else if( pOrTerm->leftCursor==iCur ){
          WhereClause tempWC;
          tempWC.pParse = pWC->pParse;
          tempWC.pMaskSet = pWC->pMaskSet;
          tempWC.pOuter = pWC;

static int termCanDriveIndex(
  WhereTerm *pTerm,              /* WHERE clause term to check */
  struct SrcList_item *pSrc,     /* Table we are trying to access */
  Bitmask notReady               /* Tables in outer loops of the join */
){
  char aff;
  if( pTerm->leftCursor!=pSrc->iCursor ) return 0;
  if( (pTerm->eOperator & WO_EQ)==0 ) return 0;
  if( (pTerm->prereqRight & notReady)!=0 ) return 0;
  aff = pSrc->pTab->aCol[pTerm->u.leftColumn].affinity;
  if( !sqlite3IndexAffinityOk(pTerm->pExpr, aff) ) return 0;
  return 1;
}
#endif

  WHERETRACE(("Recomputing index info for %s...\n", pSrc->pTab->zName));

  /* Count the number of possible WHERE clause constraints referring
  ** to this virtual table */
  for(i=nTerm=0, pTerm=pWC->a; i<pWC->nTerm; i++, pTerm++){
    if( pTerm->leftCursor != pSrc->iCursor ) continue;
    assert( IsPowerOfTwo(pTerm->eOperator & ~WO_EQUIV) );
    testcase( pTerm->eOperator & WO_IN );
    testcase( pTerm->eOperator & WO_ISNULL );
    if( pTerm->eOperator & (WO_ISNULL) ) continue;
    if( pTerm->wtFlags & TERM_VNULL ) continue;
    nTerm++;
  }

  /* If the ORDER BY clause contains only columns in the current 
  ** virtual table then allocate space for the aOrderBy part of

  *(struct sqlite3_index_orderby**)&pIdxInfo->aOrderBy = pIdxOrderBy;
  *(struct sqlite3_index_constraint_usage**)&pIdxInfo->aConstraintUsage =
                                                                   pUsage;

  for(i=j=0, pTerm=pWC->a; i<pWC->nTerm; i++, pTerm++){
    u8 op;
    if( pTerm->leftCursor != pSrc->iCursor ) continue;
    assert( IsPowerOfTwo(pTerm->eOperator & ~WO_EQUIV) );
    testcase( pTerm->eOperator & WO_IN );
    testcase( pTerm->eOperator & WO_ISNULL );
    if( pTerm->eOperator & (WO_ISNULL) ) continue;
    if( pTerm->wtFlags & TERM_VNULL ) continue;
    pIdxCons[j].iColumn = pTerm->u.leftColumn;
    pIdxCons[j].iTermOffset = i;
    op = (u8)pTerm->eOperator & WO_ALL;
    if( op==WO_IN ) op = WO_EQ;
    pIdxCons[j].op = op;
    /* The direct assignment in the previous line is possible only because
    ** the WO_ and SQLITE_INDEX_CONSTRAINT_ codes are identical.  The
    ** following asserts verify this fact. */
    assert( WO_EQ==SQLITE_INDEX_CONSTRAINT_EQ );
    assert( WO_LT==SQLITE_INDEX_CONSTRAINT_LT );

    */
    pIdxCons = *(struct sqlite3_index_constraint**)&pIdxInfo->aConstraint;
    pUsage = pIdxInfo->aConstraintUsage;
    for(i=0; i<pIdxInfo->nConstraint; i++, pIdxCons++){
      j = pIdxCons->iTermOffset;
      pTerm = &pWC->a[j];
      if( (pTerm->prereqRight&p->notReady)==0
       && (bAllowIN || (pTerm->eOperator & WO_IN)==0)
      ){
        pIdxCons->usable = 1;
      }else{
        pIdxCons->usable = 0;
      }
    }
    memset(pUsage, 0, sizeof(pUsage[0])*pIdxInfo->nConstraint);

  
    pIdxCons = *(struct sqlite3_index_constraint**)&pIdxInfo->aConstraint;
    for(i=0; i<pIdxInfo->nConstraint; i++, pIdxCons++){
      if( pUsage[i].argvIndex>0 ){
        j = pIdxCons->iTermOffset;
        pTerm = &pWC->a[j];
        p->cost.used |= pTerm->prereqRight;
        if( (pTerm->eOperator & WO_IN)!=0 && pUsage[i].omit==0 ){
          /* Do not attempt to use an IN constraint if the virtual table
          ** says that the equivalent EQ constraint cannot be safely omitted.
          ** If we do attempt to use such a constraint, some rows might be
          ** repeated in the output. */
          break;
        }
      }

    tRowcnt iUpper = p->aiRowEst[0];
    tRowcnt a[2];
    u8 aff = p->pTable->aCol[p->aiColumn[0]].affinity;

    if( pLower ){
      Expr *pExpr = pLower->pExpr->pRight;
      rc = valueFromExpr(pParse, pExpr, aff, &pRangeVal);
      assert( (pLower->eOperator & (WO_GT|WO_GE))!=0 );
      if( rc==SQLITE_OK
       && whereKeyStats(pParse, p, pRangeVal, 0, a)==SQLITE_OK
      ){
        iLower = a[0];
        if( (pLower->eOperator & WO_GT)!=0 ) iLower += a[1];
      }
      sqlite3ValueFree(pRangeVal);
    }
    if( rc==SQLITE_OK && pUpper ){
      Expr *pExpr = pUpper->pExpr->pRight;
      rc = valueFromExpr(pParse, pExpr, aff, &pRangeVal);
      assert( (pUpper->eOperator & (WO_LT|WO_LE))!=0 );
      if( rc==SQLITE_OK
       && whereKeyStats(pParse, p, pRangeVal, 1, a)==SQLITE_OK
      ){
        iUpper = a[0];
        if( (pUpper->eOperator & WO_LE)!=0 ) iUpper += a[1];
      }
      sqlite3ValueFree(pRangeVal);
    }
    if( rc==SQLITE_OK ){
      if( iUpper<=iLower ){
        *pRangeDiv = (double)p->aiRowEst[0];
      }else{

    /* If X is the column in the index and ORDER BY clause, check to see
    ** if there are any X= or X IS NULL constraints in the WHERE clause. */
    pConstraint = findTerm(p->pWC, base, iColumn, p->notReady,
                           WO_EQ|WO_ISNULL|WO_IN, pIdx);
    if( pConstraint==0 ){
      isEq = 0;
    }else if( (pConstraint->eOperator & WO_IN)!=0 ){
      /* Constraints of the form: "X IN ..." cannot be used with an ORDER BY
      ** because we do not know in what order the values on the RHS of the IN
      ** operator will occur. */
      break;
    }else if( (pConstraint->eOperator & WO_ISNULL)!=0 ){
      uniqueNotNull = 0;
      isEq = 1;  /* "X IS NULL" means X has only a single value */
    }else if( pConstraint->prereqRight==0 ){
      isEq = 1;  /* Constraint "X=constant" means X has only a single value */
    }else{
      Expr *pRight = pConstraint->pExpr->pRight;
      if( pRight->op==TK_COLUMN ){

    ** to get a better estimate on the number of rows based on
    ** VALUE and how common that value is according to the histogram.
    */
    if( pc.plan.nRow>(double)1 && pc.plan.nEq==1
     && pFirstTerm!=0 && aiRowEst[1]>1 ){
      assert( (pFirstTerm->eOperator & (WO_EQ|WO_ISNULL|WO_IN))!=0 );
      if( pFirstTerm->eOperator & (WO_EQ|WO_ISNULL) ){
        testcase( pFirstTerm->eOperator & WO_EQ );
        testcase( pFirstTerm->eOperator & WO_EQUIV );
        testcase( pFirstTerm->eOperator & WO_ISNULL );
        whereEqualScanEst(pParse, pProbe, pFirstTerm->pExpr->pRight,
                          &pc.plan.nRow);
      }else if( bInEst==0 ){
        assert( pFirstTerm->eOperator & WO_IN );
        whereInScanEst(pParse, pProbe, pFirstTerm->pExpr->x.pList,
                       &pc.plan.nRow);
      }
    }
#endif /* SQLITE_ENABLE_STAT3 */

    /* Adjust the number of output rows and downward to reflect rows

            ** set size by a factor of 3.  Indexed range constraints reduce
            ** the search space by a larger factor: 4.  We make indexed range
            ** more selective intentionally because of the subjective 
            ** observation that indexed range constraints really are more
            ** selective in practice, on average. */
            pc.plan.nRow /= 3;
          }
        }else if( (pTerm->eOperator & WO_NOOP)==0 ){
          /* Any other expression lowers the output row count by half */
          pc.plan.nRow /= 2;
        }
      }
      if( pc.plan.nRow<2 ) pc.plan.nRow = 2;
    }

    **          we reference multiple rows using a "rowid IN (...)"
    **          construct.
    */
    iReleaseReg = sqlite3GetTempReg(pParse);
    pTerm = findTerm(pWC, iCur, -1, notReady, WO_EQ|WO_IN, 0);
    assert( pTerm!=0 );
    assert( pTerm->pExpr!=0 );

    assert( omitTable==0 );
    testcase( pTerm->wtFlags & TERM_VIRTUAL ); /* EV: R-30575-11662 */
    iRowidReg = codeEqualityTerm(pParse, pTerm, pLevel, iReleaseReg);
    addrNxt = pLevel->addrNxt;
    sqlite3VdbeAddOp2(v, OP_MustBeInt, iRowidReg, addrNxt);
    sqlite3VdbeAddOp3(v, OP_NotExists, iCur, addrNxt, iRowidReg);
    sqlite3ExprCacheStore(pParse, iCur, -1, iRowidReg);

    int iRetInit;                             /* Address of regReturn init */
    int untestedTerms = 0;             /* Some terms not completely tested */
    int ii;                            /* Loop counter */
    Expr *pAndExpr = 0;                /* An ".. AND (...)" expression */
   
    pTerm = pLevel->plan.u.pTerm;
    assert( pTerm!=0 );
    assert( pTerm->eOperator & WO_OR );
    assert( (pTerm->wtFlags & TERM_ORINFO)!=0 );
    pOrWc = &pTerm->u.pOrInfo->wc;
    pLevel->op = OP_Return;
    pLevel->p1 = regReturn;

    /* Set up a new SrcList in pOrTab containing the table being scanned
    ** by this loop in the a[0] slot and all notReady tables in a[1..] slots.

      if( pAndExpr ){
        pAndExpr = sqlite3PExpr(pParse, TK_AND, 0, pAndExpr, 0);
      }
    }

    for(ii=0; ii<pOrWc->nTerm; ii++){
      WhereTerm *pOrTerm = &pOrWc->a[ii];
      if( pOrTerm->leftCursor==iCur || (pOrTerm->eOperator & WO_AND)!=0 ){
        WhereInfo *pSubWInfo;          /* Info for single OR-term scan */
        Expr *pOrExpr = pOrTerm->pExpr;
        if( pAndExpr ){
          pAndExpr->pLeft = pOrExpr;
          pOrExpr = pAndExpr;
        }
        /* Loop through table entries that match term pOrTerm. */