# 2022 October 06 # # The author disclaims copyright to this source code. In place of # a legal notice, here is a blessing: # # May you do good and not evil. # May you find forgiveness for yourself and forgive others. # May you share freely, never taking more than you give. # #*********************************************************************** # # Tests for queries that use bloom filters set testdir [file dirname $argv0] source $testdir/tester.tcl source $testdir/lock_common.tcl source $testdir/malloc_common.tcl set testprefix bloom1 # Tests 1.* verify that the bloom filter code correctly handles the # case where the RHS of an (<ipk-column> = ?) expression must be coerced # to an integer before the comparison made. # do_execsql_test 1.0 { CREATE TABLE t1(a, b); CREATE TABLE t2(c INTEGER PRIMARY KEY, d); } do_execsql_test 1.1 { INSERT INTO t1 VALUES('hello', 'world'); INSERT INTO t2 VALUES(14, 'fourteen'); } do_execsql_test 1.2 { ANALYZE sqlite_schema; INSERT INTO sqlite_stat1 VALUES('t2','idx1','6 6'); ANALYZE sqlite_schema; } do_execsql_test 1.3 { SELECT 'affinity!' FROM t1 CROSS JOIN t2 WHERE t2.c = '14'; } {affinity!} reset_db do_execsql_test 1.4 { CREATE TABLE t1(a, b TEXT); CREATE TABLE t2(c INTEGER PRIMARY KEY, d); CREATE TABLE t3(e INTEGER PRIMARY KEY, f); ANALYZE sqlite_schema; INSERT INTO sqlite_stat1 VALUES('t1','idx1','600 6'); INSERT INTO sqlite_stat1 VALUES('t2','idx1','6 6'); INSERT INTO sqlite_stat1 VALUES('t3','idx2','6 6'); ANALYZE sqlite_schema; INSERT INTO t1 VALUES(1, '123'); INSERT INTO t2 VALUES(123, 'one'); INSERT INTO t3 VALUES(123, 'two'); } do_execsql_test 1.5 { SELECT 'result' FROM t1, t2, t3 WHERE t2.c=t1.b AND t2.d!='silly' AND t3.e=t1.b AND t3.f!='silly' } {result} # 2023-02-05 # https://sqlite.org/forum/forumpost/56de336385 # # Do not employ a Bloom filter if the table being filtered or any table # wo the left of the table being filtered lacks STAT1 data, since we # cannot make a good Bloom filter usefulness determination without STAT1 # data. # reset_db do_execsql_test 2.0 { CREATE TABLE objs(c INTEGER, s INTEGER, p INTEGER, o INTEGER); CREATE UNIQUE INDEX objs_cspo ON objs(o,p,c,s); ANALYZE; DELETE FROM sqlite_stat1; INSERT INTO sqlite_stat1 VALUES('objs','objs_cspo','520138 21 20 19 1'); ANALYZE sqlite_schema; } do_eqp_test 2.1 { WITH RECURSIVE transit(x) AS ( SELECT s FROM objs WHERE p=9 AND o=32805 UNION SELECT objs.s FROM objs, transit WHERE objs.p=9 AND objs.o=transit.x ) SELECT x FROM transit; } { QUERY PLAN |--CO-ROUTINE transit | |--SETUP | | `--SEARCH objs USING COVERING INDEX objs_cspo (o=? AND p=?) | `--RECURSIVE STEP | |--SCAN transit | `--SEARCH objs USING COVERING INDEX objs_cspo (o=? AND p=?) `--SCAN transit } # 2023-02-28 # https://sqlite.org/forum/forumpost/0846211821 # # Bloom filter gives an incorrect result if the collating sequence is # anything other than binary. # reset_db do_execsql_test 3.1 { CREATE TABLE t0(x TEXT COLLATE rtrim); INSERT INTO t0(x) VALUES ('a'), ('b'), ('c'); CREATE VIEW v0(y) AS SELECT DISTINCT x FROM t0; SELECT count(*) FROM t0, v0 WHERE x='b '; } 3 do_eqp_test 3.2 { SELECT count(*) FROM t0, v0 WHERE x='b '; } { QUERY PLAN |--CO-ROUTINE v0 | |--SCAN t0 | `--USE TEMP B-TREE FOR DISTINCT |--SCAN v0 `--SEARCH t0 USING AUTOMATIC PARTIAL COVERING INDEX (x=?) } # ^^^^^--- The key feature in the previous result is that no Bloom filter # is used. In the following, a Bloom filter is used because the data type # is INT instead of TEXT. do_execsql_test 3.3 { CREATE TABLE t1(x INT COLLATE rtrim); INSERT INTO t1(x) VALUES ('a'), ('b'), ('c'); CREATE VIEW v1(y) AS SELECT DISTINCT x FROM t1; SELECT count(*) FROM t1, v1 WHERE x='b '; } 3 do_eqp_test 3.4 { SELECT count(*) FROM t1, v1 WHERE x='b '; } { QUERY PLAN |--CO-ROUTINE v1 | |--SCAN t1 | `--USE TEMP B-TREE FOR DISTINCT |--SCAN v1 |--BLOOM FILTER ON t1 (x=?) `--SEARCH t1 USING AUTOMATIC PARTIAL COVERING INDEX (x=?) } # 2023-03-14 # https://sqlite.org/forum/forumpost/d47a0e8e3a # https://sqlite.org/forum/forumpost/2e427099d5 # # Both reports are for the same problem - using a Bloom filter on an # expression index can cause issues. # reset_db do_execsql_test 4.1 { CREATE TABLE t1(x TEXT, y INT, z TEXT); INSERT INTO t1(rowid,x,y,z) VALUES(12,'aa','bb','aa'); CREATE INDEX i1x ON t1(1 IS true,z); CREATE TABLE t0(x TEXT); INSERT INTO t0(rowid,x) VALUES(4,'aa'); ANALYZE sqlite_schema; INSERT INTO sqlite_stat1 VALUES('t0',NULL,'20'); INSERT INTO sqlite_stat1 VALUES('t1','i1x','18 18 2'); ANALYZE sqlite_schema; } do_execsql_test 4.2 { SELECT * FROM t0 NATURAL JOIN t1 WHERE z=t1.x; } {aa bb aa} do_execsql_test 4.3 { DROP TABLE t0; CREATE TABLE t0(a TEXT); INSERT INTO t0 VALUES ('xyz'); CREATE INDEX t0x ON t0(a IS FALSE) WHERE false; DROP TABLE t1; CREATE TABLE t1(b INT); INSERT INTO t1 VALUES('aaa'),('bbb'),('ccc'),('ddd'),(NULL); CREATE TABLE t2(c REAL); INSERT INTO t2 VALUES(7); ANALYZE; CREATE INDEX t2x ON t2(true IN ()); } do_execsql_test 4.4 { SELECT * FROM t0 LEFT JOIN t1 LEFT JOIN t2 ON (b NOTNULL)==(c IN ()) WHERE c; } {xyz {} 7.0} finish_test