src/whereexpr.c

   1 /*
   2 ** 2015-06-08
   3 **
   4 ** The author disclaims copyright to this source code.  In place of
   5 ** a legal notice, here is a blessing:
   6 **
   7 **    May you do good and not evil.
   8 **    May you find forgiveness for yourself and forgive others.
   9 **    May you share freely, never taking more than you give.
  10 **
  11 *************************************************************************
  12 ** This module contains C code that generates VDBE code used to process
  13 ** the WHERE clause of SQL statements.
  14 **
  15 ** This file was originally part of where.c but was split out to improve
  16 ** readability and editabiliity.  This file contains utility routines for
  17 ** analyzing Expr objects in the WHERE clause.
  18 */
  19 #include "sqliteInt.h"
  20 #include "whereInt.h"
  21
  22 /* Forward declarations */
  23 static void exprAnalyze(SrcList*, WhereClause*, int);
  24
  25 /*
  26 ** Deallocate all memory associated with a WhereOrInfo object.
  27 */
  28 static void whereOrInfoDelete(sqlite3 *db, WhereOrInfo *p){
  29   sqlite3WhereClauseClear(&p->wc);
  30   sqlite3DbFree(db, p);
  31 }
  32
  33 /*
  34 ** Deallocate all memory associated with a WhereAndInfo object.
  35 */
  36 static void whereAndInfoDelete(sqlite3 *db, WhereAndInfo *p){
  37   sqlite3WhereClauseClear(&p->wc);
  38   sqlite3DbFree(db, p);
  39 }
  40
  41 /*
  42 ** Add a single new WhereTerm entry to the WhereClause object pWC.
  43 ** The new WhereTerm object is constructed from Expr p and with wtFlags.
  44 ** The index in pWC->a[] of the new WhereTerm is returned on success.
  45 ** 0 is returned if the new WhereTerm could not be added due to a memory
  46 ** allocation error.  The memory allocation failure will be recorded in
  47 ** the db->mallocFailed flag so that higher-level functions can detect it.
  48 **
  49 ** This routine will increase the size of the pWC->a[] array as necessary.
  50 **
  51 ** If the wtFlags argument includes TERM_DYNAMIC, then responsibility
  52 ** for freeing the expression p is assumed by the WhereClause object pWC.
  53 ** This is true even if this routine fails to allocate a new WhereTerm.
  54 **
  55 ** WARNING:  This routine might reallocate the space used to store
  56 ** WhereTerms.  All pointers to WhereTerms should be invalidated after
  57 ** calling this routine.  Such pointers may be reinitialized by referencing
  58 ** the pWC->a[] array.
  59 */
  60 static int whereClauseInsert(WhereClause *pWC, Expr *p, u16 wtFlags){
  61   WhereTerm *pTerm;
  62   int idx;
  63   testcase( wtFlags & TERM_VIRTUAL );
  64   if( pWC->nTerm>=pWC->nSlot ){
  65     WhereTerm *pOld = pWC->a;
  66     sqlite3 *db = pWC->pWInfo->pParse->db;
  67     pWC->a = sqlite3DbMallocRawNN(db, sizeof(pWC->a[0])*pWC->nSlot*2 );
  68     if( pWC->a==0 ){
  69       if( wtFlags & TERM_DYNAMIC ){
  70         sqlite3ExprDelete(db, p);
  71       }
  72       pWC->a = pOld;
  73       return 0;
  74     }
  75     memcpy(pWC->a, pOld, sizeof(pWC->a[0])*pWC->nTerm);
  76     if( pOld!=pWC->aStatic ){
  77       sqlite3DbFree(db, pOld);
  78     }
  79     pWC->nSlot = sqlite3DbMallocSize(db, pWC->a)/sizeof(pWC->a[0]);
  80   }
  81   pTerm = &pWC->a[idx = pWC->nTerm++];
  82   if( p && ExprHasProperty(p, EP_Unlikely) ){
  83     pTerm->truthProb = sqlite3LogEst(p->iTable) - 270;
  84   }else{
  85     pTerm->truthProb = 1;
  86   }
  87   pTerm->pExpr = sqlite3ExprSkipCollate(p);
  88   pTerm->wtFlags = wtFlags;
  89   pTerm->pWC = pWC;
  90   pTerm->iParent = -1;
  91   memset(&pTerm->eOperator, 0,
  92          sizeof(WhereTerm) - offsetof(WhereTerm,eOperator));
  93   return idx;
  94 }
  95
  96 /*
  97 ** Return TRUE if the given operator is one of the operators that is
  98 ** allowed for an indexable WHERE clause term.  The allowed operators are
  99 ** "=", "<", ">", "<=", ">=", "IN", "IS", and "IS NULL"
 100 */
 101 static int allowedOp(int op){
 102   assert( TK_GT>TK_EQ && TK_GT<TK_GE );
 103   assert( TK_LT>TK_EQ && TK_LT<TK_GE );
 104   assert( TK_LE>TK_EQ && TK_LE<TK_GE );
 105   assert( TK_GE==TK_EQ+4 );
 106   return op==TK_IN || (op>=TK_EQ && op<=TK_GE) || op==TK_ISNULL || op==TK_IS;
 107 }
 108
 109 /*
 110 ** Commute a comparison operator.  Expressions of the form "X op Y"
 111 ** are converted into "Y op X".
 112 **
 113 ** If left/right precedence rules come into play when determining the
 114 ** collating sequence, then COLLATE operators are adjusted to ensure
 115 ** that the collating sequence does not change.  For example:
 116 ** "Y collate NOCASE op X" becomes "X op Y" because any collation sequence on
 117 ** the left hand side of a comparison overrides any collation sequence
 118 ** attached to the right. For the same reason the EP_Collate flag
 119 ** is not commuted.
 120 */
 121 static void exprCommute(Parse *pParse, Expr *pExpr){
 122   u16 expRight = (pExpr->pRight->flags & EP_Collate);
 123   u16 expLeft = (pExpr->pLeft->flags & EP_Collate);
 124   assert( allowedOp(pExpr->op) && pExpr->op!=TK_IN );
 125   if( expRight==expLeft ){
 126     /* Either X and Y both have COLLATE operator or neither do */
 127     if( expRight ){
 128       /* Both X and Y have COLLATE operators.  Make sure X is always
 129       ** used by clearing the EP_Collate flag from Y. */
 130       pExpr->pRight->flags &= ~EP_Collate;
 131     }else if( sqlite3ExprCollSeq(pParse, pExpr->pLeft)!=0 ){
 132       /* Neither X nor Y have COLLATE operators, but X has a non-default
 133       ** collating sequence.  So add the EP_Collate marker on X to cause
 134       ** it to be searched first. */
 135       pExpr->pLeft->flags |= EP_Collate;
 136     }
 137   }
 138   SWAP(Expr*,pExpr->pRight,pExpr->pLeft);
 139   if( pExpr->op>=TK_GT ){
 140     assert( TK_LT==TK_GT+2 );
 141     assert( TK_GE==TK_LE+2 );
 142     assert( TK_GT>TK_EQ );
 143     assert( TK_GT<TK_LE );
 144     assert( pExpr->op>=TK_GT && pExpr->op<=TK_GE );
 145     pExpr->op = ((pExpr->op-TK_GT)^2)+TK_GT;
 146   }
 147 }
 148
 149 /*
 150 ** Translate from TK_xx operator to WO_xx bitmask.
 151 */
 152 static u16 operatorMask(int op){
 153   u16 c;
 154   assert( allowedOp(op) );
 155   if( op==TK_IN ){
 156     c = WO_IN;
 157   }else if( op==TK_ISNULL ){
 158     c = WO_ISNULL;
 159   }else if( op==TK_IS ){
 160     c = WO_IS;
 161   }else{
 162     assert( (WO_EQ<<(op-TK_EQ)) < 0x7fff );
 163     c = (u16)(WO_EQ<<(op-TK_EQ));
 164   }
 165   assert( op!=TK_ISNULL || c==WO_ISNULL );
 166   assert( op!=TK_IN || c==WO_IN );
 167   assert( op!=TK_EQ || c==WO_EQ );
 168   assert( op!=TK_LT || c==WO_LT );
 169   assert( op!=TK_LE || c==WO_LE );
 170   assert( op!=TK_GT || c==WO_GT );
 171   assert( op!=TK_GE || c==WO_GE );
 172   assert( op!=TK_IS || c==WO_IS );
 173   return c;
 174 }
 175
 176
 177 #ifndef SQLITE_OMIT_LIKE_OPTIMIZATION
 178 /*
 179 ** Check to see if the given expression is a LIKE or GLOB operator that
 180 ** can be optimized using inequality constraints.  Return TRUE if it is
 181 ** so and false if not.
 182 **
 183 ** In order for the operator to be optimizible, the RHS must be a string
 184 ** literal that does not begin with a wildcard.  The LHS must be a column
 185 ** that may only be NULL, a string, or a BLOB, never a number. (This means
 186 ** that virtual tables cannot participate in the LIKE optimization.)  The
 187 ** collating sequence for the column on the LHS must be appropriate for
 188 ** the operator.
 189 */
 190 static int isLikeOrGlob(
 191   Parse *pParse,    /* Parsing and code generating context */
 192   Expr *pExpr,      /* Test this expression */
 193   Expr **ppPrefix,  /* Pointer to TK_STRING expression with pattern prefix */
 194   int *pisComplete, /* True if the only wildcard is % in the last character */
 195   int *pnoCase      /* True if uppercase is equivalent to lowercase */
 196 ){
 197   const char *z = 0;         /* String on RHS of LIKE operator */
 198   Expr *pRight, *pLeft;      /* Right and left size of LIKE operator */
 199   ExprList *pList;           /* List of operands to the LIKE operator */
 200   int c;                     /* One character in z[] */
 201   int cnt;                   /* Number of non-wildcard prefix characters */
 202   char wc[3];                /* Wildcard characters */
 203   sqlite3 *db = pParse->db;  /* Database connection */
 204   sqlite3_value *pVal = 0;
 205   int op;                    /* Opcode of pRight */
 206   int rc;                    /* Result code to return */
 207
 208   if( !sqlite3IsLikeFunction(db, pExpr, pnoCase, wc) ){
 209     return 0;
 210   }
 211 #ifdef SQLITE_EBCDIC
 212   if( *pnoCase ) return 0;
 213 #endif
 214   pList = pExpr->x.pList;
 215   pLeft = pList->a[1].pExpr;
 216
 217   pRight = sqlite3ExprSkipCollate(pList->a[0].pExpr);
 218   op = pRight->op;
 219   if( op==TK_VARIABLE && (db->flags & SQLITE_EnableQPSG)==0 ){
 220     Vdbe *pReprepare = pParse->pReprepare;
 221     int iCol = pRight->iColumn;
 222     pVal = sqlite3VdbeGetBoundValue(pReprepare, iCol, SQLITE_AFF_BLOB);
 223     if( pVal && sqlite3_value_type(pVal)==SQLITE_TEXT ){
 224       z = (char *)sqlite3_value_text(pVal);
 225     }
 226     sqlite3VdbeSetVarmask(pParse->pVdbe, iCol);
 227     assert( pRight->op==TK_VARIABLE || pRight->op==TK_REGISTER );
 228   }else if( op==TK_STRING ){
 229     z = pRight->u.zToken;
 230   }
 231   if( z ){
 232
 233     /* If the RHS begins with a digit or a minus sign, then the LHS must
 234     ** be an ordinary column (not a virtual table column) with TEXT affinity.
 235     ** Otherwise the LHS might be numeric and "lhs >= rhs" would be false
 236     ** even though "lhs LIKE rhs" is true.  But if the RHS does not start
 237     ** with a digit or '-', then "lhs LIKE rhs" will always be false if
 238     ** the LHS is numeric and so the optimization still works.
 239     */
 240     if( sqlite3Isdigit(z[0]) || z[0]=='-' ){
 241       if( pLeft->op!=TK_COLUMN
 242        || sqlite3ExprAffinity(pLeft)!=SQLITE_AFF_TEXT
 243        || IsVirtual(pLeft->pTab)  /* Value might be numeric */
 244       ){
 245         sqlite3ValueFree(pVal);
 246         return 0;
 247       }
 248     }
 249     cnt = 0;
 250     while( (c=z[cnt])!=0 && c!=wc[0] && c!=wc[1] && c!=wc[2] ){
 251       cnt++;
 252     }
 253     if( cnt!=0 && 255!=(u8)z[cnt-1] ){
 254       Expr *pPrefix;
 255       *pisComplete = c==wc[0] && z[cnt+1]==0;
 256       pPrefix = sqlite3Expr(db, TK_STRING, z);
 257       if( pPrefix ) pPrefix->u.zToken[cnt] = 0;
 258       *ppPrefix = pPrefix;
 259       if( op==TK_VARIABLE ){
 260         Vdbe *v = pParse->pVdbe;
 261         sqlite3VdbeSetVarmask(v, pRight->iColumn);
 262         if( *pisComplete && pRight->u.zToken[1] ){
 263           /* If the rhs of the LIKE expression is a variable, and the current
 264           ** value of the variable means there is no need to invoke the LIKE
 265           ** function, then no OP_Variable will be added to the program.
 266           ** This causes problems for the sqlite3_bind_parameter_name()
 267           ** API. To work around them, add a dummy OP_Variable here.
 268           */
 269           int r1 = sqlite3GetTempReg(pParse);
 270           sqlite3ExprCodeTarget(pParse, pRight, r1);
 271           sqlite3VdbeChangeP3(v, sqlite3VdbeCurrentAddr(v)-1, 0);
 272           sqlite3ReleaseTempReg(pParse, r1);
 273         }
 274       }
 275     }else{
 276       z = 0;
 277     }
 278   }
 279
 280   rc = (z!=0);
 281   sqlite3ValueFree(pVal);
 282   return rc;
 283 }
 284 #endif /* SQLITE_OMIT_LIKE_OPTIMIZATION */
 285
 286
 287 #ifndef SQLITE_OMIT_VIRTUALTABLE
 288 /*
 289 ** Check to see if the given expression is of the form
 290 **
 291 **         column OP expr
 292 **
 293 ** where OP is one of MATCH, GLOB, LIKE or REGEXP and "column" is a
 294 ** column of a virtual table.
 295 **
 296 ** If it is then return TRUE.  If not, return FALSE.
 297 */
 298 static int isMatchOfColumn(
 299   Expr *pExpr,                    /* Test this expression */
 300   unsigned char *peOp2            /* OUT: 0 for MATCH, or else an op2 value */
 301 ){
 302   static const struct Op2 {
 303     const char *zOp;
 304     unsigned char eOp2;
 305   } aOp[] = {
 306     { "match",  SQLITE_INDEX_CONSTRAINT_MATCH },
 307     { "glob",   SQLITE_INDEX_CONSTRAINT_GLOB },
 308     { "like",   SQLITE_INDEX_CONSTRAINT_LIKE },
 309     { "regexp", SQLITE_INDEX_CONSTRAINT_REGEXP }
 310   };
 311   ExprList *pList;
 312   Expr *pCol;                     /* Column reference */
 313   int i;
 314
 315   if( pExpr->op!=TK_FUNCTION ){
 316     return 0;
 317   }
 318   pList = pExpr->x.pList;
 319   if( pList==0 || pList->nExpr!=2 ){
 320     return 0;
 321   }
 322   pCol = pList->a[1].pExpr;
 323   if( pCol->op!=TK_COLUMN || !IsVirtual(pCol->pTab) ){
 324     return 0;
 325   }
 326   for(i=0; i<ArraySize(aOp); i++){
 327     if( sqlite3StrICmp(pExpr->u.zToken, aOp[i].zOp)==0 ){
 328       *peOp2 = aOp[i].eOp2;
 329       return 1;
 330     }
 331   }
 332   return 0;
 333 }
 334 #endif /* SQLITE_OMIT_VIRTUALTABLE */
 335
 336 /*
 337 ** If the pBase expression originated in the ON or USING clause of
 338 ** a join, then transfer the appropriate markings over to derived.
 339 */
 340 static void transferJoinMarkings(Expr *pDerived, Expr *pBase){
 341   if( pDerived ){
 342     pDerived->flags |= pBase->flags & EP_FromJoin;
 343     pDerived->iRightJoinTable = pBase->iRightJoinTable;
 344   }
 345 }
 346
 347 /*
 348 ** Mark term iChild as being a child of term iParent
 349 */
 350 static void markTermAsChild(WhereClause *pWC, int iChild, int iParent){
 351   pWC->a[iChild].iParent = iParent;
 352   pWC->a[iChild].truthProb = pWC->a[iParent].truthProb;
 353   pWC->a[iParent].nChild++;
 354 }
 355
 356 /*
 357 ** Return the N-th AND-connected subterm of pTerm.  Or if pTerm is not
 358 ** a conjunction, then return just pTerm when N==0.  If N is exceeds
 359 ** the number of available subterms, return NULL.
 360 */
 361 static WhereTerm *whereNthSubterm(WhereTerm *pTerm, int N){
 362   if( pTerm->eOperator!=WO_AND ){
 363     return N==0 ? pTerm : 0;
 364   }
 365   if( N<pTerm->u.pAndInfo->wc.nTerm ){
 366     return &pTerm->u.pAndInfo->wc.a[N];
 367   }
 368   return 0;
 369 }
 370
 371 /*
 372 ** Subterms pOne and pTwo are contained within WHERE clause pWC.  The
 373 ** two subterms are in disjunction - they are OR-ed together.
 374 **
 375 ** If these two terms are both of the form:  "A op B" with the same
 376 ** A and B values but different operators and if the operators are
 377 ** compatible (if one is = and the other is <, for example) then
 378 ** add a new virtual AND term to pWC that is the combination of the
 379 ** two.
 380 **
 381 ** Some examples:
 382 **
 383 **    x<y OR x=y    -->     x<=y
 384 **    x=y OR x=y    -->     x=y
 385 **    x<=y OR x<y   -->     x<=y
 386 **
 387 ** The following is NOT generated:
 388 **
 389 **    x<y OR x>y    -->     x!=y
 390 */
 391 static void whereCombineDisjuncts(
 392   SrcList *pSrc,         /* the FROM clause */
 393   WhereClause *pWC,      /* The complete WHERE clause */
 394   WhereTerm *pOne,       /* First disjunct */
 395   WhereTerm *pTwo        /* Second disjunct */
 396 ){
 397   u16 eOp = pOne->eOperator | pTwo->eOperator;
 398   sqlite3 *db;           /* Database connection (for malloc) */
 399   Expr *pNew;            /* New virtual expression */
 400   int op;                /* Operator for the combined expression */
 401   int idxNew;            /* Index in pWC of the next virtual term */
 402
 403   if( (pOne->eOperator & (WO_EQ|WO_LT|WO_LE|WO_GT|WO_GE))==0 ) return;
 404   if( (pTwo->eOperator & (WO_EQ|WO_LT|WO_LE|WO_GT|WO_GE))==0 ) return;
 405   if( (eOp & (WO_EQ|WO_LT|WO_LE))!=eOp
 406    && (eOp & (WO_EQ|WO_GT|WO_GE))!=eOp ) return;
 407   assert( pOne->pExpr->pLeft!=0 && pOne->pExpr->pRight!=0 );
 408   assert( pTwo->pExpr->pLeft!=0 && pTwo->pExpr->pRight!=0 );
 409   if( sqlite3ExprCompare(0,pOne->pExpr->pLeft, pTwo->pExpr->pLeft, -1) ) return;
 410   if( sqlite3ExprCompare(0,pOne->pExpr->pRight, pTwo->pExpr->pRight,-1) )return;
 411   /* If we reach this point, it means the two subterms can be combined */
 412   if( (eOp & (eOp-1))!=0 ){
 413     if( eOp & (WO_LT|WO_LE) ){
 414       eOp = WO_LE;
 415     }else{
 416       assert( eOp & (WO_GT|WO_GE) );
 417       eOp = WO_GE;
 418     }
 419   }
 420   db = pWC->pWInfo->pParse->db;
 421   pNew = sqlite3ExprDup(db, pOne->pExpr, 0);
 422   if( pNew==0 ) return;
 423   for(op=TK_EQ; eOp!=(WO_EQ<<(op-TK_EQ)); op++){ assert( op<TK_GE ); }
 424   pNew->op = op;
 425   idxNew = whereClauseInsert(pWC, pNew, TERM_VIRTUAL|TERM_DYNAMIC);
 426   exprAnalyze(pSrc, pWC, idxNew);
 427 }
 428
 429 #if !defined(SQLITE_OMIT_OR_OPTIMIZATION) && !defined(SQLITE_OMIT_SUBQUERY)
 430 /*
 431 ** Analyze a term that consists of two or more OR-connected
 432 ** subterms.  So in:
 433 **
 434 **     ... WHERE  (a=5) AND (b=7 OR c=9 OR d=13) AND (d=13)
 435 **                          ^^^^^^^^^^^^^^^^^^^^
 436 **
 437 ** This routine analyzes terms such as the middle term in the above example.
 438 ** A WhereOrTerm object is computed and attached to the term under
 439 ** analysis, regardless of the outcome of the analysis.  Hence:
 440 **
 441 **     WhereTerm.wtFlags   |=  TERM_ORINFO
 442 **     WhereTerm.u.pOrInfo  =  a dynamically allocated WhereOrTerm object
 443 **
 444 ** The term being analyzed must have two or more of OR-connected subterms.
 445 ** A single subterm might be a set of AND-connected sub-subterms.
 446 ** Examples of terms under analysis:
 447 **
 448 **     (A)     t1.x=t2.y OR t1.x=t2.z OR t1.y=15 OR t1.z=t3.a+5
 449 **     (B)     x=expr1 OR expr2=x OR x=expr3
 450 **     (C)     t1.x=t2.y OR (t1.x=t2.z AND t1.y=15)
 451 **     (D)     x=expr1 OR (y>11 AND y<22 AND z LIKE '*hello*')
 452 **     (E)     (p.a=1 AND q.b=2 AND r.c=3) OR (p.x=4 AND q.y=5 AND r.z=6)
 453 **     (F)     x>A OR (x=A AND y>=B)
 454 **
 455 ** CASE 1:
 456 **
 457 ** If all subterms are of the form T.C=expr for some single column of C and
 458 ** a single table T (as shown in example B above) then create a new virtual
 459 ** term that is an equivalent IN expression.  In other words, if the term
 460 ** being analyzed is:
 461 **
 462 **      x = expr1  OR  expr2 = x  OR  x = expr3
 463 **
 464 ** then create a new virtual term like this:
 465 **
 466 **      x IN (expr1,expr2,expr3)
 467 **
 468 ** CASE 2:
 469 **
 470 ** If there are exactly two disjuncts and one side has x>A and the other side
 471 ** has x=A (for the same x and A) then add a new virtual conjunct term to the
 472 ** WHERE clause of the form "x>=A".  Example:
 473 **
 474 **      x>A OR (x=A AND y>B)    adds:    x>=A
 475 **
 476 ** The added conjunct can sometimes be helpful in query planning.
 477 **
 478 ** CASE 3:
 479 **
 480 ** If all subterms are indexable by a single table T, then set
 481 **
 482 **     WhereTerm.eOperator              =  WO_OR
 483 **     WhereTerm.u.pOrInfo->indexable  |=  the cursor number for table T
 484 **
 485 ** A subterm is "indexable" if it is of the form
 486 ** "T.C <op> <expr>" where C is any column of table T and
 487 ** <op> is one of "=", "<", "<=", ">", ">=", "IS NULL", or "IN".
 488 ** A subterm is also indexable if it is an AND of two or more
 489 ** subsubterms at least one of which is indexable.  Indexable AND
 490 ** subterms have their eOperator set to WO_AND and they have
 491 ** u.pAndInfo set to a dynamically allocated WhereAndTerm object.
 492 **
 493 ** From another point of view, "indexable" means that the subterm could
 494 ** potentially be used with an index if an appropriate index exists.
 495 ** This analysis does not consider whether or not the index exists; that
 496 ** is decided elsewhere.  This analysis only looks at whether subterms
 497 ** appropriate for indexing exist.
 498 **
 499 ** All examples A through E above satisfy case 3.  But if a term
 500 ** also satisfies case 1 (such as B) we know that the optimizer will
 501 ** always prefer case 1, so in that case we pretend that case 3 is not
 502 ** satisfied.
 503 **
 504 ** It might be the case that multiple tables are indexable.  For example,
 505 ** (E) above is indexable on tables P, Q, and R.
 506 **
 507 ** Terms that satisfy case 3 are candidates for lookup by using
 508 ** separate indices to find rowids for each subterm and composing
 509 ** the union of all rowids using a RowSet object.  This is similar
 510 ** to "bitmap indices" in other database engines.
 511 **
 512 ** OTHERWISE:
 513 **
 514 ** If none of cases 1, 2, or 3 apply, then leave the eOperator set to
 515 ** zero.  This term is not useful for search.
 516 */
 517 static void exprAnalyzeOrTerm(
 518   SrcList *pSrc,            /* the FROM clause */
 519   WhereClause *pWC,         /* the complete WHERE clause */
 520   int idxTerm               /* Index of the OR-term to be analyzed */
 521 ){
 522   WhereInfo *pWInfo = pWC->pWInfo;        /* WHERE clause processing context */
 523   Parse *pParse = pWInfo->pParse;         /* Parser context */
 524   sqlite3 *db = pParse->db;               /* Database connection */
 525   WhereTerm *pTerm = &pWC->a[idxTerm];    /* The term to be analyzed */
 526   Expr *pExpr = pTerm->pExpr;             /* The expression of the term */
 527   int i;                                  /* Loop counters */
 528   WhereClause *pOrWc;       /* Breakup of pTerm into subterms */
 529   WhereTerm *pOrTerm;       /* A Sub-term within the pOrWc */
 530   WhereOrInfo *pOrInfo;     /* Additional information associated with pTerm */
 531   Bitmask chngToIN;         /* Tables that might satisfy case 1 */
 532   Bitmask indexable;        /* Tables that are indexable, satisfying case 2 */
 533
 534   /*
 535   ** Break the OR clause into its separate subterms.  The subterms are
 536   ** stored in a WhereClause structure containing within the WhereOrInfo
 537   ** object that is attached to the original OR clause term.
 538   */
 539   assert( (pTerm->wtFlags & (TERM_DYNAMIC|TERM_ORINFO|TERM_ANDINFO))==0 );
 540   assert( pExpr->op==TK_OR );
 541   pTerm->u.pOrInfo = pOrInfo = sqlite3DbMallocZero(db, sizeof(*pOrInfo));
 542   if( pOrInfo==0 ) return;
 543   pTerm->wtFlags |= TERM_ORINFO;
 544   pOrWc = &pOrInfo->wc;
 545   memset(pOrWc->aStatic, 0, sizeof(pOrWc->aStatic));
 546   sqlite3WhereClauseInit(pOrWc, pWInfo);
 547   sqlite3WhereSplit(pOrWc, pExpr, TK_OR);
 548   sqlite3WhereExprAnalyze(pSrc, pOrWc);
 549   if( db->mallocFailed ) return;
 550   assert( pOrWc->nTerm>=2 );
 551
 552   /*
 553   ** Compute the set of tables that might satisfy cases 1 or 3.
 554   */
 555   indexable = ~(Bitmask)0;
 556   chngToIN = ~(Bitmask)0;
 557   for(i=pOrWc->nTerm-1, pOrTerm=pOrWc->a; i>=0 && indexable; i--, pOrTerm++){
 558     if( (pOrTerm->eOperator & WO_SINGLE)==0 ){
 559       WhereAndInfo *pAndInfo;
 560       assert( (pOrTerm->wtFlags & (TERM_ANDINFO|TERM_ORINFO))==0 );
 561       chngToIN = 0;
 562       pAndInfo = sqlite3DbMallocRawNN(db, sizeof(*pAndInfo));
 563       if( pAndInfo ){
 564         WhereClause *pAndWC;
 565         WhereTerm *pAndTerm;
 566         int j;
 567         Bitmask b = 0;
 568         pOrTerm->u.pAndInfo = pAndInfo;
 569         pOrTerm->wtFlags |= TERM_ANDINFO;
 570         pOrTerm->eOperator = WO_AND;
 571         pAndWC = &pAndInfo->wc;
 572         memset(pAndWC->aStatic, 0, sizeof(pAndWC->aStatic));
 573         sqlite3WhereClauseInit(pAndWC, pWC->pWInfo);
 574         sqlite3WhereSplit(pAndWC, pOrTerm->pExpr, TK_AND);
 575         sqlite3WhereExprAnalyze(pSrc, pAndWC);
 576         pAndWC->pOuter = pWC;
 577         if( !db->mallocFailed ){
 578           for(j=0, pAndTerm=pAndWC->a; j<pAndWC->nTerm; j++, pAndTerm++){
 579             assert( pAndTerm->pExpr );
 580             if( allowedOp(pAndTerm->pExpr->op)
 581              || pAndTerm->eOperator==WO_MATCH
 582             ){
 583               b |= sqlite3WhereGetMask(&pWInfo->sMaskSet, pAndTerm->leftCursor);
 584             }
 585           }
 586         }
 587         indexable &= b;
 588       }
 589     }else if( pOrTerm->wtFlags & TERM_COPIED ){
 590       /* Skip this term for now.  We revisit it when we process the
 591       ** corresponding TERM_VIRTUAL term */
 592     }else{
 593       Bitmask b;
 594       b = sqlite3WhereGetMask(&pWInfo->sMaskSet, pOrTerm->leftCursor);
 595       if( pOrTerm->wtFlags & TERM_VIRTUAL ){
 596         WhereTerm *pOther = &pOrWc->a[pOrTerm->iParent];
 597         b |= sqlite3WhereGetMask(&pWInfo->sMaskSet, pOther->leftCursor);
 598       }
 599       indexable &= b;
 600       if( (pOrTerm->eOperator & WO_EQ)==0 ){
 601         chngToIN = 0;
 602       }else{
 603         chngToIN &= b;
 604       }
 605     }
 606   }
 607
 608   /*
 609   ** Record the set of tables that satisfy case 3.  The set might be
 610   ** empty.
 611   */
 612   pOrInfo->indexable = indexable;
 613   pTerm->eOperator = indexable==0 ? 0 : WO_OR;
 614
 615   /* For a two-way OR, attempt to implementation case 2.
 616   */
 617   if( indexable && pOrWc->nTerm==2 ){
 618     int iOne = 0;
 619     WhereTerm *pOne;
 620     while( (pOne = whereNthSubterm(&pOrWc->a[0],iOne++))!=0 ){
 621       int iTwo = 0;
 622       WhereTerm *pTwo;
 623       while( (pTwo = whereNthSubterm(&pOrWc->a[1],iTwo++))!=0 ){
 624         whereCombineDisjuncts(pSrc, pWC, pOne, pTwo);
 625       }
 626     }
 627   }
 628
 629   /*
 630   ** chngToIN holds a set of tables that *might* satisfy case 1.  But
 631   ** we have to do some additional checking to see if case 1 really
 632   ** is satisfied.
 633   **
 634   ** chngToIN will hold either 0, 1, or 2 bits.  The 0-bit case means
 635   ** that there is no possibility of transforming the OR clause into an
 636   ** IN operator because one or more terms in the OR clause contain
 637   ** something other than == on a column in the single table.  The 1-bit
 638   ** case means that every term of the OR clause is of the form
 639   ** "table.column=expr" for some single table.  The one bit that is set
 640   ** will correspond to the common table.  We still need to check to make
 641   ** sure the same column is used on all terms.  The 2-bit case is when
 642   ** the all terms are of the form "table1.column=table2.column".  It
 643   ** might be possible to form an IN operator with either table1.column
 644   ** or table2.column as the LHS if either is common to every term of
 645   ** the OR clause.
 646   **
 647   ** Note that terms of the form "table.column1=table.column2" (the
 648   ** same table on both sizes of the ==) cannot be optimized.
 649   */
 650   if( chngToIN ){
 651     int okToChngToIN = 0;     /* True if the conversion to IN is valid */
 652     int iColumn = -1;         /* Column index on lhs of IN operator */
 653     int iCursor = -1;         /* Table cursor common to all terms */
 654     int j = 0;                /* Loop counter */
 655
 656     /* Search for a table and column that appears on one side or the
 657     ** other of the == operator in every subterm.  That table and column
 658     ** will be recorded in iCursor and iColumn.  There might not be any
 659     ** such table and column.  Set okToChngToIN if an appropriate table
 660     ** and column is found but leave okToChngToIN false if not found.
 661     */
 662     for(j=0; j<2 && !okToChngToIN; j++){
 663       pOrTerm = pOrWc->a;
 664       for(i=pOrWc->nTerm-1; i>=0; i--, pOrTerm++){
 665         assert( pOrTerm->eOperator & WO_EQ );
 666         pOrTerm->wtFlags &= ~TERM_OR_OK;
 667         if( pOrTerm->leftCursor==iCursor ){
 668           /* This is the 2-bit case and we are on the second iteration and
 669           ** current term is from the first iteration.  So skip this term. */
 670           assert( j==1 );
 671           continue;
 672         }
 673         if( (chngToIN & sqlite3WhereGetMask(&pWInfo->sMaskSet,
 674                                             pOrTerm->leftCursor))==0 ){
 675           /* This term must be of the form t1.a==t2.b where t2 is in the
 676           ** chngToIN set but t1 is not.  This term will be either preceded
 677           ** or follwed by an inverted copy (t2.b==t1.a).  Skip this term
 678           ** and use its inversion. */
 679           testcase( pOrTerm->wtFlags & TERM_COPIED );
 680           testcase( pOrTerm->wtFlags & TERM_VIRTUAL );
 681           assert( pOrTerm->wtFlags & (TERM_COPIED|TERM_VIRTUAL) );
 682           continue;
 683         }
 684         iColumn = pOrTerm->u.leftColumn;
 685         iCursor = pOrTerm->leftCursor;
 686         break;
 687       }
 688       if( i<0 ){
 689         /* No candidate table+column was found.  This can only occur
 690         ** on the second iteration */
 691         assert( j==1 );
 692         assert( IsPowerOfTwo(chngToIN) );
 693         assert( chngToIN==sqlite3WhereGetMask(&pWInfo->sMaskSet, iCursor) );
 694         break;
 695       }
 696       testcase( j==1 );
 697
 698       /* We have found a candidate table and column.  Check to see if that
 699       ** table and column is common to every term in the OR clause */
 700       okToChngToIN = 1;
 701       for(; i>=0 && okToChngToIN; i--, pOrTerm++){
 702         assert( pOrTerm->eOperator & WO_EQ );
 703         if( pOrTerm->leftCursor!=iCursor ){
 704           pOrTerm->wtFlags &= ~TERM_OR_OK;
 705         }else if( pOrTerm->u.leftColumn!=iColumn ){
 706           okToChngToIN = 0;
 707         }else{
 708           int affLeft, affRight;
 709           /* If the right-hand side is also a column, then the affinities
 710           ** of both right and left sides must be such that no type
 711           ** conversions are required on the right.  (Ticket #2249)
 712           */
 713           affRight = sqlite3ExprAffinity(pOrTerm->pExpr->pRight);
 714           affLeft = sqlite3ExprAffinity(pOrTerm->pExpr->pLeft);
 715           if( affRight!=0 && affRight!=affLeft ){
 716             okToChngToIN = 0;
 717           }else{
 718             pOrTerm->wtFlags |= TERM_OR_OK;
 719           }
 720         }
 721       }
 722     }
 723
 724     /* At this point, okToChngToIN is true if original pTerm satisfies
 725     ** case 1.  In that case, construct a new virtual term that is
 726     ** pTerm converted into an IN operator.
 727     */
 728     if( okToChngToIN ){
 729       Expr *pDup;            /* A transient duplicate expression */
 730       ExprList *pList = 0;   /* The RHS of the IN operator */
 731       Expr *pLeft = 0;       /* The LHS of the IN operator */
 732       Expr *pNew;            /* The complete IN operator */
 733
 734       for(i=pOrWc->nTerm-1, pOrTerm=pOrWc->a; i>=0; i--, pOrTerm++){
 735         if( (pOrTerm->wtFlags & TERM_OR_OK)==0 ) continue;
 736         assert( pOrTerm->eOperator & WO_EQ );
 737         assert( pOrTerm->leftCursor==iCursor );
 738         assert( pOrTerm->u.leftColumn==iColumn );
 739         pDup = sqlite3ExprDup(db, pOrTerm->pExpr->pRight, 0);
 740         pList = sqlite3ExprListAppend(pWInfo->pParse, pList, pDup);
 741         pLeft = pOrTerm->pExpr->pLeft;
 742       }
 743       assert( pLeft!=0 );
 744       pDup = sqlite3ExprDup(db, pLeft, 0);
 745       pNew = sqlite3PExpr(pParse, TK_IN, pDup, 0);
 746       if( pNew ){
 747         int idxNew;
 748         transferJoinMarkings(pNew, pExpr);
 749         assert( !ExprHasProperty(pNew, EP_xIsSelect) );
 750         pNew->x.pList = pList;
 751         idxNew = whereClauseInsert(pWC, pNew, TERM_VIRTUAL|TERM_DYNAMIC);
 752         testcase( idxNew==0 );
 753         exprAnalyze(pSrc, pWC, idxNew);
 754         pTerm = &pWC->a[idxTerm];
 755         markTermAsChild(pWC, idxNew, idxTerm);
 756       }else{
 757         sqlite3ExprListDelete(db, pList);
 758       }
 759       pTerm->eOperator = WO_NOOP;  /* case 1 trumps case 3 */
 760     }
 761   }
 762 }
 763 #endif /* !SQLITE_OMIT_OR_OPTIMIZATION && !SQLITE_OMIT_SUBQUERY */
 764
 765 /*
 766 ** We already know that pExpr is a binary operator where both operands are
 767 ** column references.  This routine checks to see if pExpr is an equivalence
 768 ** relation:
 769 **   1.  The SQLITE_Transitive optimization must be enabled
 770 **   2.  Must be either an == or an IS operator
 771 **   3.  Not originating in the ON clause of an OUTER JOIN
 772 **   4.  The affinities of A and B must be compatible
 773 **   5a. Both operands use the same collating sequence OR
 774 **   5b. The overall collating sequence is BINARY
 775 ** If this routine returns TRUE, that means that the RHS can be substituted
 776 ** for the LHS anyplace else in the WHERE clause where the LHS column occurs.
 777 ** This is an optimization.  No harm comes from returning 0.  But if 1 is
 778 ** returned when it should not be, then incorrect answers might result.
 779 */
 780 static int termIsEquivalence(Parse *pParse, Expr *pExpr){
 781   char aff1, aff2;
 782   CollSeq *pColl;
 783   const char *zColl1, *zColl2;
 784   if( !OptimizationEnabled(pParse->db, SQLITE_Transitive) ) return 0;
 785   if( pExpr->op!=TK_EQ && pExpr->op!=TK_IS ) return 0;
 786   if( ExprHasProperty(pExpr, EP_FromJoin) ) return 0;
 787   aff1 = sqlite3ExprAffinity(pExpr->pLeft);
 788   aff2 = sqlite3ExprAffinity(pExpr->pRight);
 789   if( aff1!=aff2
 790    && (!sqlite3IsNumericAffinity(aff1) || !sqlite3IsNumericAffinity(aff2))
 791   ){
 792     return 0;
 793   }
 794   pColl = sqlite3BinaryCompareCollSeq(pParse, pExpr->pLeft, pExpr->pRight);
 795   if( pColl==0 || sqlite3StrICmp(pColl->zName, "BINARY")==0 ) return 1;
 796   pColl = sqlite3ExprCollSeq(pParse, pExpr->pLeft);
 797   zColl1 = pColl ? pColl->zName : 0;
 798   pColl = sqlite3ExprCollSeq(pParse, pExpr->pRight);
 799   zColl2 = pColl ? pColl->zName : 0;
 800   return sqlite3_stricmp(zColl1, zColl2)==0;
 801 }
 802
 803 /*
 804 ** Recursively walk the expressions of a SELECT statement and generate
 805 ** a bitmask indicating which tables are used in that expression
 806 ** tree.
 807 */
 808 static Bitmask exprSelectUsage(WhereMaskSet *pMaskSet, Select *pS){
 809   Bitmask mask = 0;
 810   while( pS ){
 811     SrcList *pSrc = pS->pSrc;
 812     mask |= sqlite3WhereExprListUsage(pMaskSet, pS->pEList);
 813     mask |= sqlite3WhereExprListUsage(pMaskSet, pS->pGroupBy);
 814     mask |= sqlite3WhereExprListUsage(pMaskSet, pS->pOrderBy);
 815     mask |= sqlite3WhereExprUsage(pMaskSet, pS->pWhere);
 816     mask |= sqlite3WhereExprUsage(pMaskSet, pS->pHaving);
 817     if( ALWAYS(pSrc!=0) ){
 818       int i;
 819       for(i=0; i<pSrc->nSrc; i++){
 820         mask |= exprSelectUsage(pMaskSet, pSrc->a[i].pSelect);
 821         mask |= sqlite3WhereExprUsage(pMaskSet, pSrc->a[i].pOn);
 822       }
 823     }
 824     pS = pS->pPrior;
 825   }
 826   return mask;
 827 }
 828
 829 /*
 830 ** Expression pExpr is one operand of a comparison operator that might
 831 ** be useful for indexing.  This routine checks to see if pExpr appears
 832 ** in any index.  Return TRUE (1) if pExpr is an indexed term and return
 833 ** FALSE (0) if not.  If TRUE is returned, also set aiCurCol[0] to the cursor
 834 ** number of the table that is indexed and aiCurCol[1] to the column number
 835 ** of the column that is indexed, or XN_EXPR (-2) if an expression is being
 836 ** indexed.
 837 **
 838 ** If pExpr is a TK_COLUMN column reference, then this routine always returns
 839 ** true even if that particular column is not indexed, because the column
 840 ** might be added to an automatic index later.
 841 */
 842 static SQLITE_NOINLINE int exprMightBeIndexed2(
 843   SrcList *pFrom,        /* The FROM clause */
 844   Bitmask mPrereq,       /* Bitmask of FROM clause terms referenced by pExpr */
 845   int *aiCurCol,         /* Write the referenced table cursor and column here */
 846   Expr *pExpr            /* An operand of a comparison operator */
 847 ){
 848   Index *pIdx;
 849   int i;
 850   int iCur;
 851   for(i=0; mPrereq>1; i++, mPrereq>>=1){}
 852   iCur = pFrom->a[i].iCursor;
 853   for(pIdx=pFrom->a[i].pTab->pIndex; pIdx; pIdx=pIdx->pNext){
 854     if( pIdx->aColExpr==0 ) continue;
 855     for(i=0; i<pIdx->nKeyCol; i++){
 856       if( pIdx->aiColumn[i]!=XN_EXPR ) continue;
 857       if( sqlite3ExprCompareSkip(pExpr, pIdx->aColExpr->a[i].pExpr, iCur)==0 ){
 858         aiCurCol[0] = iCur;
 859         aiCurCol[1] = XN_EXPR;
 860         return 1;
 861       }
 862     }
 863   }
 864   return 0;
 865 }
 866 static int exprMightBeIndexed(
 867   SrcList *pFrom,        /* The FROM clause */
 868   Bitmask mPrereq,       /* Bitmask of FROM clause terms referenced by pExpr */
 869   int *aiCurCol,         /* Write the referenced table cursor & column here */
 870   Expr *pExpr,           /* An operand of a comparison operator */
 871   int op                 /* The specific comparison operator */
 872 ){
 873   /* If this expression is a vector to the left or right of a
 874   ** inequality constraint (>, <, >= or <=), perform the processing
 875   ** on the first element of the vector.  */
 876   assert( TK_GT+1==TK_LE && TK_GT+2==TK_LT && TK_GT+3==TK_GE );
 877   assert( TK_IS<TK_GE && TK_ISNULL<TK_GE && TK_IN<TK_GE );
 878   assert( op<=TK_GE );
 879   if( pExpr->op==TK_VECTOR && (op>=TK_GT && ALWAYS(op<=TK_GE)) ){
 880     pExpr = pExpr->x.pList->a[0].pExpr;
 881   }
 882
 883   if( pExpr->op==TK_COLUMN ){
 884     aiCurCol[0] = pExpr->iTable;
 885     aiCurCol[1] = pExpr->iColumn;
 886     return 1;
 887   }
 888   if( mPrereq==0 ) return 0;                 /* No table references */
 889   if( (mPrereq&(mPrereq-1))!=0 ) return 0;   /* Refs more than one table */
 890   return exprMightBeIndexed2(pFrom,mPrereq,aiCurCol,pExpr);
 891 }
 892
 893 /*
 894 ** The input to this routine is an WhereTerm structure with only the
 895 ** "pExpr" field filled in.  The job of this routine is to analyze the
 896 ** subexpression and populate all the other fields of the WhereTerm
 897 ** structure.
 898 **
 899 ** If the expression is of the form "<expr> <op> X" it gets commuted
 900 ** to the standard form of "X <op> <expr>".
 901 **
 902 ** If the expression is of the form "X <op> Y" where both X and Y are
 903 ** columns, then the original expression is unchanged and a new virtual
 904 ** term of the form "Y <op> X" is added to the WHERE clause and
 905 ** analyzed separately.  The original term is marked with TERM_COPIED
 906 ** and the new term is marked with TERM_DYNAMIC (because it's pExpr
 907 ** needs to be freed with the WhereClause) and TERM_VIRTUAL (because it
 908 ** is a commuted copy of a prior term.)  The original term has nChild=1
 909 ** and the copy has idxParent set to the index of the original term.
 910 */
 911 static void exprAnalyze(
 912   SrcList *pSrc,            /* the FROM clause */
 913   WhereClause *pWC,         /* the WHERE clause */
 914   int idxTerm               /* Index of the term to be analyzed */
 915 ){
 916   WhereInfo *pWInfo = pWC->pWInfo; /* WHERE clause processing context */
 917   WhereTerm *pTerm;                /* The term to be analyzed */
 918   WhereMaskSet *pMaskSet;          /* Set of table index masks */
 919   Expr *pExpr;                     /* The expression to be analyzed */
 920   Bitmask prereqLeft;              /* Prerequesites of the pExpr->pLeft */
 921   Bitmask prereqAll;               /* Prerequesites of pExpr */
 922   Bitmask extraRight = 0;          /* Extra dependencies on LEFT JOIN */
 923   Expr *pStr1 = 0;                 /* RHS of LIKE/GLOB operator */
 924   int isComplete = 0;              /* RHS of LIKE/GLOB ends with wildcard */
 925   int noCase = 0;                  /* uppercase equivalent to lowercase */
 926   int op;                          /* Top-level operator.  pExpr->op */
 927   Parse *pParse = pWInfo->pParse;  /* Parsing context */
 928   sqlite3 *db = pParse->db;        /* Database connection */
 929   unsigned char eOp2;              /* op2 value for LIKE/REGEXP/GLOB */
 930   int nLeft;                       /* Number of elements on left side vector */
 931
 932   if( db->mallocFailed ){
 933     return;
 934   }
 935   pTerm = &pWC->a[idxTerm];
 936   pMaskSet = &pWInfo->sMaskSet;
 937   pExpr = pTerm->pExpr;
 938   assert( pExpr->op!=TK_AS && pExpr->op!=TK_COLLATE );
 939   prereqLeft = sqlite3WhereExprUsage(pMaskSet, pExpr->pLeft);
 940   op = pExpr->op;
 941   if( op==TK_IN ){
 942     assert( pExpr->pRight==0 );
 943     if( sqlite3ExprCheckIN(pParse, pExpr) ) return;
 944     if( ExprHasProperty(pExpr, EP_xIsSelect) ){
 945       pTerm->prereqRight = exprSelectUsage(pMaskSet, pExpr->x.pSelect);
 946     }else{
 947       pTerm->prereqRight = sqlite3WhereExprListUsage(pMaskSet, pExpr->x.pList);
 948     }
 949   }else if( op==TK_ISNULL ){
 950     pTerm->prereqRight = 0;
 951   }else{
 952     pTerm->prereqRight = sqlite3WhereExprUsage(pMaskSet, pExpr->pRight);
 953   }
 954   pMaskSet->bVarSelect = 0;
 955   prereqAll = sqlite3WhereExprUsage(pMaskSet, pExpr);
 956   if( pMaskSet->bVarSelect ) pTerm->wtFlags |= TERM_VARSELECT;
 957   if( ExprHasProperty(pExpr, EP_FromJoin) ){
 958     Bitmask x = sqlite3WhereGetMask(pMaskSet, pExpr->iRightJoinTable);
 959     prereqAll |= x;
 960     extraRight = x-1;  /* ON clause terms may not be used with an index
 961                        ** on left table of a LEFT JOIN.  Ticket #3015 */
 962     if( (prereqAll>>1)>=x ){
 963       sqlite3ErrorMsg(pParse, "ON clause references tables to its right");
 964       return;
 965     }
 966   }
 967   pTerm->prereqAll = prereqAll;
 968   pTerm->leftCursor = -1;
 969   pTerm->iParent = -1;
 970   pTerm->eOperator = 0;
 971   if( allowedOp(op) ){
 972     int aiCurCol[2];
 973     Expr *pLeft = sqlite3ExprSkipCollate(pExpr->pLeft);
 974     Expr *pRight = sqlite3ExprSkipCollate(pExpr->pRight);
 975     u16 opMask = (pTerm->prereqRight & prereqLeft)==0 ? WO_ALL : WO_EQUIV;
 976
 977     if( pTerm->iField>0 ){
 978       assert( op==TK_IN );
 979       assert( pLeft->op==TK_VECTOR );
 980       pLeft = pLeft->x.pList->a[pTerm->iField-1].pExpr;
 981     }
 982
 983     if( exprMightBeIndexed(pSrc, prereqLeft, aiCurCol, pLeft, op) ){
 984       pTerm->leftCursor = aiCurCol[0];
 985       pTerm->u.leftColumn = aiCurCol[1];
 986       pTerm->eOperator = operatorMask(op) & opMask;
 987     }
 988     if( op==TK_IS ) pTerm->wtFlags |= TERM_IS;
 989     if( pRight
 990      && exprMightBeIndexed(pSrc, pTerm->prereqRight, aiCurCol, pRight, op)
 991     ){
 992       WhereTerm *pNew;
 993       Expr *pDup;
 994       u16 eExtraOp = 0;        /* Extra bits for pNew->eOperator */
 995       assert( pTerm->iField==0 );
 996       if( pTerm->leftCursor>=0 ){
 997         int idxNew;
 998         pDup = sqlite3ExprDup(db, pExpr, 0);
 999         if( db->mallocFailed ){
1000           sqlite3ExprDelete(db, pDup);
1001           return;
1002         }
1003         idxNew = whereClauseInsert(pWC, pDup, TERM_VIRTUAL|TERM_DYNAMIC);
1004         if( idxNew==0 ) return;
1005         pNew = &pWC->a[idxNew];
1006         markTermAsChild(pWC, idxNew, idxTerm);
1007         if( op==TK_IS ) pNew->wtFlags |= TERM_IS;
1008         pTerm = &pWC->a[idxTerm];
1009         pTerm->wtFlags |= TERM_COPIED;
1010
1011         if( termIsEquivalence(pParse, pDup) ){
1012           pTerm->eOperator |= WO_EQUIV;
1013           eExtraOp = WO_EQUIV;
1014         }
1015       }else{
1016         pDup = pExpr;
1017         pNew = pTerm;
1018       }
1019       exprCommute(pParse, pDup);
1020       pNew->leftCursor = aiCurCol[0];
1021       pNew->u.leftColumn = aiCurCol[1];
1022       testcase( (prereqLeft | extraRight) != prereqLeft );
1023       pNew->prereqRight = prereqLeft | extraRight;
1024       pNew->prereqAll = prereqAll;
1025       pNew->eOperator = (operatorMask(pDup->op) + eExtraOp) & opMask;
1026     }
1027   }
1028
1029 #ifndef SQLITE_OMIT_BETWEEN_OPTIMIZATION
1030   /* If a term is the BETWEEN operator, create two new virtual terms
1031   ** that define the range that the BETWEEN implements.  For example:
1032   **
1033   **      a BETWEEN b AND c
1034   **
1035   ** is converted into:
1036   **
1037   **      (a BETWEEN b AND c) AND (a>=b) AND (a<=c)
1038   **
1039   ** The two new terms are added onto the end of the WhereClause object.
1040   ** The new terms are "dynamic" and are children of the original BETWEEN
1041   ** term.  That means that if the BETWEEN term is coded, the children are
1042   ** skipped.  Or, if the children are satisfied by an index, the original
1043   ** BETWEEN term is skipped.
1044   */
1045   else if( pExpr->op==TK_BETWEEN && pWC->op==TK_AND ){
1046     ExprList *pList = pExpr->x.pList;
1047     int i;
1048     static const u8 ops[] = {TK_GE, TK_LE};
1049     assert( pList!=0 );
1050     assert( pList->nExpr==2 );
1051     for(i=0; i<2; i++){
1052       Expr *pNewExpr;
1053       int idxNew;
1054       pNewExpr = sqlite3PExpr(pParse, ops[i],
1055                              sqlite3ExprDup(db, pExpr->pLeft, 0),
1056                              sqlite3ExprDup(db, pList->a[i].pExpr, 0));
1057       transferJoinMarkings(pNewExpr, pExpr);
1058       idxNew = whereClauseInsert(pWC, pNewExpr, TERM_VIRTUAL|TERM_DYNAMIC);
1059       testcase( idxNew==0 );
1060       exprAnalyze(pSrc, pWC, idxNew);
1061       pTerm = &pWC->a[idxTerm];
1062       markTermAsChild(pWC, idxNew, idxTerm);
1063     }
1064   }
1065 #endif /* SQLITE_OMIT_BETWEEN_OPTIMIZATION */
1066
1067 #if !defined(SQLITE_OMIT_OR_OPTIMIZATION) && !defined(SQLITE_OMIT_SUBQUERY)
1068   /* Analyze a term that is composed of two or more subterms connected by
1069   ** an OR operator.
1070   */
1071   else if( pExpr->op==TK_OR ){
1072     assert( pWC->op==TK_AND );
1073     exprAnalyzeOrTerm(pSrc, pWC, idxTerm);
1074     pTerm = &pWC->a[idxTerm];
1075   }
1076 #endif /* SQLITE_OMIT_OR_OPTIMIZATION */
1077
1078 #ifndef SQLITE_OMIT_LIKE_OPTIMIZATION
1079   /* Add constraints to reduce the search space on a LIKE or GLOB
1080   ** operator.
1081   **
1082   ** A like pattern of the form "x LIKE 'aBc%'" is changed into constraints
1083   **
1084   **          x>='ABC' AND x<'abd' AND x LIKE 'aBc%'
1085   **
1086   ** The last character of the prefix "abc" is incremented to form the
1087   ** termination condition "abd".  If case is not significant (the default
1088   ** for LIKE) then the lower-bound is made all uppercase and the upper-
1089   ** bound is made all lowercase so that the bounds also work when comparing
1090   ** BLOBs.
1091   */
1092   if( pWC->op==TK_AND
1093    && isLikeOrGlob(pParse, pExpr, &pStr1, &isComplete, &noCase)
1094   ){
1095     Expr *pLeft;       /* LHS of LIKE/GLOB operator */
1096     Expr *pStr2;       /* Copy of pStr1 - RHS of LIKE/GLOB operator */
1097     Expr *pNewExpr1;
1098     Expr *pNewExpr2;
1099     int idxNew1;
1100     int idxNew2;
1101     const char *zCollSeqName;     /* Name of collating sequence */
1102     const u16 wtFlags = TERM_LIKEOPT | TERM_VIRTUAL | TERM_DYNAMIC;
1103
1104     pLeft = pExpr->x.pList->a[1].pExpr;
1105     pStr2 = sqlite3ExprDup(db, pStr1, 0);
1106
1107     /* Convert the lower bound to upper-case and the upper bound to
1108     ** lower-case (upper-case is less than lower-case in ASCII) so that
1109     ** the range constraints also work for BLOBs
1110     */
1111     if( noCase && !pParse->db->mallocFailed ){
1112       int i;
1113       char c;
1114       pTerm->wtFlags |= TERM_LIKE;
1115       for(i=0; (c = pStr1->u.zToken[i])!=0; i++){
1116         pStr1->u.zToken[i] = sqlite3Toupper(c);
1117         pStr2->u.zToken[i] = sqlite3Tolower(c);
1118       }
1119     }
1120
1121     if( !db->mallocFailed ){
1122       u8 c, *pC;       /* Last character before the first wildcard */
1123       pC = (u8*)&pStr2->u.zToken[sqlite3Strlen30(pStr2->u.zToken)-1];
1124       c = *pC;
1125       if( noCase ){
1126         /* The point is to increment the last character before the first
1127         ** wildcard.  But if we increment '@', that will push it into the
1128         ** alphabetic range where case conversions will mess up the
1129         ** inequality.  To avoid this, make sure to also run the full
1130         ** LIKE on all candidate expressions by clearing the isComplete flag
1131         */
1132         if( c=='A'-1 ) isComplete = 0;
1133         c = sqlite3UpperToLower[c];
1134       }
1135       *pC = c + 1;
1136     }
1137     zCollSeqName = noCase ? "NOCASE" : "BINARY";
1138     pNewExpr1 = sqlite3ExprDup(db, pLeft, 0);
1139     pNewExpr1 = sqlite3PExpr(pParse, TK_GE,
1140            sqlite3ExprAddCollateString(pParse,pNewExpr1,zCollSeqName),
1141            pStr1);
1142     transferJoinMarkings(pNewExpr1, pExpr);
1143     idxNew1 = whereClauseInsert(pWC, pNewExpr1, wtFlags);
1144     testcase( idxNew1==0 );
1145     exprAnalyze(pSrc, pWC, idxNew1);
1146     pNewExpr2 = sqlite3ExprDup(db, pLeft, 0);
1147     pNewExpr2 = sqlite3PExpr(pParse, TK_LT,
1148            sqlite3ExprAddCollateString(pParse,pNewExpr2,zCollSeqName),
1149            pStr2);
1150     transferJoinMarkings(pNewExpr2, pExpr);
1151     idxNew2 = whereClauseInsert(pWC, pNewExpr2, wtFlags);
1152     testcase( idxNew2==0 );
1153     exprAnalyze(pSrc, pWC, idxNew2);
1154     pTerm = &pWC->a[idxTerm];
1155     if( isComplete ){
1156       markTermAsChild(pWC, idxNew1, idxTerm);
1157       markTermAsChild(pWC, idxNew2, idxTerm);
1158     }
1159   }
1160 #endif /* SQLITE_OMIT_LIKE_OPTIMIZATION */
1161
1162 #ifndef SQLITE_OMIT_VIRTUALTABLE
1163   /* Add a WO_MATCH auxiliary term to the constraint set if the
1164   ** current expression is of the form:  column MATCH expr.
1165   ** This information is used by the xBestIndex methods of
1166   ** virtual tables.  The native query optimizer does not attempt
1167   ** to do anything with MATCH functions.
1168   */
1169   if( pWC->op==TK_AND && isMatchOfColumn(pExpr, &eOp2) ){
1170     int idxNew;
1171     Expr *pRight, *pLeft;
1172     WhereTerm *pNewTerm;
1173     Bitmask prereqColumn, prereqExpr;
1174
1175     pRight = pExpr->x.pList->a[0].pExpr;
1176     pLeft = pExpr->x.pList->a[1].pExpr;
1177     prereqExpr = sqlite3WhereExprUsage(pMaskSet, pRight);
1178     prereqColumn = sqlite3WhereExprUsage(pMaskSet, pLeft);
1179     if( (prereqExpr & prereqColumn)==0 ){
1180       Expr *pNewExpr;
1181       pNewExpr = sqlite3PExpr(pParse, TK_MATCH,
1182                               0, sqlite3ExprDup(db, pRight, 0));
1183       if( ExprHasProperty(pExpr, EP_FromJoin) && pNewExpr ){
1184         ExprSetProperty(pNewExpr, EP_FromJoin);
1185       }
1186       idxNew = whereClauseInsert(pWC, pNewExpr, TERM_VIRTUAL|TERM_DYNAMIC);
1187       testcase( idxNew==0 );
1188       pNewTerm = &pWC->a[idxNew];
1189       pNewTerm->prereqRight = prereqExpr;
1190       pNewTerm->leftCursor = pLeft->iTable;
1191       pNewTerm->u.leftColumn = pLeft->iColumn;
1192       pNewTerm->eOperator = WO_MATCH;
1193       pNewTerm->eMatchOp = eOp2;
1194       markTermAsChild(pWC, idxNew, idxTerm);
1195       pTerm = &pWC->a[idxTerm];
1196       pTerm->wtFlags |= TERM_COPIED;
1197       pNewTerm->prereqAll = pTerm->prereqAll;
1198     }
1199   }
1200 #endif /* SQLITE_OMIT_VIRTUALTABLE */
1201
1202   /* If there is a vector == or IS term - e.g. "(a, b) == (?, ?)" - create
1203   ** new terms for each component comparison - "a = ?" and "b = ?".  The
1204   ** new terms completely replace the original vector comparison, which is
1205   ** no longer used.
1206   **
1207   ** This is only required if at least one side of the comparison operation
1208   ** is not a sub-select.  */
1209   if( pWC->op==TK_AND
1210   && (pExpr->op==TK_EQ || pExpr->op==TK_IS)
1211   && (nLeft = sqlite3ExprVectorSize(pExpr->pLeft))>1
1212   && sqlite3ExprVectorSize(pExpr->pRight)==nLeft
1213   && ( (pExpr->pLeft->flags & EP_xIsSelect)==0
1214     || (pExpr->pRight->flags & EP_xIsSelect)==0)
1215   ){
1216     int i;
1217     for(i=0; i<nLeft; i++){
1218       int idxNew;
1219       Expr *pNew;
1220       Expr *pLeft = sqlite3ExprForVectorField(pParse, pExpr->pLeft, i);
1221       Expr *pRight = sqlite3ExprForVectorField(pParse, pExpr->pRight, i);
1222
1223       pNew = sqlite3PExpr(pParse, pExpr->op, pLeft, pRight);
1224       transferJoinMarkings(pNew, pExpr);
1225       idxNew = whereClauseInsert(pWC, pNew, TERM_DYNAMIC);
1226       exprAnalyze(pSrc, pWC, idxNew);
1227     }
1228     pTerm = &pWC->a[idxTerm];
1229     pTerm->wtFlags = TERM_CODED|TERM_VIRTUAL;  /* Disable the original */
1230     pTerm->eOperator = 0;
1231   }
1232
1233   /* If there is a vector IN term - e.g. "(a, b) IN (SELECT ...)" - create
1234   ** a virtual term for each vector component. The expression object
1235   ** used by each such virtual term is pExpr (the full vector IN(...)
1236   ** expression). The WhereTerm.iField variable identifies the index within
1237   ** the vector on the LHS that the virtual term represents.
1238   **
1239   ** This only works if the RHS is a simple SELECT, not a compound
1240   */
1241   if( pWC->op==TK_AND && pExpr->op==TK_IN && pTerm->iField==0
1242    && pExpr->pLeft->op==TK_VECTOR
1243    && pExpr->x.pSelect->pPrior==0
1244   ){
1245     int i;
1246     for(i=0; i<sqlite3ExprVectorSize(pExpr->pLeft); i++){
1247       int idxNew;
1248       idxNew = whereClauseInsert(pWC, pExpr, TERM_VIRTUAL);
1249       pWC->a[idxNew].iField = i+1;
1250       exprAnalyze(pSrc, pWC, idxNew);
1251       markTermAsChild(pWC, idxNew, idxTerm);
1252     }
1253   }
1254
1255 #ifdef SQLITE_ENABLE_STAT3_OR_STAT4
1256   /* When sqlite_stat3 histogram data is available an operator of the
1257   ** form "x IS NOT NULL" can sometimes be evaluated more efficiently
1258   ** as "x>NULL" if x is not an INTEGER PRIMARY KEY.  So construct a
1259   ** virtual term of that form.
1260   **
1261   ** Note that the virtual term must be tagged with TERM_VNULL.
1262   */
1263   if( pExpr->op==TK_NOTNULL
1264    && pExpr->pLeft->op==TK_COLUMN
1265    && pExpr->pLeft->iColumn>=0
1266    && OptimizationEnabled(db, SQLITE_Stat34)
1267   ){
1268     Expr *pNewExpr;
1269     Expr *pLeft = pExpr->pLeft;
1270     int idxNew;
1271     WhereTerm *pNewTerm;
1272
1273     pNewExpr = sqlite3PExpr(pParse, TK_GT,
1274                             sqlite3ExprDup(db, pLeft, 0),
1275                             sqlite3ExprAlloc(db, TK_NULL, 0, 0));
1276
1277     idxNew = whereClauseInsert(pWC, pNewExpr,
1278                               TERM_VIRTUAL|TERM_DYNAMIC|TERM_VNULL);
1279     if( idxNew ){
1280       pNewTerm = &pWC->a[idxNew];
1281       pNewTerm->prereqRight = 0;
1282       pNewTerm->leftCursor = pLeft->iTable;
1283       pNewTerm->u.leftColumn = pLeft->iColumn;
1284       pNewTerm->eOperator = WO_GT;
1285       markTermAsChild(pWC, idxNew, idxTerm);
1286       pTerm = &pWC->a[idxTerm];
1287       pTerm->wtFlags |= TERM_COPIED;
1288       pNewTerm->prereqAll = pTerm->prereqAll;
1289     }
1290   }
1291 #endif /* SQLITE_ENABLE_STAT3_OR_STAT4 */
1292
1293   /* Prevent ON clause terms of a LEFT JOIN from being used to drive
1294   ** an index for tables to the left of the join.
1295   */
1296   testcase( pTerm!=&pWC->a[idxTerm] );
1297   pTerm = &pWC->a[idxTerm];
1298   pTerm->prereqRight |= extraRight;
1299 }
1300
1301 /***************************************************************************
1302 ** Routines with file scope above.  Interface to the rest of the where.c
1303 ** subsystem follows.
1304 ***************************************************************************/
1305
1306 /*
1307 ** This routine identifies subexpressions in the WHERE clause where
1308 ** each subexpression is separated by the AND operator or some other
1309 ** operator specified in the op parameter.  The WhereClause structure
1310 ** is filled with pointers to subexpressions.  For example:
1311 **
1312 **    WHERE  a=='hello' AND coalesce(b,11)<10 AND (c+12!=d OR c==22)
1313 **           \________/     \_______________/     \________________/
1314 **            slot[0]            slot[1]               slot[2]
1315 **
1316 ** The original WHERE clause in pExpr is unaltered.  All this routine
1317 ** does is make slot[] entries point to substructure within pExpr.
1318 **
1319 ** In the previous sentence and in the diagram, "slot[]" refers to
1320 ** the WhereClause.a[] array.  The slot[] array grows as needed to contain
1321 ** all terms of the WHERE clause.
1322 */
1323 void sqlite3WhereSplit(WhereClause *pWC, Expr *pExpr, u8 op){
1324   Expr *pE2 = sqlite3ExprSkipCollate(pExpr);
1325   pWC->op = op;
1326   if( pE2==0 ) return;
1327   if( pE2->op!=op ){
1328     whereClauseInsert(pWC, pExpr, 0);
1329   }else{
1330     sqlite3WhereSplit(pWC, pE2->pLeft, op);
1331     sqlite3WhereSplit(pWC, pE2->pRight, op);
1332   }
1333 }
1334
1335 /*
1336 ** Initialize a preallocated WhereClause structure.
1337 */
1338 void sqlite3WhereClauseInit(
1339   WhereClause *pWC,        /* The WhereClause to be initialized */
1340   WhereInfo *pWInfo        /* The WHERE processing context */
1341 ){
1342   pWC->pWInfo = pWInfo;
1343   pWC->pOuter = 0;
1344   pWC->nTerm = 0;
1345   pWC->nSlot = ArraySize(pWC->aStatic);
1346   pWC->a = pWC->aStatic;
1347 }
1348
1349 /*
1350 ** Deallocate a WhereClause structure.  The WhereClause structure
1351 ** itself is not freed.  This routine is the inverse of
1352 ** sqlite3WhereClauseInit().
1353 */
1354 void sqlite3WhereClauseClear(WhereClause *pWC){
1355   int i;
1356   WhereTerm *a;
1357   sqlite3 *db = pWC->pWInfo->pParse->db;
1358   for(i=pWC->nTerm-1, a=pWC->a; i>=0; i--, a++){
1359     if( a->wtFlags & TERM_DYNAMIC ){
1360       sqlite3ExprDelete(db, a->pExpr);
1361     }
1362     if( a->wtFlags & TERM_ORINFO ){
1363       whereOrInfoDelete(db, a->u.pOrInfo);
1364     }else if( a->wtFlags & TERM_ANDINFO ){
1365       whereAndInfoDelete(db, a->u.pAndInfo);
1366     }
1367   }
1368   if( pWC->a!=pWC->aStatic ){
1369     sqlite3DbFree(db, pWC->a);
1370   }
1371 }
1372
1373
1374 /*
1375 ** These routines walk (recursively) an expression tree and generate
1376 ** a bitmask indicating which tables are used in that expression
1377 ** tree.
1378 */
1379 Bitmask sqlite3WhereExprUsage(WhereMaskSet *pMaskSet, Expr *p){
1380   Bitmask mask;
1381   if( p==0 ) return 0;
1382   if( p->op==TK_COLUMN ){
1383     return sqlite3WhereGetMask(pMaskSet, p->iTable);
1384   }
1385   mask = (p->op==TK_IF_NULL_ROW) ? sqlite3WhereGetMask(pMaskSet, p->iTable) : 0;
1386   assert( !ExprHasProperty(p, EP_TokenOnly) );
1387   if( p->pLeft ) mask |= sqlite3WhereExprUsage(pMaskSet, p->pLeft);
1388   if( p->pRight ){
1389     mask |= sqlite3WhereExprUsage(pMaskSet, p->pRight);
1390     assert( p->x.pList==0 );
1391   }else if( ExprHasProperty(p, EP_xIsSelect) ){
1392     if( ExprHasProperty(p, EP_VarSelect) ) pMaskSet->bVarSelect = 1;
1393     mask |= exprSelectUsage(pMaskSet, p->x.pSelect);
1394   }else if( p->x.pList ){
1395     mask |= sqlite3WhereExprListUsage(pMaskSet, p->x.pList);
1396   }
1397   return mask;
1398 }
1399 Bitmask sqlite3WhereExprListUsage(WhereMaskSet *pMaskSet, ExprList *pList){
1400   int i;
1401   Bitmask mask = 0;
1402   if( pList ){
1403     for(i=0; i<pList->nExpr; i++){
1404       mask |= sqlite3WhereExprUsage(pMaskSet, pList->a[i].pExpr);
1405     }
1406   }
1407   return mask;
1408 }
1409
1410
1411 /*
1412 ** Call exprAnalyze on all terms in a WHERE clause.
1413 **
1414 ** Note that exprAnalyze() might add new virtual terms onto the
1415 ** end of the WHERE clause.  We do not want to analyze these new
1416 ** virtual terms, so start analyzing at the end and work forward
1417 ** so that the added virtual terms are never processed.
1418 */
1419 void sqlite3WhereExprAnalyze(
1420   SrcList *pTabList,       /* the FROM clause */
1421   WhereClause *pWC         /* the WHERE clause to be analyzed */
1422 ){
1423   int i;
1424   for(i=pWC->nTerm-1; i>=0; i--){
1425     exprAnalyze(pTabList, pWC, i);
1426   }
1427 }
1428
1429 /*
1430 ** For table-valued-functions, transform the function arguments into
1431 ** new WHERE clause terms.
1432 **
1433 ** Each function argument translates into an equality constraint against
1434 ** a HIDDEN column in the table.
1435 */
1436 void sqlite3WhereTabFuncArgs(
1437   Parse *pParse,                    /* Parsing context */
1438   struct SrcList_item *pItem,       /* The FROM clause term to process */
1439   WhereClause *pWC                  /* Xfer function arguments to here */
1440 ){
1441   Table *pTab;
1442   int j, k;
1443   ExprList *pArgs;
1444   Expr *pColRef;
1445   Expr *pTerm;
1446   if( pItem->fg.isTabFunc==0 ) return;
1447   pTab = pItem->pTab;
1448   assert( pTab!=0 );
1449   pArgs = pItem->u1.pFuncArg;
1450   if( pArgs==0 ) return;
1451   for(j=k=0; j<pArgs->nExpr; j++){
1452     while( k<pTab->nCol && (pTab->aCol[k].colFlags & COLFLAG_HIDDEN)==0 ){k++;}
1453     if( k>=pTab->nCol ){
1454       sqlite3ErrorMsg(pParse, "too many arguments on %s() - max %d",
1455                       pTab->zName, j);
1456       return;
1457     }
1458     pColRef = sqlite3ExprAlloc(pParse->db, TK_COLUMN, 0, 0);
1459     if( pColRef==0 ) return;
1460     pColRef->iTable = pItem->iCursor;
1461     pColRef->iColumn = k++;
1462     pColRef->pTab = pTab;
1463     pTerm = sqlite3PExpr(pParse, TK_EQ, pColRef,
1464                          sqlite3ExprDup(pParse->db, pArgs->a[j].pExpr, 0));
1465     whereClauseInsert(pWC, pTerm, TERM_DYNAMIC);
1466   }
1467 }