| 1 |
# 2010 September 24 |
| 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 to verify that the "testable statements" in |
| 13 |
# the lang_vacuum.html document are correct. |
| 14 |
# |
| 15 |
|
| 16 |
set testdir [file dirname $argv0] |
| 17 |
source $testdir/tester.tcl |
| 18 |
|
| 19 |
sqlite3_test_control_pending_byte 0x1000000 |
| 20 |
|
| 21 |
proc create_db {{sql ""}} { |
| 22 |
catch { db close } |
| 23 |
forcedelete test.db |
| 24 |
sqlite3 db test.db |
| 25 |
|
| 26 |
db transaction { |
| 27 |
execsql { PRAGMA page_size = 1024; } |
| 28 |
execsql $sql |
| 29 |
execsql { |
| 30 |
CREATE TABLE t1(a PRIMARY KEY, b UNIQUE); |
| 31 |
INSERT INTO t1 VALUES(1, randomblob(400)); |
| 32 |
INSERT INTO t1 SELECT a+1, randomblob(400) FROM t1; |
| 33 |
INSERT INTO t1 SELECT a+2, randomblob(400) FROM t1; |
| 34 |
INSERT INTO t1 SELECT a+4, randomblob(400) FROM t1; |
| 35 |
INSERT INTO t1 SELECT a+8, randomblob(400) FROM t1; |
| 36 |
INSERT INTO t1 SELECT a+16, randomblob(400) FROM t1; |
| 37 |
INSERT INTO t1 SELECT a+32, randomblob(400) FROM t1; |
| 38 |
INSERT INTO t1 SELECT a+64, randomblob(400) FROM t1; |
| 39 |
|
| 40 |
CREATE TABLE t2(a PRIMARY KEY, b UNIQUE); |
| 41 |
INSERT INTO t2 SELECT * FROM t1; |
| 42 |
} |
| 43 |
} |
| 44 |
|
| 45 |
return [expr {[file size test.db] / 1024}] |
| 46 |
} |
| 47 |
|
| 48 |
# This proc returns the number of contiguous blocks of pages that make up |
| 49 |
# the table or index named by the only argument. For example, if the table |
| 50 |
# occupies database pages 3, 4, 8 and 9, then this command returns 2 (there |
| 51 |
# are 2 fragments - one consisting of pages 3 and 4, the other of fragments |
| 52 |
# 8 and 9). |
| 53 |
# |
| 54 |
proc fragment_count {name} { |
| 55 |
execsql { CREATE VIRTUAL TABLE temp.stat USING dbstat } |
| 56 |
set nFrag 1 |
| 57 |
db eval {SELECT pageno FROM stat WHERE name = 't1' ORDER BY pageno} { |
| 58 |
if {[info exists prevpageno] && $prevpageno != $pageno-1} { |
| 59 |
incr nFrag |
| 60 |
} |
| 61 |
set prevpageno $pageno |
| 62 |
} |
| 63 |
execsql { DROP TABLE temp.stat } |
| 64 |
set nFrag |
| 65 |
} |
| 66 |
|
| 67 |
|
| 68 |
# -- syntax diagram vacuum-stmt |
| 69 |
# |
| 70 |
do_execsql_test e_vacuum-0.1 { VACUUM } {} |
| 71 |
|
| 72 |
# EVIDENCE-OF: R-51469-36013 Unless SQLite is running in |
| 73 |
# "auto_vacuum=FULL" mode, when a large amount of data is deleted from |
| 74 |
# the database file it leaves behind empty space, or "free" database |
| 75 |
# pages. |
| 76 |
# |
| 77 |
# EVIDENCE-OF: R-60541-63059 Running VACUUM to rebuild the database |
| 78 |
# reclaims this space and reduces the size of the database file. |
| 79 |
# |
| 80 |
foreach {tn avmode sz} { |
| 81 |
1 none 7 |
| 82 |
2 full 8 |
| 83 |
3 incremental 8 |
| 84 |
} { |
| 85 |
set nPage [create_db "PRAGMA auto_vacuum = $avmode"] |
| 86 |
|
| 87 |
do_execsql_test e_vacuum-1.1.$tn.1 { |
| 88 |
DELETE FROM t1; |
| 89 |
DELETE FROM t2; |
| 90 |
} {} |
| 91 |
|
| 92 |
if {$avmode == "full"} { |
| 93 |
# This branch tests the "unless ... auto_vacuum=FULL" in the requirement |
| 94 |
# above. If auto_vacuum is set to FULL, then no empty space is left in |
| 95 |
# the database file. |
| 96 |
do_execsql_test e_vacuum-1.1.$tn.2 {PRAGMA freelist_count} 0 |
| 97 |
} else { |
| 98 |
set freelist [expr {$nPage - $sz}] |
| 99 |
if {$avmode == "incremental"} { |
| 100 |
# The page size is 1024 bytes. Therefore, assuming the database contains |
| 101 |
# somewhere between 207 and 411 pages (it does), there are 2 pointer-map |
| 102 |
# pages. |
| 103 |
incr freelist -2 |
| 104 |
} |
| 105 |
do_execsql_test e_vacuum-1.1.$tn.3 {PRAGMA freelist_count} $freelist |
| 106 |
do_execsql_test e_vacuum-1.1.$tn.4 {VACUUM} {} |
| 107 |
} |
| 108 |
|
| 109 |
do_test e_vacuum-1.1.$tn.5 { expr {[file size test.db] / 1024} } $sz |
| 110 |
} |
| 111 |
|
| 112 |
# EVIDENCE-OF: R-50943-18433 Frequent inserts, updates, and deletes can |
| 113 |
# cause the database file to become fragmented - where data for a single |
| 114 |
# table or index is scattered around the database file. |
| 115 |
# |
| 116 |
# EVIDENCE-OF: R-05791-54928 Running VACUUM ensures that each table and |
| 117 |
# index is largely stored contiguously within the database file. |
| 118 |
# |
| 119 |
# e_vacuum-1.2.1 - Perform many INSERT, UPDATE and DELETE ops on table t1. |
| 120 |
# e_vacuum-1.2.2 - Verify that t1 and its indexes are now quite fragmented. |
| 121 |
# e_vacuum-1.2.3 - Run VACUUM. |
| 122 |
# e_vacuum-1.2.4 - Verify that t1 and its indexes are now much |
| 123 |
# less fragmented. |
| 124 |
# |
| 125 |
ifcapable vtab&&compound { |
| 126 |
create_db |
| 127 |
register_dbstat_vtab db |
| 128 |
do_execsql_test e_vacuum-1.2.1 { |
| 129 |
DELETE FROM t1 WHERE a%2; |
| 130 |
INSERT INTO t1 SELECT b, a FROM t2 WHERE a%2; |
| 131 |
UPDATE t1 SET b=randomblob(600) WHERE (a%2)==0; |
| 132 |
} {} |
| 133 |
|
| 134 |
do_test e_vacuum-1.2.2.1 { expr [fragment_count t1]>100 } 1 |
| 135 |
do_test e_vacuum-1.2.2.2 { expr [fragment_count sqlite_autoindex_t1_1]>100 } 1 |
| 136 |
do_test e_vacuum-1.2.2.3 { expr [fragment_count sqlite_autoindex_t1_2]>100 } 1 |
| 137 |
|
| 138 |
do_execsql_test e_vacuum-1.2.3 { VACUUM } {} |
| 139 |
|
| 140 |
# In practice, the tables and indexes each end up stored as two fragments - |
| 141 |
# one containing the root page and another containing all other pages. |
| 142 |
# |
| 143 |
do_test e_vacuum-1.2.4.1 { fragment_count t1 } 2 |
| 144 |
do_test e_vacuum-1.2.4.2 { fragment_count sqlite_autoindex_t1_1 } 2 |
| 145 |
do_test e_vacuum-1.2.4.3 { fragment_count sqlite_autoindex_t1_2 } 2 |
| 146 |
} |
| 147 |
|
| 148 |
# EVIDENCE-OF: R-20474-44465 Normally, the database page_size and |
| 149 |
# whether or not the database supports auto_vacuum must be configured |
| 150 |
# before the database file is actually created. |
| 151 |
# |
| 152 |
do_test e_vacuum-1.3.1.1 { |
| 153 |
create_db "PRAGMA page_size = 1024 ; PRAGMA auto_vacuum = FULL" |
| 154 |
execsql { PRAGMA page_size ; PRAGMA auto_vacuum } |
| 155 |
} {1024 1} |
| 156 |
do_test e_vacuum-1.3.1.2 { |
| 157 |
execsql { PRAGMA page_size = 2048 } |
| 158 |
execsql { PRAGMA auto_vacuum = NONE } |
| 159 |
execsql { PRAGMA page_size ; PRAGMA auto_vacuum } |
| 160 |
} {1024 1} |
| 161 |
|
| 162 |
if {![nonzero_reserved_bytes]} { |
| 163 |
# EVIDENCE-OF: R-08570-19916 However, when not in write-ahead log mode, |
| 164 |
# the page_size and/or auto_vacuum properties of an existing database |
| 165 |
# may be changed by using the page_size and/or pragma auto_vacuum |
| 166 |
# pragmas and then immediately VACUUMing the database. |
| 167 |
# |
| 168 |
do_test e_vacuum-1.3.2.1 { |
| 169 |
execsql { PRAGMA journal_mode = delete } |
| 170 |
execsql { PRAGMA page_size = 2048 } |
| 171 |
execsql { PRAGMA auto_vacuum = NONE } |
| 172 |
execsql VACUUM |
| 173 |
execsql { PRAGMA page_size ; PRAGMA auto_vacuum } |
| 174 |
} {2048 0} |
| 175 |
|
| 176 |
# EVIDENCE-OF: R-48521-51450 When in write-ahead log mode, only the |
| 177 |
# auto_vacuum support property can be changed using VACUUM. |
| 178 |
# |
| 179 |
if {[wal_is_capable]} { |
| 180 |
do_test e_vacuum-1.3.3.1 { |
| 181 |
execsql { PRAGMA journal_mode = wal } |
| 182 |
execsql { PRAGMA page_size ; PRAGMA auto_vacuum } |
| 183 |
} {2048 0} |
| 184 |
do_test e_vacuum-1.3.3.2 { |
| 185 |
execsql { PRAGMA page_size = 1024 } |
| 186 |
execsql { PRAGMA auto_vacuum = FULL } |
| 187 |
execsql VACUUM |
| 188 |
execsql { PRAGMA page_size ; PRAGMA auto_vacuum } |
| 189 |
} {2048 1} |
| 190 |
} |
| 191 |
} |
| 192 |
|
| 193 |
# EVIDENCE-OF: R-40347-36128 By default, VACUUM operates on the main |
| 194 |
# database. |
| 195 |
forcedelete test.db2 |
| 196 |
create_db { PRAGMA auto_vacuum = NONE } |
| 197 |
do_execsql_test e_vacuum-2.1.1 { |
| 198 |
ATTACH 'test.db2' AS aux; |
| 199 |
PRAGMA aux.page_size = 1024; |
| 200 |
CREATE TABLE aux.t3 AS SELECT * FROM t1; |
| 201 |
DELETE FROM t3; |
| 202 |
} {} |
| 203 |
set original_size [file size test.db2] |
| 204 |
|
| 205 |
# Vacuuming the main database does not affect aux |
| 206 |
do_execsql_test e_vacuum-2.1.3 { VACUUM } {} |
| 207 |
do_test e_vacuum-2.1.6 { expr {[file size test.db2]==$::original_size} } 1 |
| 208 |
|
| 209 |
# EVIDENCE-OF: R-36598-60500 Attached databases can be vacuumed by |
| 210 |
# appending the appropriate schema-name to the VACUUM statement. |
| 211 |
do_execsql_test e_vacuum-2.1.7 { VACUUM aux; } {} |
| 212 |
do_test e_vacuum-2.1.8 { expr {[file size test.db2]<$::original_size} } 1 |
| 213 |
|
| 214 |
# EVIDENCE-OF: R-17495-17419 The VACUUM command may change the ROWIDs of |
| 215 |
# entries in any tables that do not have an explicit INTEGER PRIMARY |
| 216 |
# KEY. |
| 217 |
# |
| 218 |
# Tests e_vacuum-3.1.1 - 3.1.2 demonstrate that rowids can change when |
| 219 |
# a database is VACUUMed. Tests e_vacuum-3.1.3 - 3.1.4 show that adding |
| 220 |
# an INTEGER PRIMARY KEY column to a table stops this from happening. |
| 221 |
# |
| 222 |
# Update 2019-01-07: Rowids are now preserved by VACUUM. |
| 223 |
# |
| 224 |
do_execsql_test e_vacuum-3.1.1 { |
| 225 |
CREATE TABLE t4(x); |
| 226 |
INSERT INTO t4(x) VALUES('x'); |
| 227 |
INSERT INTO t4(x) VALUES('y'); |
| 228 |
INSERT INTO t4(x) VALUES('z'); |
| 229 |
DELETE FROM t4 WHERE x = 'y'; |
| 230 |
SELECT rowid, x FROM t4; |
| 231 |
} {1 x 3 z} |
| 232 |
do_execsql_test e_vacuum-3.1.2 { |
| 233 |
VACUUM; |
| 234 |
SELECT rowid, x FROM t4; |
| 235 |
} {1 x 2 z} |
| 236 |
|
| 237 |
# Rowids are preserved if an INTEGER PRIMARY KEY is used |
| 238 |
do_execsql_test e_vacuum-3.1.3 { |
| 239 |
CREATE TABLE t5(x, y INTEGER PRIMARY KEY); |
| 240 |
INSERT INTO t5(x) VALUES('x'); |
| 241 |
INSERT INTO t5(x) VALUES('y'); |
| 242 |
INSERT INTO t5(x) VALUES('z'); |
| 243 |
DELETE FROM t5 WHERE x = 'y'; |
| 244 |
SELECT rowid, x FROM t5; |
| 245 |
} {1 x 3 z} |
| 246 |
do_execsql_test e_vacuum-3.1.4 { |
| 247 |
VACUUM; |
| 248 |
SELECT rowid, x FROM t5; |
| 249 |
} {1 x 3 z} |
| 250 |
|
| 251 |
# Rowid is preserved for VACUUM INTO |
| 252 |
do_execsql_test e_vacuum-3.1.5 { |
| 253 |
DROP TABLE t5; |
| 254 |
CREATE TABLE t5(x); |
| 255 |
INSERT INTO t5(x) VALUES('x'); |
| 256 |
INSERT INTO t5(x) VALUES('y'); |
| 257 |
INSERT INTO t5(x) VALUES('z'); |
| 258 |
DELETE FROM t5 WHERE x = 'y'; |
| 259 |
SELECT rowid, x FROM t5; |
| 260 |
} {1 x 3 z} |
| 261 |
forcedelete test2.db |
| 262 |
do_execsql_test e_vacuum-3.1.6 { |
| 263 |
VACUUM INTO 'test2.db'; |
| 264 |
ATTACH 'test2.db' AS aux1; |
| 265 |
SELECT rowid, x FROM aux1.t5; |
| 266 |
DETACH aux1; |
| 267 |
} {1 x 3 z} |
| 268 |
|
| 269 |
# Rowids are not renumbered if the table being vacuumed |
| 270 |
# has indexes. |
| 271 |
do_execsql_test e_vacuum-3.1.7 { |
| 272 |
DROP TABLE t5; |
| 273 |
CREATE TABLE t5(x,y,z); |
| 274 |
INSERT INTO t5(x) VALUES('x'); |
| 275 |
INSERT INTO t5(x) VALUES('y'); |
| 276 |
INSERT INTO t5(x) VALUES('z'); |
| 277 |
UPDATE t5 SET y=x, z=random(); |
| 278 |
DELETE FROM t5 WHERE x = 'y'; |
| 279 |
CREATE INDEX t5x ON t5(x); |
| 280 |
CREATE UNIQUE INDEX t5y ON t5(y); |
| 281 |
CREATE INDEX t5zxy ON t5(z,x,y); |
| 282 |
SELECT rowid, x FROM t5; |
| 283 |
} {1 x 3 z} |
| 284 |
do_execsql_test e_vacuum-3.1.8 { |
| 285 |
VACUUM; |
| 286 |
SELECT rowid, x FROM t5; |
| 287 |
} {1 x 3 z} |
| 288 |
|
| 289 |
# EVIDENCE-OF: R-12218-18073 A VACUUM will fail if there is an open |
| 290 |
# transaction on the database connection that is attempting to run the |
| 291 |
# VACUUM. |
| 292 |
# |
| 293 |
do_execsql_test e_vacuum-3.2.1.1 { BEGIN } {} |
| 294 |
do_catchsql_test e_vacuum-3.2.1.2 { |
| 295 |
VACUUM |
| 296 |
} {1 {cannot VACUUM from within a transaction}} |
| 297 |
do_execsql_test e_vacuum-3.2.1.3 { COMMIT } {} |
| 298 |
do_execsql_test e_vacuum-3.2.1.4 { VACUUM } {} |
| 299 |
do_execsql_test e_vacuum-3.2.1.5 { SAVEPOINT x } {} |
| 300 |
do_catchsql_test e_vacuum-3.2.1.6 { |
| 301 |
VACUUM |
| 302 |
} {1 {cannot VACUUM from within a transaction}} |
| 303 |
do_execsql_test e_vacuum-3.2.1.7 { COMMIT } {} |
| 304 |
do_execsql_test e_vacuum-3.2.1.8 { VACUUM } {} |
| 305 |
|
| 306 |
create_db |
| 307 |
do_test e_vacuum-3.2.2.1 { |
| 308 |
set res "" |
| 309 |
db eval { SELECT a FROM t1 } { |
| 310 |
if {$a == 10} { set res [catchsql VACUUM] } |
| 311 |
} |
| 312 |
set res |
| 313 |
} {1 {cannot VACUUM - SQL statements in progress}} |
| 314 |
|
| 315 |
|
| 316 |
# EVIDENCE-OF: R-55138-13241 An alternative to using the VACUUM command |
| 317 |
# to reclaim space after data has been deleted is auto-vacuum mode, |
| 318 |
# enabled using the auto_vacuum pragma. |
| 319 |
# |
| 320 |
do_test e_vacuum-3.3.1 { |
| 321 |
create_db { PRAGMA auto_vacuum = FULL } |
| 322 |
execsql { PRAGMA auto_vacuum } |
| 323 |
} {1} |
| 324 |
|
| 325 |
# EVIDENCE-OF: R-64844-34873 When auto_vacuum is enabled for a database |
| 326 |
# free pages may be reclaimed after deleting data, causing the file to |
| 327 |
# shrink, without rebuilding the entire database using VACUUM. |
| 328 |
# |
| 329 |
do_test e_vacuum-3.3.2.1 { |
| 330 |
create_db { PRAGMA auto_vacuum = FULL } |
| 331 |
execsql { |
| 332 |
DELETE FROM t1; |
| 333 |
DELETE FROM t2; |
| 334 |
} |
| 335 |
expr {[file size test.db] / 1024} |
| 336 |
} {8} |
| 337 |
do_test e_vacuum-3.3.2.2 { |
| 338 |
create_db { PRAGMA auto_vacuum = INCREMENTAL } |
| 339 |
execsql { |
| 340 |
DELETE FROM t1; |
| 341 |
DELETE FROM t2; |
| 342 |
PRAGMA incremental_vacuum; |
| 343 |
} |
| 344 |
expr {[file size test.db] / 1024} |
| 345 |
} {8} |
| 346 |
|
| 347 |
finish_test |