sqlite/test/sort4.test

# 2014 May 6.
#
# 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.
#
#***********************************************************************
# This file implements regression tests for SQLite library. 
#
# The tests in this file are brute force tests of the multi-threaded
# sorter.
#

set testdir [file dirname $argv0]
source $testdir/tester.tcl
set testprefix sort4

# Configure the sorter to use 3 background threads.
db eval {PRAGMA threads=3}

# Minimum number of seconds to run for. If the value is 0, each test
# is run exactly once. Otherwise, tests are repeated until the timeout
# expires.
set SORT4TIMEOUT 0
if {[permutation] == "multithread"} { set SORT4TIMEOUT 300 }

#--------------------------------------------------------------------
# Set up a table "t1" containing $nRow rows. Each row contains also
# contains blob fields that collectively contain at least $nPayload 
# bytes of content. The table schema is as follows:
#
#   CREATE TABLE t1(a INTEGER, <extra-columns>, b INTEGER);
#
# For each row, the values of columns "a" and "b" are set to the same
# pseudo-randomly selected integer. The "extra-columns", of which there
# are at most eight, are named c0, c1, c2 etc. Column c0 contains a 4
# byte string. Column c1 an 8 byte string. Field c2 16 bytes, and so on.
#
# This table is intended to be used for testing queries of the form: 
#
#   SELECT a, <cols>, b FROM t1 ORDER BY a;
#
# The test code checks that rows are returned in order, and that the 
# values of "a" and "b" are the same for each row (the idea being that
# if field "b" at the end of the sorter record has not been corrupted, 
# the rest of the record is probably Ok as well).
#
proc populate_table {nRow nPayload} {
  set nCol 0

  set n 0
  for {set nCol 0} {$n < $nPayload} {incr nCol} {
    incr n [expr (4 << $nCol)]
  }

  set cols [lrange [list xxx c0 c1 c2 c3 c4 c5 c6 c7] 1 $nCol]
  set data [lrange [list xxx \
      randomblob(4) randomblob(8) randomblob(16) randomblob(32) \
      randomblob(64) randomblob(128) randomblob(256) randomblob(512) \
  ] 1 $nCol]

  execsql { DROP TABLE IF EXISTS t1 }

  db transaction {
    execsql "CREATE TABLE t1(a, [join $cols ,], b);"
    set insert "INSERT INTO t1 VALUES(:k, [join $data ,], :k)"
    for {set i 0} {$i < $nRow} {incr i} {
      set k [expr int(rand()*1000000000)]
      execsql $insert
    }
  }
}

# Helper for [do_sorter_test]
#
proc sorter_test {nRow nRead nPayload} {
  set res [list]

  set nLoad [expr ($nRow > $nRead) ? $nRead : $nRow]

  set nPayload [expr (($nPayload+3)/4) * 4]
  set cols [list]
  foreach {mask col} { 
    0x04  c0 0x08  c1 0x10  c2 0x20  c3 
    0x40  c4 0x80  c5 0x100 c6 0x200 c7 
  } {
    if {$nPayload & $mask} { lappend cols $col }
  }

  # Create two SELECT statements. Statement $sql1 uses the sorter to sort
  # $nRow records of a bit over $nPayload bytes each read from the "t1"
  # table created by [populate_table] proc above. Rows are sorted in order
  # of the integer field in each "t1" record.
  #
  # The second SQL statement sorts the same set of rows as the first, but
  # uses a LIMIT clause, causing SQLite to use a temp table instead of the
  # sorter for sorting.
  #
  set sql1 "SELECT a, [join $cols ,], b FROM t1 WHERE rowid<=$nRow ORDER BY a"
  set sql2 "SELECT a FROM t1 WHERE rowid<=$nRow ORDER BY a LIMIT $nRead"

  # Pass the two SQL statements to a helper command written in C. This
  # command steps statement $sql1 $nRead times and compares the integer
  # values in the rows returned with the results of executing $sql2. If
  # the comparison fails (indicating some bug in the sorter), a Tcl
  # exception is thrown.
  #
  sorter_test_sort4_helper db $sql1 $nRead $sql2
  set {} {} 
}

# Usage:
#
#   do_sorter_test <testname> <args>...
#
# where <args> are any of the following switches:
#
#   -rows N          (number of rows to have sorter sort)
#   -read N          (number of rows to read out of sorter)
#   -payload N       (bytes of payload to read with each row)
#   -cachesize N     (Value for "PRAGMA cache_size = ?")
#   -repeats N       (number of times to repeat test)
#   -fakeheap BOOL   (true to use separate allocations for in-memory records)
#
proc do_sorter_test {tn args} {
  set a(-rows)      1000
  set a(-repeats)   1
  set a(-read)      100
  set a(-payload)   100
  set a(-cachesize) 100
  set a(-fakeheap)  0

  foreach {s val} $args {
    if {[info exists a($s)]==0} { 
      unset a(-cachesize)
      set optlist "[join [array names a] ,] or -cachesize"
      error "Unknown option $s, expected $optlist"
    }
    set a($s) $val
  }
  if {[permutation] == "memsys3" || [permutation] == "memsys5"} {
    set a(-fakeheap) 0
  }
  if {$a(-fakeheap)} { sorter_test_fakeheap 1 }


  db eval "PRAGMA cache_size = $a(-cachesize)"
  do_test $tn [subst -nocommands {
    for {set i 0} {[set i] < $a(-repeats)} {incr i} {
      sorter_test $a(-rows) $a(-read) $a(-payload)
    }
  }] {}

  if {$a(-fakeheap)} { sorter_test_fakeheap 0 }
}

proc clock_seconds {} {
  db one {SELECT strftime('%s')}
}

#-------------------------------------------------------------------------
# Begin tests here.

# Create a test database.
do_test 1 {
  execsql "PRAGMA page_size = 4096"
  populate_table 100000 500
} {}

set iTimeLimit [expr [clock_seconds] + $SORT4TIMEOUT]

for {set t 2} {1} {incr tn} {
  do_sorter_test $t.2 -repeats 10 -rows 1000   -read 100
  do_sorter_test $t.3 -repeats 10 -rows 100000 -read 1000
  do_sorter_test $t.4 -repeats 10 -rows 100000 -read 1000 -payload 500
  do_sorter_test $t.5 -repeats 10 -rows 100000 -read 100000 -payload 8
  do_sorter_test $t.6 -repeats 10 -rows 100000 -read 10 -payload 8
  do_sorter_test $t.7 -repeats 10 -rows 10000 -read 10000 -payload 8 -fakeheap 1
  do_sorter_test $t.8 -repeats 10 -rows 100000 -read 10000 -cachesize 250

  set iNow [clock_seconds]
  if {$iNow>=$iTimeLimit} break
  do_test "$testprefix-([expr $iTimeLimit-$iNow] seconds remain)" {} {}
}

finish_test