Annotation of embedaddon/sqlite3/test/autoindex1.test, revision 1.1.1.1

1.1       misho       1: # 2010 April 07
                      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: # This file implements regression tests for SQLite library.  The
                     12: # focus of this script is testing automatic index creation logic.
                     13: #
                     14: 
                     15: set testdir [file dirname $argv0]
                     16: source $testdir/tester.tcl
                     17: 
                     18: # If the library is not compiled with automatic index support then
                     19: # skip all tests in this file.
                     20: #
                     21: ifcapable {!autoindex} {
                     22:   finish_test
                     23:   return
                     24: }
                     25: 
                     26: # With automatic index turned off, we do a full scan of the T2 table
                     27: do_test autoindex1-100 {
                     28:   db eval {
                     29:     CREATE TABLE t1(a,b);
                     30:     INSERT INTO t1 VALUES(1,11);
                     31:     INSERT INTO t1 VALUES(2,22);
                     32:     INSERT INTO t1 SELECT a+2, b+22 FROM t1;
                     33:     INSERT INTO t1 SELECT a+4, b+44 FROM t1;
                     34:     CREATE TABLE t2(c,d);
                     35:     INSERT INTO t2 SELECT a, 900+b FROM t1;
                     36:   }
                     37:   db eval {
                     38:     PRAGMA automatic_index=OFF;
                     39:     SELECT b, d FROM t1 JOIN t2 ON a=c ORDER BY b;
                     40:   }
                     41: } {11 911 22 922 33 933 44 944 55 955 66 966 77 977 88 988}
                     42: do_test autoindex1-101 {
                     43:   db status step
                     44: } {63}
                     45: do_test autoindex1-102 {
                     46:   db status autoindex
                     47: } {0}
                     48: 
                     49: # With autoindex turned on, we build an index once and then use that index
                     50: # to find T2 values.
                     51: do_test autoindex1-110 {
                     52:   db eval {
                     53:     PRAGMA automatic_index=ON;
                     54:     SELECT b, d FROM t1 JOIN t2 ON a=c ORDER BY b;
                     55:   }
                     56: } {11 911 22 922 33 933 44 944 55 955 66 966 77 977 88 988}
                     57: do_test autoindex1-111 {
                     58:   db status step
                     59: } {7}
                     60: do_test autoindex1-112 {
                     61:   db status autoindex
                     62: } {7}
                     63: 
                     64: # The same test as above, but this time the T2 query is a subquery rather
                     65: # than a join.
                     66: do_test autoindex1-200 {
                     67:   db eval {
                     68:     PRAGMA automatic_index=OFF;
                     69:     SELECT b, (SELECT d FROM t2 WHERE c=a) FROM t1;
                     70:   }
                     71: } {11 911 22 922 33 933 44 944 55 955 66 966 77 977 88 988}
                     72: do_test autoindex1-201 {
                     73:   db status step
                     74: } {35}
                     75: do_test autoindex1-202 {
                     76:   db status autoindex
                     77: } {0}
                     78: do_test autoindex1-210 {
                     79:   db eval {
                     80:     PRAGMA automatic_index=ON;
                     81:     SELECT b, (SELECT d FROM t2 WHERE c=a) FROM t1;
                     82:   }
                     83: } {11 911 22 922 33 933 44 944 55 955 66 966 77 977 88 988}
                     84: do_test autoindex1-211 {
                     85:   db status step
                     86: } {7}
                     87: do_test autoindex1-212 {
                     88:   db status autoindex
                     89: } {7}
                     90: 
                     91: 
                     92: # Modify the second table of the join while the join is in progress
                     93: #
                     94: do_test autoindex1-300 {
                     95:   set r {}
                     96:   db eval {SELECT b, d FROM t1 JOIN t2 ON (c=a)} {
                     97:     lappend r $b $d
                     98:     db eval {UPDATE t2 SET d=d+1}
                     99:   }
                    100:   set r
                    101: } {11 911 22 922 33 933 44 944 55 955 66 966 77 977 88 988}
                    102: do_test autoindex1-310 {
                    103:   db eval {SELECT d FROM t2 ORDER BY d}
                    104: } {919 930 941 952 963 974 985 996}
                    105: 
                    106: # The next test does a 10-way join on unindexed tables.  Without
                    107: # automatic indices, the join will take a long time to complete.
                    108: # With automatic indices, it should only take about a second.
                    109: #
                    110: do_test autoindex1-400 {
                    111:   db eval {
                    112:     CREATE TABLE t4(a, b);
                    113:     INSERT INTO t4 VALUES(1,2);
                    114:     INSERT INTO t4 VALUES(2,3);
                    115:   }
                    116:   for {set n 2} {$n<4096} {set n [expr {$n+$n}]} {
                    117:     db eval {INSERT INTO t4 SELECT a+$n, b+$n FROM t4}
                    118:   }
                    119:   db eval {
                    120:     SELECT count(*) FROM t4;
                    121:   }
                    122: } {4096}
                    123: do_test autoindex1-401 {
                    124:   db eval {
                    125:     SELECT count(*)
                    126:       FROM t4 AS x1
                    127:       JOIN t4 AS x2 ON x2.a=x1.b
                    128:       JOIN t4 AS x3 ON x3.a=x2.b
                    129:       JOIN t4 AS x4 ON x4.a=x3.b
                    130:       JOIN t4 AS x5 ON x5.a=x4.b
                    131:       JOIN t4 AS x6 ON x6.a=x5.b
                    132:       JOIN t4 AS x7 ON x7.a=x6.b
                    133:       JOIN t4 AS x8 ON x8.a=x7.b
                    134:       JOIN t4 AS x9 ON x9.a=x8.b
                    135:       JOIN t4 AS x10 ON x10.a=x9.b;
                    136:   }
                    137: } {4087}
                    138: 
                    139: # Ticket [8011086c85c6c404014c947fcf3eb9f42b184a0d] from 2010-07-08
                    140: # Make sure automatic indices are not created for the RHS of an IN expression
                    141: # that is not a correlated subquery.
                    142: #
                    143: do_execsql_test autoindex1-500 {
                    144:   CREATE TABLE t501(a INTEGER PRIMARY KEY, b);
                    145:   CREATE TABLE t502(x INTEGER PRIMARY KEY, y);
                    146:   EXPLAIN QUERY PLAN
                    147:   SELECT b FROM t501
                    148:    WHERE t501.a IN (SELECT x FROM t502 WHERE y=?);
                    149: } {
                    150:   0 0 0 {SEARCH TABLE t501 USING INTEGER PRIMARY KEY (rowid=?) (~25 rows)} 
                    151:   0 0 0 {EXECUTE LIST SUBQUERY 1} 
                    152:   1 0 0 {SCAN TABLE t502 (~100000 rows)}
                    153: }
                    154: do_execsql_test autoindex1-501 {
                    155:   EXPLAIN QUERY PLAN
                    156:   SELECT b FROM t501
                    157:    WHERE t501.a IN (SELECT x FROM t502 WHERE y=t501.b);
                    158: } {
                    159:   0 0 0 {SCAN TABLE t501 (~500000 rows)} 
                    160:   0 0 0 {EXECUTE CORRELATED LIST SUBQUERY 1} 
                    161:   1 0 0 {SEARCH TABLE t502 USING AUTOMATIC COVERING INDEX (y=?) (~7 rows)}
                    162: }
                    163: do_execsql_test autoindex1-502 {
                    164:   EXPLAIN QUERY PLAN
                    165:   SELECT b FROM t501
                    166:    WHERE t501.a=123
                    167:      AND t501.a IN (SELECT x FROM t502 WHERE y=t501.b);
                    168: } {
                    169:   0 0 0 {SEARCH TABLE t501 USING INTEGER PRIMARY KEY (rowid=?) (~1 rows)} 
                    170:   0 0 0 {EXECUTE CORRELATED LIST SUBQUERY 1} 
                    171:   1 0 0 {SCAN TABLE t502 (~100000 rows)}
                    172: }
                    173: 
                    174: 
                    175: # The following code checks a performance regression reported on the
                    176: # mailing list on 2010-10-19.  The problem is that the nRowEst field
                    177: # of ephermeral tables was not being initialized correctly and so no
                    178: # automatic index was being created for the emphemeral table when it was
                    179: # used as part of a join.
                    180: #
                    181: do_execsql_test autoindex1-600 {
                    182:   CREATE TABLE flock_owner(
                    183:     owner_rec_id INTEGER CONSTRAINT flock_owner_key PRIMARY KEY,
                    184:     flock_no VARCHAR(6) NOT NULL REFERENCES flock (flock_no),
                    185:     owner_person_id INTEGER NOT NULL REFERENCES person (person_id),
                    186:     owner_change_date TEXT, last_changed TEXT NOT NULL,
                    187:     CONSTRAINT fo_owner_date UNIQUE (flock_no, owner_change_date)
                    188:   );
                    189:   CREATE TABLE sheep (
                    190:     Sheep_No char(7) NOT NULL,
                    191:     Date_of_Birth char(8),
                    192:     Sort_DoB text,
                    193:     Flock_Book_Vol char(2),
                    194:     Breeder_No char(6),
                    195:     Breeder_Person integer,
                    196:     Originating_Flock char(6),
                    197:     Registering_Flock char(6),
                    198:     Tag_Prefix char(9),
                    199:     Tag_No char(15),
                    200:     Sort_Tag_No integer,
                    201:     Breeders_Temp_Tag char(15),
                    202:     Sex char(1),
                    203:     Sheep_Name char(32),
                    204:     Sire_No char(7),
                    205:     Dam_No char(7),
                    206:     Register_Code char(1),
                    207:     Colour char(48),
                    208:     Colour_Code char(2),
                    209:     Pattern_Code char(8),
                    210:     Horns char(1),
                    211:     Litter_Size char(1),
                    212:     Coeff_of_Inbreeding real,
                    213:     Date_of_Registration text,
                    214:     Date_Last_Changed text,
                    215:     UNIQUE(Sheep_No));
                    216:   CREATE INDEX fo_flock_no_index  
                    217:               ON flock_owner (flock_no);
                    218:   CREATE INDEX fo_owner_change_date_index  
                    219:               ON flock_owner (owner_change_date);
                    220:   CREATE INDEX fo_owner_person_id_index  
                    221:               ON flock_owner (owner_person_id);
                    222:   CREATE INDEX sheep_org_flock_index  
                    223:            ON sheep (originating_flock);
                    224:   CREATE INDEX sheep_reg_flock_index  
                    225:            ON sheep (registering_flock);
                    226:   EXPLAIN QUERY PLAN
                    227:   SELECT x.sheep_no, x.registering_flock, x.date_of_registration
                    228:    FROM sheep x LEFT JOIN
                    229:        (SELECT s.sheep_no, prev.flock_no, prev.owner_person_id,
                    230:        s.date_of_registration, prev.owner_change_date
                    231:        FROM sheep s JOIN flock_owner prev ON s.registering_flock =
                    232:    prev.flock_no
                    233:        AND (prev.owner_change_date <= s.date_of_registration || ' 00:00:00')
                    234:        WHERE NOT EXISTS
                    235:            (SELECT 'x' FROM flock_owner later
                    236:            WHERE prev.flock_no = later.flock_no
                    237:            AND later.owner_change_date > prev.owner_change_date
                    238:            AND later.owner_change_date <= s.date_of_registration||' 00:00:00')
                    239:        ) y ON x.sheep_no = y.sheep_no
                    240:    WHERE y.sheep_no IS NULL
                    241:    ORDER BY x.registering_flock;
                    242: } {
                    243:   1 0 0 {SCAN TABLE sheep AS s (~1000000 rows)} 
                    244:   1 1 1 {SEARCH TABLE flock_owner AS prev USING INDEX sqlite_autoindex_flock_owner_1 (flock_no=? AND owner_change_date<?) (~2 rows)} 
                    245:   1 0 0 {EXECUTE CORRELATED SCALAR SUBQUERY 2} 
                    246:   2 0 0 {SEARCH TABLE flock_owner AS later USING COVERING INDEX sqlite_autoindex_flock_owner_1 (flock_no=? AND owner_change_date>? AND owner_change_date<?) (~1 rows)} 
                    247:   0 0 0 {SCAN TABLE sheep AS x USING INDEX sheep_reg_flock_index (~1000000 rows)} 
                    248:   0 1 1 {SEARCH SUBQUERY 1 AS y USING AUTOMATIC COVERING INDEX (sheep_no=?) (~8 rows)}
                    249: }
                    250: 
                    251: 
                    252: do_execsql_test autoindex1-700 {
                    253:   CREATE TABLE t5(a, b, c);
                    254:   EXPLAIN QUERY PLAN SELECT a FROM t5 WHERE b=10 ORDER BY c;
                    255: } {
                    256:   0 0 0 {SCAN TABLE t5 (~100000 rows)} 
                    257:   0 0 0 {USE TEMP B-TREE FOR ORDER BY}
                    258: }
                    259: 
                    260: 
                    261: finish_test

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>