SQLite

Check-in [e68b427afb]
Login

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

Overview
Comment:Convert expressions of the form "X IN (?)" with exactly one value on the RHS of the IN into equality tests: "X=?". Add test cases to verify that statements work correctly on this corner case. Fix for ticket [e39d032577df6942].
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: e68b427afbc82e201c64474117851aa4c9eb0c92
User & Date: drh 2014-03-20 13:26:47.152
References
2019-08-27
17:01
Omit the "x IN (y)" to "x==y" optimization of check-in [e68b427afbc82e20] (and ticket [e39d032577df6942]) as it causes difficult affinity problems as demonstrated by ticket [dbaf8a6820be1ece] and the original assertion fault is no longer a factor due to countless other changes of the previous 5 years. (check-in: 7f5168a76a user: drh tags: trunk)
2014-03-20
17:03
The "x IN (?)" optimization in check-ins [2ff3b25f40] and [e68b427afb] is incorrect, as demonstrated by the in4-5.1 test case in this check-in. The "COLLATE binary" that was being added to the RHS of IN was overriding the implicit collating sequence of the LHS. This change defines the EP_Generic expression node property that blocks all affinity or collating sequence information in the expression subtree and adds that property to the expression taken from RHS of the IN operator. (check-in: 2ea4a9f75f user: drh tags: trunk)
Context
2014-03-20
14:56
Previous check-in is not quite correct. "x IN (?)" is not exactly the same as "x==?" do to collation and affinity issues. The correct converstion should be to "x==(+? COLLATE binary)". The current check-in fixes this problem and provides test cases. Ticket [e39d032577df69] (check-in: 2ff3b25f40 user: drh tags: trunk)
13:26
Convert expressions of the form "X IN (?)" with exactly one value on the RHS of the IN into equality tests: "X=?". Add test cases to verify that statements work correctly on this corner case. Fix for ticket [e39d032577df6942]. (check-in: e68b427afb user: drh tags: trunk)
12:17
Fix an unnecessarily obtuse use of a bitmask flag. (check-in: ca31408131 user: drh tags: trunk)
Changes
Unified Diff Ignore Whitespace Patch
Changes to src/parse.y.
1016
1017
1018
1019
1020
1021
1022















1023
1024
1025
1026
1027
1028
1029
      **      expr1 NOT IN ()
      **
      ** simplify to constants 0 (false) and 1 (true), respectively,
      ** regardless of the value of expr1.
      */
      A.pExpr = sqlite3PExpr(pParse, TK_INTEGER, 0, 0, &sqlite3IntTokens[N]);
      sqlite3ExprDelete(pParse->db, X.pExpr);















    }else{
      A.pExpr = sqlite3PExpr(pParse, TK_IN, X.pExpr, 0, 0);
      if( A.pExpr ){
        A.pExpr->x.pList = Y;
        sqlite3ExprSetHeight(pParse, A.pExpr);
      }else{
        sqlite3ExprListDelete(pParse->db, Y);







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







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
      **      expr1 NOT IN ()
      **
      ** simplify to constants 0 (false) and 1 (true), respectively,
      ** regardless of the value of expr1.
      */
      A.pExpr = sqlite3PExpr(pParse, TK_INTEGER, 0, 0, &sqlite3IntTokens[N]);
      sqlite3ExprDelete(pParse->db, X.pExpr);
    }else if( Y->nExpr==1 ){
      /* Expressions of the form:
      **
      **      expr1 IN (?1)
      **      expr1 NOT IN (?2)
      **
      ** with exactly one value on the RHS can be simplified to:
      **
      **      expr1 == ?1
      **      expr1 <> ?2
      */
      Expr *pRHS = Y->a[0].pExpr;
      Y->a[0].pExpr = 0;
      sqlite3ExprListDelete(pParse->db, Y);
      A.pExpr = sqlite3PExpr(pParse, N ? TK_NE : TK_EQ, X.pExpr, pRHS, 0);
    }else{
      A.pExpr = sqlite3PExpr(pParse, TK_IN, X.pExpr, 0, 0);
      if( A.pExpr ){
        A.pExpr->x.pList = Y;
        sqlite3ExprSetHeight(pParse, A.pExpr);
      }else{
        sqlite3ExprListDelete(pParse->db, Y);
Changes to src/where.c.
4005
4006
4007
4008
4009
4010
4011


4012
4013
4014
4015
4016
4017
4018
      if( ExprHasProperty(pExpr, EP_xIsSelect) ){
        /* "x IN (SELECT ...)":  TUNING: the SELECT returns 25 rows */
        nIn = 46;  assert( 46==sqlite3LogEst(25) );
      }else if( ALWAYS(pExpr->x.pList && pExpr->x.pList->nExpr) ){
        /* "x IN (value, value, ...)" */
        nIn = sqlite3LogEst(pExpr->x.pList->nExpr);
      }


      pNew->rRun += nIn;
      pNew->u.btree.nEq++;
      pNew->nOut = nRowEst + nInMul + nIn;
    }else if( pTerm->eOperator & (WO_EQ) ){
      assert(
        (pNew->wsFlags & (WHERE_COLUMN_NULL|WHERE_COLUMN_IN|WHERE_SKIPSCAN))!=0
        || nInMul==0







>
>







4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
      if( ExprHasProperty(pExpr, EP_xIsSelect) ){
        /* "x IN (SELECT ...)":  TUNING: the SELECT returns 25 rows */
        nIn = 46;  assert( 46==sqlite3LogEst(25) );
      }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=?". */
      pNew->rRun += nIn;
      pNew->u.btree.nEq++;
      pNew->nOut = nRowEst + nInMul + nIn;
    }else if( pTerm->eOperator & (WO_EQ) ){
      assert(
        (pNew->wsFlags & (WHERE_COLUMN_NULL|WHERE_COLUMN_IN|WHERE_SKIPSCAN))!=0
        || nInMul==0
Changes to test/in4.test.
154
155
156
157
158
159
160
























































































161
162
} {}
do_test in4-3.11 {
  execsql { SELECT * FROM t3 WHERE x IN (1, 2) OR y IN ()}
} {1 1 1}
do_test in4-3.12 {
  execsql { SELECT * FROM t3 WHERE x IN (1, 2) AND y IN ()}
} {}

























































































finish_test







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


154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
} {}
do_test in4-3.11 {
  execsql { SELECT * FROM t3 WHERE x IN (1, 2) OR y IN ()}
} {1 1 1}
do_test in4-3.12 {
  execsql { SELECT * FROM t3 WHERE x IN (1, 2) AND y IN ()}
} {}

# Tests for "... IN (?)" and "... NOT IN (?)".  In other words, tests
# for when the RHS of IN is a single expression.  This should work the
# same as the == and <> operators.
#
do_execsql_test in4-3.21 {
  SELECT * FROM t3 WHERE x=10 AND y IN (10);
} {10 10 10}
do_execsql_test in4-3.22 {
  SELECT * FROM t3 WHERE x IN (10) AND y=10;
} {10 10 10}
do_execsql_test in4-3.23 {
  SELECT * FROM t3 WHERE x IN (10) AND y IN (10);
} {10 10 10}
do_execsql_test in4-3.24 {
  SELECT * FROM t3 WHERE x=1 AND y NOT IN (10);
} {1 1 1}
do_execsql_test in4-3.25 {
  SELECT * FROM t3 WHERE x  NOT IN (10) AND y=1;
} {1 1 1}
do_execsql_test in4-3.26 {
  SELECT * FROM t3 WHERE x NOT IN (10) AND y NOT IN (10);
} {1 1 1}

# The query planner recognizes that "x IN (?)" only generates a
# single match and can use this information to optimize-out ORDER BY
# clauses.
#
do_execsql_test in4-3.31 {
  DROP INDEX t3i1;
  CREATE UNIQUE INDEX t3xy ON t3(x,y);

  SELECT *, '|' FROM t3 A, t3 B
   WHERE A.x=10 AND A.y IN (10)
     AND B.x=1 AND B.y IN (1);
} {10 10 10 1 1 1 |}
do_execsql_test in4-3.32 {
  EXPLAIN QUERY PLAN
  SELECT *, '|' FROM t3 A, t3 B
   WHERE A.x=10 AND A.y IN (10)
     AND B.x=1 AND B.y IN (1);
} {~/B-TREE/}  ;# No separate sorting pass
do_execsql_test in4-3.33 {
  SELECT *, '|' FROM t3 A, t3 B
   WHERE A.x IN (10) AND A.y=10
     AND B.x IN (1) AND B.y=1;
} {10 10 10 1 1 1 |}
do_execsql_test in4-3.34 {
  EXPLAIN QUERY PLAN
  SELECT *, '|' FROM t3 A, t3 B
   WHERE A.x IN (10) AND A.y=10
     AND B.x IN (1) AND B.y=1;
} {~/B-TREE/}  ;# No separate sorting pass

# An expression of the form "x IN (?,?)" creates an ephemeral table to
# hold the list of values on the RHS.  But "x IN (?)" does not create
# an ephemeral table.
#
do_execsql_test in4-3.41 {
  SELECT * FROM t3 WHERE x IN (10,11);
} {10 10 10}
do_execsql_test in4-3.42 {
  EXPLAIN
  SELECT * FROM t3 WHERE x IN (10,11);
} {/OpenEphemeral/}
do_execsql_test in4-3.43 {
  SELECT * FROM t3 WHERE x IN (10);
} {10 10 10}
do_execsql_test in4-3.44 {
  EXPLAIN
  SELECT * FROM t3 WHERE x IN (10);
} {~/OpenEphemeral/}
do_execsql_test in4-3.45 {
  SELECT * FROM t3 WHERE x NOT IN (10,11);
} {1 1 1}
do_execsql_test in4-3.46 {
  EXPLAIN
  SELECT * FROM t3 WHERE x NOT IN (10,11);
} {/OpenEphemeral/}
do_execsql_test in4-3.47 {
  SELECT * FROM t3 WHERE x NOT IN (10);
} {1 1 1}
do_execsql_test in4-3.48 {
  EXPLAIN
  SELECT * FROM t3 WHERE x NOT IN (10);
} {~/OpenEphemeral/}



finish_test