sqlite3/www/speed.tcl

   1 #
   2 # Run this Tcl script to generate the speed.html file.
   3 #
   4 set rcsid {$Id: speed.tcl,v 1.17 2005/03/12 15:55:11 drh Exp $ }
   5 source common.tcl
   6 header {SQLite Database Speed Comparison}
   7
   8 puts {
   9 <h2>Database Speed Comparison</h2>
  10
  11 <font color="red"><b>
  12 Note:  This document is old.  It describes a speed comparison between
  13 an older version of SQLite against archaic versions of MySQL and PostgreSQL.
  14 Readers are invited to contribute more up-to-date speed comparisons
  15 on the <a href="http://www.sqlite.org/cvstrac/wiki">SQLite Wiki</a>.
  16 <p>
  17 The numbers here are old enough to be nearly meaningless.  Until it is
  18 updated, use this document only as proof that SQLite is not a
  19 sluggard.
  20 </b></font>
  21
  22 <h3>Executive Summary</h3>
  23
  24 <p>A series of tests were run to measure the relative performance of
  25 SQLite 2.7.6, PostgreSQL 7.1.3, and MySQL 3.23.41.
  26 The following are general
  27 conclusions drawn from these experiments:
  28 </p>
  29
  30 <ul>
  31 <li><p>
  32   SQLite 2.7.6 is significantly faster (sometimes as much as 10 or
  33   20 times faster) than the default PostgreSQL 7.1.3 installation
  34   on RedHat 7.2 for most common operations.
  35 </p></li>
  36 <li><p>
  37   SQLite 2.7.6 is often faster (sometimes
  38   more than twice as fast) than MySQL 3.23.41
  39   for most common operations.
  40 </p></li>
  41 <li><p>
  42   SQLite does not execute CREATE INDEX or DROP TABLE as fast as
  43   the other databases.  But this is not seen as a problem because
  44   those are infrequent operations.
  45 </p></li>
  46 <li><p>
  47   SQLite works best if you group multiple operations together into
  48   a single transaction.
  49 </p></li>
  50 </ul>
  51
  52 <p>
  53 The results presented here come with the following caveats:
  54 </p>
  55
  56 <ul>
  57 <li><p>
  58   These tests did not attempt to measure multi-user performance or
  59   optimization of complex queries involving multiple joins and subqueries.
  60 </p></li>
  61 <li><p>
  62   These tests are on a relatively small (approximately 14 megabyte) database.
  63   They do not measure how well the database engines scale to larger problems.
  64 </p></li>
  65 </ul>
  66
  67 <h3>Test Environment</h3>
  68
  69 <p>
  70 The platform used for these tests is a 1.6GHz Athlon with 1GB or memory
  71 and an IDE disk drive.  The operating system is RedHat Linux 7.2 with
  72 a stock kernel.
  73 </p>
  74
  75 <p>
  76 The PostgreSQL and MySQL servers used were as delivered by default on
  77 RedHat 7.2.  (PostgreSQL version 7.1.3 and MySQL version 3.23.41.)
  78 No effort was made to tune these engines.  Note in particular
  79 the the default MySQL configuration on RedHat 7.2 does not support
  80 transactions.  Not having to support transactions gives MySQL a
  81 big speed advantage, but SQLite is still able to hold its own on most
  82 tests.
  83 </p>
  84
  85 <p>
  86 I am told that the default PostgreSQL configuration in RedHat 7.3
  87 is unnecessarily conservative (it is designed to
  88 work on a machine with 8MB of RAM) and that PostgreSQL could
  89 be made to run a lot faster with some knowledgeable configuration
  90 tuning.
  91 Matt Sergeant reports that he has tuned his PostgreSQL installation
  92 and rerun the tests shown below.  His results show that
  93 PostgreSQL and MySQL run at about the same speed.  For Matt's
  94 results, visit
  95 </p>
  96
  97 <blockquote>
  98 <a href="http://www.sergeant.org/sqlite_vs_pgsync.html">
  99 http://www.sergeant.org/sqlite_vs_pgsync.html</a>
 100 </blockquote>
 101
 102 <p>
 103 SQLite was tested in the same configuration that it appears
 104 on the website.  It was compiled with -O6 optimization and with
 105 the -DNDEBUG=1 switch which disables the many "assert()" statements
 106 in the SQLite code.  The -DNDEBUG=1 compiler option roughly doubles
 107 the speed of SQLite.
 108 </p>
 109
 110 <p>
 111 All tests are conducted on an otherwise quiescent machine.
 112 A simple Tcl script was used to generate and run all the tests.
 113 A copy of this Tcl script can be found in the SQLite source tree
 114 in the file <b>tools/speedtest.tcl</b>.
 115 </p>
 116
 117 <p>
 118 The times reported on all tests represent wall-clock time
 119 in seconds.  Two separate time values are reported for SQLite.
 120 The first value is for SQLite in its default configuration with
 121 full disk synchronization turned on.  With synchronization turned
 122 on, SQLite executes
 123 an <b>fsync()</b> system call (or the equivalent) at key points
 124 to make certain that critical data has
 125 actually been written to the disk drive surface.  Synchronization
 126 is necessary to guarantee the integrity of the database if the
 127 operating system crashes or the computer powers down unexpectedly
 128 in the middle of a database update.  The second time reported for SQLite is
 129 when synchronization is turned off.  With synchronization off,
 130 SQLite is sometimes much faster, but there is a risk that an
 131 operating system crash or an unexpected power failure could
 132 damage the database.  Generally speaking, the synchronous SQLite
 133 times are for comparison against PostgreSQL (which is also
 134 synchronous) and the asynchronous SQLite times are for
 135 comparison against the asynchronous MySQL engine.
 136 </p>
 137
 138 <h3>Test 1: 1000 INSERTs</h3>
 139 <blockquote>
 140 CREATE TABLE t1(a INTEGER, b INTEGER, c VARCHAR(100));<br>
 141 INSERT INTO t1 VALUES(1,13153,'thirteen thousand one hundred fifty three');<br>
 142 INSERT INTO t1 VALUES(2,75560,'seventy five thousand five hundred sixty');<br>
 143 <i>... 995 lines omitted</i><br>
 144 INSERT INTO t1 VALUES(998,66289,'sixty six thousand two hundred eighty nine');<br>
 145 INSERT INTO t1 VALUES(999,24322,'twenty four thousand three hundred twenty two');<br>
 146 INSERT INTO t1 VALUES(1000,94142,'ninety four thousand one hundred forty two');<br>
 147
 148 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 149 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;4.373</td></tr>
 150 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.114</td></tr>
 151 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;13.061</td></tr>
 152 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.223</td></tr>
 153 </table>
 154
 155 <p>
 156 Because it does not have a central server to coordinate access,
 157 SQLite must close and reopen the database file, and thus invalidate
 158 its cache, for each transaction.  In this test, each SQL statement
 159 is a separate transaction so the database file must be opened and closed
 160 and the cache must be flushed 1000 times.  In spite of this, the asynchronous
 161 version of SQLite is still nearly as fast as MySQL.  Notice how much slower
 162 the synchronous version is, however.  SQLite calls <b>fsync()</b> after
 163 each synchronous transaction to make sure that all data is safely on
 164 the disk surface before continuing.  For most of the 13 seconds in the
 165 synchronous test, SQLite was sitting idle waiting on disk I/O to complete.</p>
 166
 167
 168 <h3>Test 2: 25000 INSERTs in a transaction</h3>
 169 <blockquote>
 170 BEGIN;<br>
 171 CREATE TABLE t2(a INTEGER, b INTEGER, c VARCHAR(100));<br>
 172 INSERT INTO t2 VALUES(1,59672,'fifty nine thousand six hundred seventy two');<br>
 173 <i>... 24997 lines omitted</i><br>
 174 INSERT INTO t2 VALUES(24999,89569,'eighty nine thousand five hundred sixty nine');<br>
 175 INSERT INTO t2 VALUES(25000,94666,'ninety four thousand six hundred sixty six');<br>
 176 COMMIT;<br>
 177
 178 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 179 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;4.900</td></tr>
 180 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;2.184</td></tr>
 181 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;0.914</td></tr>
 182 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.757</td></tr>
 183 </table>
 184
 185 <p>
 186 When all the INSERTs are put in a transaction, SQLite no longer has to
 187 close and reopen the database or invalidate its cache between each statement.
 188 It also does not
 189 have to do any fsync()s until the very end.  When unshackled in
 190 this way, SQLite is much faster than either PostgreSQL and MySQL.
 191 </p>
 192
 193 <h3>Test 3: 25000 INSERTs into an indexed table</h3>
 194 <blockquote>
 195 BEGIN;<br>
 196 CREATE TABLE t3(a INTEGER, b INTEGER, c VARCHAR(100));<br>
 197 CREATE INDEX i3 ON t3(c);<br>
 198 <i>... 24998 lines omitted</i><br>
 199 INSERT INTO t3 VALUES(24999,88509,'eighty eight thousand five hundred nine');<br>
 200 INSERT INTO t3 VALUES(25000,84791,'eighty four thousand seven hundred ninety one');<br>
 201 COMMIT;<br>
 202
 203 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 204 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;8.175</td></tr>
 205 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;3.197</td></tr>
 206 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;1.555</td></tr>
 207 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;1.402</td></tr>
 208 </table>
 209
 210 <p>
 211 There were reports that SQLite did not perform as well on an indexed table.
 212 This test was recently added to disprove those rumors.  It is true that
 213 SQLite is not as fast at creating new index entries as the other engines
 214 (see Test 6 below) but its overall speed is still better.
 215 </p>
 216
 217 <h3>Test 4: 100 SELECTs without an index</h3>
 218 <blockquote>
 219 BEGIN;<br>
 220 SELECT count(*), avg(b) FROM t2 WHERE b>=0 AND b<1000;<br>
 221 SELECT count(*), avg(b) FROM t2 WHERE b>=100 AND b<1100;<br>
 222 <i>... 96 lines omitted</i><br>
 223 SELECT count(*), avg(b) FROM t2 WHERE b>=9800 AND b<10800;<br>
 224 SELECT count(*), avg(b) FROM t2 WHERE b>=9900 AND b<10900;<br>
 225 COMMIT;<br>
 226
 227 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 228 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;3.629</td></tr>
 229 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;2.760</td></tr>
 230 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;2.494</td></tr>
 231 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;2.526</td></tr>
 232 </table>
 233
 234
 235 <p>
 236 This test does 100 queries on a 25000 entry table without an index,
 237 thus requiring a full table scan.   Prior versions of SQLite used to
 238 be slower than PostgreSQL and MySQL on this test, but recent performance
 239 enhancements have increased its speed so that it is now the fastest
 240 of the group.
 241 </p>
 242
 243 <h3>Test 5: 100 SELECTs on a string comparison</h3>
 244 <blockquote>
 245 BEGIN;<br>
 246 SELECT count(*), avg(b) FROM t2 WHERE c LIKE '%one%';<br>
 247 SELECT count(*), avg(b) FROM t2 WHERE c LIKE '%two%';<br>
 248 <i>... 96 lines omitted</i><br>
 249 SELECT count(*), avg(b) FROM t2 WHERE c LIKE '%ninety nine%';<br>
 250 SELECT count(*), avg(b) FROM t2 WHERE c LIKE '%one hundred%';<br>
 251 COMMIT;<br>
 252
 253 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 254 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;13.409</td></tr>
 255 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;4.640</td></tr>
 256 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;3.362</td></tr>
 257 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;3.372</td></tr>
 258 </table>
 259
 260 <p>
 261 This test still does 100 full table scans but it uses
 262 uses string comparisons instead of numerical comparisons.
 263 SQLite is over three times faster than PostgreSQL here and about 30%
 264 faster than MySQL.
 265 </p>
 266
 267 <h3>Test 6: Creating an index</h3>
 268 <blockquote>
 269 CREATE INDEX i2a ON t2(a);<br>CREATE INDEX i2b ON t2(b);
 270 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 271 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.381</td></tr>
 272 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.318</td></tr>
 273 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;0.777</td></tr>
 274 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.659</td></tr>
 275 </table>
 276
 277 <p>
 278 SQLite is slower at creating new indices.  This is not a huge problem
 279 (since new indices are not created very often) but it is something that
 280 is being worked on.  Hopefully, future versions of SQLite will do better
 281 here.
 282 </p>
 283
 284 <h3>Test 7: 5000 SELECTs with an index</h3>
 285 <blockquote>
 286 SELECT count(*), avg(b) FROM t2 WHERE b>=0 AND b<100;<br>
 287 SELECT count(*), avg(b) FROM t2 WHERE b>=100 AND b<200;<br>
 288 SELECT count(*), avg(b) FROM t2 WHERE b>=200 AND b<300;<br>
 289 <i>... 4994 lines omitted</i><br>
 290 SELECT count(*), avg(b) FROM t2 WHERE b>=499700 AND b<499800;<br>
 291 SELECT count(*), avg(b) FROM t2 WHERE b>=499800 AND b<499900;<br>
 292 SELECT count(*), avg(b) FROM t2 WHERE b>=499900 AND b<500000;<br>
 293
 294 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 295 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;4.614</td></tr>
 296 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.270</td></tr>
 297 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;1.121</td></tr>
 298 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;1.162</td></tr>
 299 </table>
 300
 301 <p>
 302 All three database engines run faster when they have indices to work with.
 303 But SQLite is still the fastest.
 304 </p>
 305
 306 <h3>Test 8: 1000 UPDATEs without an index</h3>
 307 <blockquote>
 308 BEGIN;<br>
 309 UPDATE t1 SET b=b*2 WHERE a>=0 AND a<10;<br>
 310 UPDATE t1 SET b=b*2 WHERE a>=10 AND a<20;<br>
 311 <i>... 996 lines omitted</i><br>
 312 UPDATE t1 SET b=b*2 WHERE a>=9980 AND a<9990;<br>
 313 UPDATE t1 SET b=b*2 WHERE a>=9990 AND a<10000;<br>
 314 COMMIT;<br>
 315
 316 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 317 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.739</td></tr>
 318 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;8.410</td></tr>
 319 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;0.637</td></tr>
 320 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.638</td></tr>
 321 </table>
 322
 323 <p>
 324 For this particular UPDATE test, MySQL is consistently
 325 five or ten times
 326 slower than PostgreSQL and SQLite.  I do not know why.  MySQL is
 327 normally a very fast engine.  Perhaps this problem has been addressed
 328 in later versions of MySQL.
 329 </p>
 330
 331 <h3>Test 9: 25000 UPDATEs with an index</h3>
 332 <blockquote>
 333 BEGIN;<br>
 334 UPDATE t2 SET b=468026 WHERE a=1;<br>
 335 UPDATE t2 SET b=121928 WHERE a=2;<br>
 336 <i>... 24996 lines omitted</i><br>
 337 UPDATE t2 SET b=35065 WHERE a=24999;<br>
 338 UPDATE t2 SET b=347393 WHERE a=25000;<br>
 339 COMMIT;<br>
 340
 341 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 342 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;18.797</td></tr>
 343 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;8.134</td></tr>
 344 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;3.520</td></tr>
 345 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;3.104</td></tr>
 346 </table>
 347
 348 <p>
 349 As recently as version 2.7.0, SQLite ran at about the same speed as
 350 MySQL on this test.  But recent optimizations to SQLite have more
 351 than doubled speed of UPDATEs.
 352 </p>
 353
 354 <h3>Test 10: 25000 text UPDATEs with an index</h3>
 355 <blockquote>
 356 BEGIN;<br>
 357 UPDATE t2 SET c='one hundred forty eight thousand three hundred eighty two' WHERE a=1;<br>
 358 UPDATE t2 SET c='three hundred sixty six thousand five hundred two' WHERE a=2;<br>
 359 <i>... 24996 lines omitted</i><br>
 360 UPDATE t2 SET c='three hundred eighty three thousand ninety nine' WHERE a=24999;<br>
 361 UPDATE t2 SET c='two hundred fifty six thousand eight hundred thirty' WHERE a=25000;<br>
 362 COMMIT;<br>
 363
 364 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 365 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;48.133</td></tr>
 366 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;6.982</td></tr>
 367 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;2.408</td></tr>
 368 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;1.725</td></tr>
 369 </table>
 370
 371 <p>
 372 Here again, version 2.7.0 of SQLite used to run at about the same speed
 373 as MySQL.  But now version 2.7.6 is over two times faster than MySQL and
 374 over twenty times faster than PostgreSQL.
 375 </p>
 376
 377 <p>
 378 In fairness to PostgreSQL, it started thrashing on this test.  A
 379 knowledgeable administrator might be able to get PostgreSQL to run a lot
 380 faster here by tweaking and tuning the server a little.
 381 </p>
 382
 383 <h3>Test 11: INSERTs from a SELECT</h3>
 384 <blockquote>
 385 BEGIN;<br>INSERT INTO t1 SELECT b,a,c FROM t2;<br>INSERT INTO t2 SELECT b,a,c FROM t1;<br>COMMIT;
 386 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 387 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;61.364</td></tr>
 388 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.537</td></tr>
 389 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;2.787</td></tr>
 390 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;1.599</td></tr>
 391 </table>
 392
 393 <p>
 394 The asynchronous SQLite is just a shade slower than MySQL on this test.
 395 (MySQL seems to be especially adept at INSERT...SELECT statements.)
 396 The PostgreSQL engine is still thrashing - most of the 61 seconds it used
 397 were spent waiting on disk I/O.
 398 </p>
 399
 400 <h3>Test 12: DELETE without an index</h3>
 401 <blockquote>
 402 DELETE FROM t2 WHERE c LIKE '%fifty%';
 403 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 404 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.509</td></tr>
 405 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.975</td></tr>
 406 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;4.004</td></tr>
 407 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.560</td></tr>
 408 </table>
 409
 410 <p>
 411 The synchronous version of SQLite is the slowest of the group in this test,
 412 but the asynchronous version is the fastest.
 413 The difference is the extra time needed to execute fsync().
 414 </p>
 415
 416 <h3>Test 13: DELETE with an index</h3>
 417 <blockquote>
 418 DELETE FROM t2 WHERE a>10 AND a<20000;
 419 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 420 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.316</td></tr>
 421 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;2.262</td></tr>
 422 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;2.068</td></tr>
 423 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.752</td></tr>
 424 </table>
 425
 426 <p>
 427 This test is significant because it is one of the few where
 428 PostgreSQL is faster than MySQL.  The asynchronous SQLite is,
 429 however, faster then both the other two.
 430 </p>
 431
 432 <h3>Test 14: A big INSERT after a big DELETE</h3>
 433 <blockquote>
 434 INSERT INTO t2 SELECT * FROM t1;
 435 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 436 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;13.168</td></tr>
 437 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.815</td></tr>
 438 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;3.210</td></tr>
 439 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;1.485</td></tr>
 440 </table>
 441
 442 <p>
 443 Some older versions of SQLite (prior to version 2.4.0)
 444 would show decreasing performance after a
 445 sequence of DELETEs followed by new INSERTs.  As this test shows, the
 446 problem has now been resolved.
 447 </p>
 448
 449 <h3>Test 15: A big DELETE followed by many small INSERTs</h3>
 450 <blockquote>
 451 BEGIN;<br>
 452 DELETE FROM t1;<br>
 453 INSERT INTO t1 VALUES(1,10719,'ten thousand seven hundred nineteen');<br>
 454 <i>... 11997 lines omitted</i><br>
 455 INSERT INTO t1 VALUES(11999,72836,'seventy two thousand eight hundred thirty six');<br>
 456 INSERT INTO t1 VALUES(12000,64231,'sixty four thousand two hundred thirty one');<br>
 457 COMMIT;<br>
 458
 459 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 460 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;4.556</td></tr>
 461 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;1.704</td></tr>
 462 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;0.618</td></tr>
 463 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.406</td></tr>
 464 </table>
 465
 466 <p>
 467 SQLite is very good at doing INSERTs within a transaction, which probably
 468 explains why it is so much faster than the other databases at this test.
 469 </p>
 470
 471 <h3>Test 16: DROP TABLE</h3>
 472 <blockquote>
 473 DROP TABLE t1;<br>DROP TABLE t2;<br>DROP TABLE t3;
 474 </blockquote><table border=0 cellpadding=0 cellspacing=0>
 475 <tr><td>PostgreSQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.135</td></tr>
 476 <tr><td>MySQL:</td><td align="right">&nbsp;&nbsp;&nbsp;0.015</td></tr>
 477 <tr><td>SQLite 2.7.6:</td><td align="right">&nbsp;&nbsp;&nbsp;0.939</td></tr>
 478 <tr><td>SQLite 2.7.6 (nosync):</td><td align="right">&nbsp;&nbsp;&nbsp;0.254</td></tr>
 479 </table>
 480
 481 <p>
 482 SQLite is slower than the other databases when it comes to dropping tables.
 483 This probably is because when SQLite drops a table, it has to go through and
 484 erase the records in the database file that deal with that table.  MySQL and
 485 PostgreSQL, on the other hand, use separate files to represent each table
 486 so they can drop a table simply by deleting a file, which is much faster.
 487 </p>
 488
 489 <p>
 490 On the other hand, dropping tables is not a very common operation
 491 so if SQLite takes a little longer, that is not seen as a big problem.
 492 </p>
 493
 494 }
 495 footer $rcsid