Debian Codegraph @

test/analyze5.test

raw
1 # 2011 January 19
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 # This file implements tests for SQLite library. The focus of the tests
13 # in this file is the use of the sqlite_stat4 histogram data on tables
14 # with many repeated values and only a few distinct values.
15 #
16
17 set testdir [file dirname $argv0]
18 source $testdir/tester.tcl
19
20 ifcapable !stat4 {
21 finish_test
22 return
23 }
24
25 set testprefix analyze5
26
27 proc eqp {sql {db db}} {
28 uplevel execsql [list "EXPLAIN QUERY PLAN $sql"] $db
29 }
30
31 proc alpha {blob} {
32 set ret ""
33 foreach c [split $blob {}] {
34 if {[string is alpha $c]} {append ret $c}
35 }
36 return $ret
37 }
38 db func alpha alpha
39
40 db func lindex lindex
41
42 unset -nocomplain i t u v w x y z
43 do_test analyze5-1.0 {
44 db eval {CREATE TABLE t1(t,u,v TEXT COLLATE nocase,w,x,y,z)}
45 for {set i 0} {$i < 1000} {incr i} {
46 set y [expr {$i>=25 && $i<=50}]
47 set z [expr {($i>=400) + ($i>=700) + ($i>=875)}]
48 set x $z
49 set w $z
50 set t [expr {$z+0.5}]
51 switch $z {
52 0 {set u "alpha"; unset x}
53 1 {set u "bravo"}
54 2 {set u "charlie"}
55 3 {set u "delta"; unset w}
56 }
57 if {$i%2} {set v $u} {set v [string toupper $u]}
58 db eval {INSERT INTO t1 VALUES($t,$u,$v,$w,$x,$y,$z)}
59 }
60 db eval {
61 CREATE INDEX t1t ON t1(t); -- 0.5, 1.5, 2.5, and 3.5
62 CREATE INDEX t1u ON t1(u); -- text
63 CREATE INDEX t1v ON t1(v); -- mixed case text
64 CREATE INDEX t1w ON t1(w); -- integers 0, 1, 2 and a few NULLs
65 CREATE INDEX t1x ON t1(x); -- integers 1, 2, 3 and many NULLs
66 CREATE INDEX t1y ON t1(y); -- integers 0 and very few 1s
67 CREATE INDEX t1z ON t1(z); -- integers 0, 1, 2, and 3
68 ANALYZE;
69 }
70 db eval {
71 SELECT DISTINCT lindex(test_decode(sample),0)
72 FROM sqlite_stat4 WHERE idx='t1u' ORDER BY nlt;
73 }
74 } {alpha bravo charlie delta}
75
76 do_test analyze5-1.1 {
77 db eval {
78 SELECT DISTINCT lower(lindex(test_decode(sample), 0))
79 FROM sqlite_stat4 WHERE idx='t1v' ORDER BY 1
80 }
81 } {alpha bravo charlie delta}
82 do_test analyze5-1.2 {
83 db eval {SELECT idx, count(*) FROM sqlite_stat4 GROUP BY 1 ORDER BY 1}
84 } {t1t 8 t1u 8 t1v 8 t1w 8 t1x 8 t1y 9 t1z 8}
85
86 # Verify that range queries generate the correct row count estimates
87 #
88 foreach {testid where index rows} {
89 1 {z>=0 AND z<=0} t1z 400
90 2 {z>=1 AND z<=1} t1z 300
91 3 {z>=2 AND z<=2} t1z 175
92 4 {z>=3 AND z<=3} t1z 125
93 5 {z>=4 AND z<=4} t1z 1
94 6 {z>=-1 AND z<=-1} t1z 1
95 7 {z>1 AND z<3} t1z 175
96 8 {z>0 AND z<100} t1z 600
97 9 {z>=1 AND z<100} t1z 600
98 10 {z>1 AND z<100} t1z 300
99 11 {z>=2 AND z<100} t1z 300
100 12 {z>2 AND z<100} t1z 125
101 13 {z>=3 AND z<100} t1z 125
102 14 {z>3 AND z<100} t1z 1
103 15 {z>=4 AND z<100} t1z 1
104 16 {z>=-100 AND z<=-1} t1z 1
105 17 {z>=-100 AND z<=0} t1z 400
106 18 {z>=-100 AND z<0} t1z 1
107 19 {z>=-100 AND z<=1} t1z 700
108 20 {z>=-100 AND z<2} t1z 700
109 21 {z>=-100 AND z<=2} t1z 875
110 22 {z>=-100 AND z<3} t1z 875
111
112 31 {z>=0.0 AND z<=0.0} t1z 400
113 32 {z>=1.0 AND z<=1.0} t1z 300
114 33 {z>=2.0 AND z<=2.0} t1z 175
115 34 {z>=3.0 AND z<=3.0} t1z 125
116 35 {z>=4.0 AND z<=4.0} t1z 1
117 36 {z>=-1.0 AND z<=-1.0} t1z 1
118 37 {z>1.5 AND z<3.0} t1z 174
119 38 {z>0.5 AND z<100} t1z 599
120 39 {z>=1.0 AND z<100} t1z 600
121 40 {z>1.5 AND z<100} t1z 299
122 41 {z>=2.0 AND z<100} t1z 300
123 42 {z>2.1 AND z<100} t1z 124
124 43 {z>=3.0 AND z<100} t1z 125
125 44 {z>3.2 AND z<100} t1z 1
126 45 {z>=4.0 AND z<100} t1z 1
127 46 {z>=-100 AND z<=-1.0} t1z 1
128 47 {z>=-100 AND z<=0.0} t1z 400
129 48 {z>=-100 AND z<0.0} t1z 1
130 49 {z>=-100 AND z<=1.0} t1z 700
131 50 {z>=-100 AND z<2.0} t1z 700
132 51 {z>=-100 AND z<=2.0} t1z 875
133 52 {z>=-100 AND z<3.0} t1z 875
134
135 101 {z=-1} t1z 1
136 102 {z=0} t1z 400
137 103 {z=1} t1z 300
138 104 {z=2} t1z 175
139 105 {z=3} t1z 125
140 106 {z=4} t1z 1
141 107 {z=-10.0} t1z 1
142 108 {z=0.0} t1z 400
143 109 {z=1.0} t1z 300
144 110 {z=2.0} t1z 175
145 111 {z=3.0} t1z 125
146 112 {z=4.0} t1z 1
147 113 {z=1.5} t1z 1
148 114 {z=2.5} t1z 1
149
150 201 {z IN (-1)} t1z 1
151 202 {z IN (0)} t1z 400
152 203 {z IN (1)} t1z 300
153 204 {z IN (2)} t1z 175
154 205 {z IN (3)} t1z 125
155 206 {z IN (4)} t1z 1
156 207 {z IN (0.5)} t1z 1
157 208 {z IN (0,1)} t1z 700
158 209 {z IN (0,1,2)} t1z 875
159 210 {z IN (0,1,2,3)} {} 100
160 211 {z IN (0,1,2,3,4,5)} {} 100
161 212 {z IN (1,2)} t1z 475
162 213 {z IN (2,3)} t1z 300
163 214 {z=3 OR z=2} t1z 300
164 215 {z IN (-1,3)} t1z 126
165 216 {z=-1 OR z=3} t1z 126
166
167 300 {y=0} t1y 974
168 301 {y=1} t1y 26
169 302 {y=0.1} t1y 1
170
171 400 {x IS NULL} t1x 400
172
173 } {
174 # Verify that the expected index is used with the expected row count
175 # No longer valid due to an EXPLAIN QUERY PLAN output format change
176 # do_test analyze5-1.${testid}a {
177 # set x [lindex [eqp "SELECT * FROM t1 WHERE $where"] 3]
178 # set idx {}
179 # regexp {INDEX (t1.) } $x all idx
180 # regexp {~([0-9]+) rows} $x all nrow
181 # list $idx $nrow
182 # } [list $index $rows]
183
184 # Verify that the same result is achieved regardless of whether or not
185 # the index is used
186 do_test analyze5-1.${testid}b {
187 set w2 [string map {y +y z +z} $where]
188 set a1 [db eval "SELECT rowid FROM t1 NOT INDEXED WHERE $w2\
189 ORDER BY +rowid"]
190 set a2 [db eval "SELECT rowid FROM t1 WHERE $where ORDER BY +rowid"]
191 if {$a1==$a2} {
192 set res ok
193 } else {
194 set res "a1=\[$a1\] a2=\[$a2\]"
195 }
196 set res
197 } {ok}
198 }
199
200 # Increase the number of NULLs in column x
201 #
202 db eval {
203 UPDATE t1 SET x=NULL;
204 UPDATE t1 SET x=rowid
205 WHERE rowid IN (SELECT rowid FROM t1 ORDER BY random() LIMIT 5);
206 ANALYZE;
207 }
208
209 # Verify that range queries generate the correct row count estimates
210 #
211 foreach {testid where index rows} {
212 500 {x IS NULL AND u='charlie'} t1u 17
213 501 {x=1 AND u='charlie'} t1x 1
214 502 {x IS NULL} t1x 995
215 503 {x=1} t1x 1
216 504 {x IS NOT NULL} t1x 2
217 505 {+x IS NOT NULL} {} 500
218 506 {upper(x) IS NOT NULL} {} 500
219
220 } {
221 # Verify that the expected index is used with the expected row count
222 # No longer valid due to an EXPLAIN QUERY PLAN format change
223 # do_test analyze5-1.${testid}a {
224 # set x [lindex [eqp "SELECT * FROM t1 WHERE $where"] 3]
225 # set idx {}
226 # regexp {INDEX (t1.) } $x all idx
227 # regexp {~([0-9]+) rows} $x all nrow
228 # list $idx $nrow
229 # } [list $index $rows]
230
231 # Verify that the same result is achieved regardless of whether or not
232 # the index is used
233 do_test analyze5-1.${testid}b {
234 set w2 [string map {y +y z +z} $where]
235 set a1 [db eval "SELECT rowid FROM t1 NOT INDEXED WHERE $w2\
236 ORDER BY +rowid"]
237 set a2 [db eval "SELECT rowid FROM t1 WHERE $where ORDER BY +rowid"]
238 if {$a1==$a2} {
239 set res ok
240 } else {
241 set res "a1=\[$a1\] a2=\[$a2\]"
242 }
243 set res
244 } {ok}
245 }
246
247 finish_test