| 1 |
# 2017 April 30 |
| 2 |
# |
| 3 |
# The author disclaims copyright to this source code. In place of |
| 4 |
# a legal notice, here is a blessing: |
| 5 |
# |
| 6 |
# May you do good and not evil. |
| 7 |
# May you find forgiveness for yourself and forgive others. |
| 8 |
# May you share freely, never taking more than you give. |
| 9 |
# |
| 10 |
#*********************************************************************** |
| 11 |
# |
| 12 |
# Test the HAVING->WHERE optimization. |
| 13 |
# |
| 14 |
|
| 15 |
set testdir [file dirname $argv0] |
| 16 |
source $testdir/tester.tcl |
| 17 |
set testprefix having |
| 18 |
|
| 19 |
do_execsql_test 1.0 { |
| 20 |
CREATE TABLE t2(c, d); |
| 21 |
|
| 22 |
CREATE TABLE t1(a, b); |
| 23 |
INSERT INTO t1 VALUES(1, 1); |
| 24 |
INSERT INTO t1 VALUES(2, 2); |
| 25 |
INSERT INTO t1 VALUES(1, 3); |
| 26 |
INSERT INTO t1 VALUES(2, 4); |
| 27 |
INSERT INTO t1 VALUES(1, 5); |
| 28 |
INSERT INTO t1 VALUES(2, 6); |
| 29 |
} {} |
| 30 |
|
| 31 |
foreach {tn sql res} { |
| 32 |
1 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING a=2" {2 12} |
| 33 |
2 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING a=2 AND sum(b)>10" {2 12} |
| 34 |
3 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING sum(b)>12" {} |
| 35 |
} { |
| 36 |
do_execsql_test 1.$tn $sql $res |
| 37 |
} |
| 38 |
|
| 39 |
# Run an EXPLAIN command for both SQL statements. Return true if |
| 40 |
# the outputs are identical, or false otherwise. |
| 41 |
# |
| 42 |
proc compare_vdbe {sql1 sql2} { |
| 43 |
set r1 [list] |
| 44 |
set r2 [list] |
| 45 |
db eval "explain $sql1" { lappend r1 $opcode $p1 $p2 $p3 $p4 $p5} |
| 46 |
db eval "explain $sql2" { lappend r2 $opcode $p1 $p2 $p3 $p4 $p5} |
| 47 |
return [expr {$r1==$r2}] |
| 48 |
} |
| 49 |
|
| 50 |
proc do_compare_vdbe_test {tn sql1 sql2 res} { |
| 51 |
uplevel [list do_test $tn [list compare_vdbe $sql1 $sql2] $res] |
| 52 |
} |
| 53 |
|
| 54 |
#------------------------------------------------------------------------- |
| 55 |
# Test that various statements that are eligible for the optimization |
| 56 |
# produce the same VDBE code as optimizing by hand does. |
| 57 |
# |
| 58 |
foreach {tn sql1 sql2} { |
| 59 |
1 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING a=2" |
| 60 |
"SELECT a, sum(b) FROM t1 WHERE a=2 GROUP BY a" |
| 61 |
|
| 62 |
2 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING sum(b)>5 AND a=2" |
| 63 |
"SELECT a, sum(b) FROM t1 WHERE a=2 GROUP BY a HAVING sum(b)>5" |
| 64 |
|
| 65 |
3 "SELECT a, sum(b) FROM t1 GROUP BY a COLLATE binary HAVING a=2" |
| 66 |
"SELECT a, sum(b) FROM t1 WHERE a=2 GROUP BY a COLLATE binary" |
| 67 |
|
| 68 |
5 "SELECT a, sum(b) FROM t1 GROUP BY a COLLATE binary HAVING 1" |
| 69 |
"SELECT a, sum(b) FROM t1 WHERE 1 GROUP BY a COLLATE binary" |
| 70 |
|
| 71 |
6 "SELECT count(*) FROM t1,t2 WHERE a=c GROUP BY b, d HAVING b=d" |
| 72 |
"SELECT count(*) FROM t1,t2 WHERE a=c AND b=d GROUP BY b, d" |
| 73 |
|
| 74 |
7 { |
| 75 |
SELECT count(*) FROM t1,t2 WHERE a=c GROUP BY b, d |
| 76 |
HAVING b=d COLLATE nocase |
| 77 |
} { |
| 78 |
SELECT count(*) FROM t1,t2 WHERE a=c AND b=d COLLATE nocase |
| 79 |
GROUP BY b, d |
| 80 |
} |
| 81 |
|
| 82 |
8 "SELECT a, sum(b) FROM t1 GROUP BY a||b HAVING substr(a||b, 1, 1)='a'" |
| 83 |
"SELECT a, sum(b) FROM t1 WHERE substr(a||b, 1, 1)='a' GROUP BY a||b" |
| 84 |
} { |
| 85 |
do_compare_vdbe_test 2.$tn $sql1 $sql2 1 |
| 86 |
} |
| 87 |
|
| 88 |
# The (4) test in the above set used to generate identical bytecode, but |
| 89 |
# that is no longer the case. The byte code is equivalent, though. |
| 90 |
# |
| 91 |
do_execsql_test 2.4a { |
| 92 |
SELECT x,y FROM ( |
| 93 |
SELECT a AS x, sum(b) AS y FROM t1 |
| 94 |
GROUP BY a |
| 95 |
) WHERE x BETWEEN 2 AND 9999 |
| 96 |
} {2 12} |
| 97 |
do_execsql_test 2.4b { |
| 98 |
SELECT x,y FROM ( |
| 99 |
SELECT a AS x, sum(b) AS y FROM t1 |
| 100 |
WHERE x BETWEEN 2 AND 9999 |
| 101 |
GROUP BY a |
| 102 |
) |
| 103 |
} {2 12} |
| 104 |
|
| 105 |
|
| 106 |
#------------------------------------------------------------------------- |
| 107 |
# 1: Test that the optimization is only applied if the GROUP BY term |
| 108 |
# uses BINARY collation. |
| 109 |
# |
| 110 |
# 2: Not applied if there is a non-deterministic function in the HAVING |
| 111 |
# term. |
| 112 |
# |
| 113 |
foreach {tn sql1 sql2} { |
| 114 |
1 "SELECT a, sum(b) FROM t1 GROUP BY a COLLATE nocase HAVING a=2" |
| 115 |
"SELECT a, sum(b) FROM t1 WHERE a=2 GROUP BY a COLLATE nocase" |
| 116 |
|
| 117 |
2 "SELECT a, sum(b) FROM t1 GROUP BY a HAVING randomblob(a)<X'88'" |
| 118 |
"SELECT a, sum(b) FROM t1 WHERE randomblob(a)<X'88' GROUP BY a" |
| 119 |
} { |
| 120 |
do_compare_vdbe_test 3.$tn $sql1 $sql2 0 |
| 121 |
} |
| 122 |
|
| 123 |
|
| 124 |
#------------------------------------------------------------------------- |
| 125 |
# Test that non-deterministic functions disqualify a term from being |
| 126 |
# moved from the HAVING to WHERE clause. |
| 127 |
# |
| 128 |
do_execsql_test 4.1 { |
| 129 |
CREATE TABLE t3(a, b); |
| 130 |
INSERT INTO t3 VALUES(1, 1); |
| 131 |
INSERT INTO t3 VALUES(1, 2); |
| 132 |
INSERT INTO t3 VALUES(1, 3); |
| 133 |
INSERT INTO t3 VALUES(2, 1); |
| 134 |
INSERT INTO t3 VALUES(2, 2); |
| 135 |
INSERT INTO t3 VALUES(2, 3); |
| 136 |
} |
| 137 |
|
| 138 |
proc nondeter {args} { |
| 139 |
incr ::nondeter_ret |
| 140 |
expr {$::nondeter_ret % 2} |
| 141 |
} |
| 142 |
db func nondeter nondeter |
| 143 |
|
| 144 |
set ::nondeter_ret 0 |
| 145 |
do_execsql_test 4.2 { |
| 146 |
SELECT a, sum(b) FROM t3 GROUP BY a HAVING nondeter(a) |
| 147 |
} {1 6} |
| 148 |
|
| 149 |
# If the term where moved, the query above would return the same |
| 150 |
# result as the following. But it does not. |
| 151 |
# |
| 152 |
set ::nondeter_ret 0 |
| 153 |
do_execsql_test 4.3 { |
| 154 |
SELECT a, sum(b) FROM t3 WHERE nondeter(a) GROUP BY a |
| 155 |
} {1 4 2 2} |
| 156 |
|
| 157 |
#------------------------------------------------------------------------- |
| 158 |
reset_db |
| 159 |
do_execsql_test 5.0 { |
| 160 |
CREATE TABLE t1(a, b); |
| 161 |
CREATE TABLE t2(x, y); |
| 162 |
INSERT INTO t1 VALUES('a', 'b'); |
| 163 |
} |
| 164 |
|
| 165 |
# The WHERE clause (a=2), uses an aggregate column from the outer query. |
| 166 |
# If the HAVING term (0) is moved into the WHERE clause in this case, |
| 167 |
# SQLite would at one point optimize (a=2 AND 0) to simply (0). Which |
| 168 |
# is logically correct, but happened to cause problems in aggregate |
| 169 |
# processing for the outer query. This test case verifies that those |
| 170 |
# problems are no longer present. |
| 171 |
do_execsql_test 5.1 { |
| 172 |
SELECT min(b), ( |
| 173 |
SELECT x FROM t2 WHERE a=2 GROUP BY y HAVING 0 |
| 174 |
) FROM t1; |
| 175 |
} {b {}} |
| 176 |
|
| 177 |
# From chromium |
| 178 |
# https://bugs.chromium.org/p/chromium/issues/detail?id=1161869 |
| 179 |
# |
| 180 |
do_execsql_test 5.2 { |
| 181 |
SELECT EXISTS ( |
| 182 |
SELECT * FROM ( |
| 183 |
SELECT * FROM ( |
| 184 |
SELECT 1 |
| 185 |
) WHERE Col0 = 1 GROUP BY 1 |
| 186 |
) WHERE 0 |
| 187 |
) |
| 188 |
FROM (SELECT 1 Col0) GROUP BY 1 |
| 189 |
} {0} |
| 190 |
|
| 191 |
finish_test |