xref: /sqlite-3.40.0/test/analyze9.test (revision f71b8123)

# 2013 August 3
#
# 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 contains automated tests used to verify that the sqlite_stat4
# functionality is working.
#

set testdir [file dirname $argv0]
source $testdir/tester.tcl
set testprefix analyze9

ifcapable !stat4 {
  finish_test
  return
}

proc s {blob} {
  set ret ""
  binary scan $blob c* bytes
  foreach b $bytes {
    set t [binary format c $b]
    if {[string is print $t]} {
      append ret $t
    } else {
      append ret .
    }
  }
  return $ret
}
db function s s

do_execsql_test 1.0 {
  CREATE TABLE t1(a TEXT, b TEXT);
  INSERT INTO t1 VALUES('(0)', '(0)');
  INSERT INTO t1 VALUES('(1)', '(1)');
  INSERT INTO t1 VALUES('(2)', '(2)');
  INSERT INTO t1 VALUES('(3)', '(3)');
  INSERT INTO t1 VALUES('(4)', '(4)');
  CREATE INDEX i1 ON t1(a, b);
} {}


do_execsql_test 1.1 {
  ANALYZE;
} {}

do_execsql_test 1.2 {
  SELECT tbl,idx,nEq,nLt,nDLt,test_decode(sample) FROM sqlite_stat4;
} {
  t1 i1 {1 1 1} {0 0 0} {0 0 0} {(0) (0) 1}
  t1 i1 {1 1 1} {1 1 1} {1 1 1} {(1) (1) 2}
  t1 i1 {1 1 1} {2 2 2} {2 2 2} {(2) (2) 3}
  t1 i1 {1 1 1} {3 3 3} {3 3 3} {(3) (3) 4}
  t1 i1 {1 1 1} {4 4 4} {4 4 4} {(4) (4) 5}
}

if {[permutation] != "utf16"} {
  do_execsql_test 1.3 {
    SELECT tbl,idx,nEq,nLt,nDLt,s(sample) FROM sqlite_stat4;
  } {
    t1 i1 {1 1 1} {0 0 0} {0 0 0} ....(0)(0)
    t1 i1 {1 1 1} {1 1 1} {1 1 1} ....(1)(1).
    t1 i1 {1 1 1} {2 2 2} {2 2 2} ....(2)(2).
    t1 i1 {1 1 1} {3 3 3} {3 3 3} ....(3)(3).
    t1 i1 {1 1 1} {4 4 4} {4 4 4} ....(4)(4).
  }
}


#-------------------------------------------------------------------------
# This is really just to test SQL user function "test_decode".
#
reset_db
do_execsql_test 2.1 {
  CREATE TABLE t1(a, b, c);
  INSERT INTO t1 VALUES('some text', 14, NULL);
  INSERT INTO t1 VALUES(22.0, NULL, x'656667');
  CREATE INDEX i1 ON t1(a, b, c);
  ANALYZE;
  SELECT test_decode(sample) FROM sqlite_stat4;
} {
  {22.0 NULL x'656667' 2}
  {{some text} 14 NULL 1}
}

#-------------------------------------------------------------------------
#
reset_db
do_execsql_test 3.1 {
  CREATE TABLE t2(a, b);
  CREATE INDEX i2 ON t2(a, b);
  BEGIN;
}

do_test 3.2 {
  for {set i 0} {$i < 1000} {incr i} {
    set a [expr $i / 10]
    set b [expr int(rand() * 15.0)]
    execsql { INSERT INTO t2 VALUES($a, $b) }
  }
  execsql COMMIT
} {}

db func lindex lindex

# Each value of "a" occurs exactly 10 times in the table.
#
do_execsql_test 3.3.1 {
  SELECT count(*) FROM t2 GROUP BY a;
} [lrange [string repeat "10 " 100] 0 99]

# The first element in the "nEq" list of all samples should therefore be 10.
#
do_execsql_test 3.3.2 {
  ANALYZE;
  SELECT lindex(nEq, 0) FROM sqlite_stat4;
} [lrange [string repeat "10 " 100] 0 23]

#-------------------------------------------------------------------------
#
do_execsql_test 3.4 {
  DROP TABLE IF EXISTS t1;
  CREATE TABLE t1(a INTEGER PRIMARY KEY, b, c);
  INSERT INTO t1 VALUES(1, 1, 'one-a');
  INSERT INTO t1 VALUES(11, 1, 'one-b');
  INSERT INTO t1 VALUES(21, 1, 'one-c');
  INSERT INTO t1 VALUES(31, 1, 'one-d');
  INSERT INTO t1 VALUES(41, 1, 'one-e');
  INSERT INTO t1 VALUES(51, 1, 'one-f');
  INSERT INTO t1 VALUES(61, 1, 'one-g');
  INSERT INTO t1 VALUES(71, 1, 'one-h');
  INSERT INTO t1 VALUES(81, 1, 'one-i');
  INSERT INTO t1 VALUES(91, 1, 'one-j');
  INSERT INTO t1 SELECT a+1,2,'two' || substr(c,4) FROM t1;
  INSERT INTO t1 SELECT a+2,3,'three'||substr(c,4) FROM t1 WHERE c GLOB 'one-*';
  INSERT INTO t1 SELECT a+3,4,'four'||substr(c,4) FROM t1 WHERE c GLOB 'one-*';
  INSERT INTO t1 SELECT a+4,5,'five'||substr(c,4) FROM t1 WHERE c GLOB 'one-*';
  INSERT INTO t1 SELECT a+5,6,'six'||substr(c,4) FROM t1 WHERE c GLOB 'one-*';
  CREATE INDEX t1b ON t1(b);
  ANALYZE;
  SELECT c FROM t1 WHERE b=3 AND a BETWEEN 30 AND 60;
} {three-d three-e three-f}


#-------------------------------------------------------------------------
# These tests verify that the sample selection for stat4 appears to be
# working as designed.
#

reset_db
db func lindex lindex
db func lrange lrange

do_execsql_test 4.0 {
  DROP TABLE IF EXISTS t1;
  CREATE TABLE t1(a, b, c);
  CREATE INDEX i1 ON t1(c, b, a);
}


proc insert_filler_rows_n {iStart args} {
  set A(-ncopy) 1
  set A(-nval) 1

  foreach {k v} $args {
    if {[info exists A($k)]==0} { error "no such option: $k" }
    set A($k) $v
  }
  if {[llength $args] % 2} {
    error "option requires an argument: [lindex $args end]"
  }

  for {set i 0} {$i < $A(-nval)} {incr i} {
    set iVal [expr $iStart+$i]
    for {set j 0} {$j < $A(-ncopy)} {incr j} {
      execsql { INSERT INTO t1 VALUES($iVal, $iVal, $iVal) }
    }
  }
}

do_test 4.1 {
  execsql { BEGIN }
  insert_filler_rows_n  0  -ncopy 10 -nval 19
  insert_filler_rows_n 20  -ncopy  1 -nval 100

  execsql {
    INSERT INTO t1(c, b, a) VALUES(200, 1, 'a');
    INSERT INTO t1(c, b, a) VALUES(200, 1, 'b');
    INSERT INTO t1(c, b, a) VALUES(200, 1, 'c');

    INSERT INTO t1(c, b, a) VALUES(200, 2, 'e');
    INSERT INTO t1(c, b, a) VALUES(200, 2, 'f');

    INSERT INTO t1(c, b, a) VALUES(201, 3, 'g');
    INSERT INTO t1(c, b, a) VALUES(201, 4, 'h');

    ANALYZE;
    SELECT count(*) FROM sqlite_stat4;
    SELECT count(*) FROM t1;
  }
} {24 297}

do_execsql_test 4.2 {
  SELECT
    neq,
    lrange(nlt, 0, 2),
    lrange(ndlt, 0, 2),
    lrange(test_decode(sample), 0, 2)
    FROM sqlite_stat4
  ORDER BY rowid LIMIT 16;
} {
  {10 10 10 1} {0 0 0} {0 0 0} {0 0 0}
  {10 10 10 1} {10 10 10} {1 1 1} {1 1 1}
  {10 10 10 1} {20 20 20} {2 2 2} {2 2 2}
  {10 10 10 1} {30 30 30} {3 3 3} {3 3 3}
  {10 10 10 1} {40 40 40} {4 4 4} {4 4 4}
  {10 10 10 1} {50 50 50} {5 5 5} {5 5 5}
  {10 10 10 1} {60 60 60} {6 6 6} {6 6 6}
  {10 10 10 1} {70 70 70} {7 7 7} {7 7 7}
  {10 10 10 1} {80 80 80} {8 8 8} {8 8 8}
  {10 10 10 1} {90 90 90} {9 9 9} {9 9 9}
  {10 10 10 1} {100 100 100} {10 10 10} {10 10 10}
  {10 10 10 1} {110 110 110} {11 11 11} {11 11 11}
  {10 10 10 1} {120 120 120} {12 12 12} {12 12 12}
  {10 10 10 1} {130 130 130} {13 13 13} {13 13 13}
  {10 10 10 1} {140 140 140} {14 14 14} {14 14 14}
  {10 10 10 1} {150 150 150} {15 15 15} {15 15 15}
}

do_execsql_test 4.3 {
  SELECT
    neq,
    lrange(nlt, 0, 2),
    lrange(ndlt, 0, 2),
    lrange(test_decode(sample), 0, 1)
    FROM sqlite_stat4
  ORDER BY rowid DESC LIMIT 2;
} {
  {2 1 1 1} {295 296 296} {120 122 125} {201 4}
  {5 3 1 1} {290 290 290} {119 119 119} {200 1}
}

do_execsql_test 4.4 { SELECT count(DISTINCT c) FROM t1 WHERE c<201 } 120
do_execsql_test 4.5 { SELECT count(DISTINCT c) FROM t1 WHERE c<200 } 119

# Check that the perioidic samples are present.
do_execsql_test 4.6 {
  SELECT count(*) FROM sqlite_stat4
  WHERE lindex(test_decode(sample), 3) IN
    ('34', '68', '102', '136', '170', '204', '238', '272')
} {8}

reset_db
do_test 4.7 {
  execsql {
    BEGIN;
    CREATE TABLE t1(o,t INTEGER PRIMARY KEY);
    CREATE INDEX i1 ON t1(o);
  }
  for {set i 0} {$i<10000} {incr i [expr (($i<1000)?1:10)]} {
    execsql { INSERT INTO t1 VALUES('x', $i) }
  }
  execsql {
    COMMIT;
    ANALYZE;
    SELECT count(*) FROM sqlite_stat4;
  }
} {8}
do_execsql_test 4.8 {
  SELECT test_decode(sample) FROM sqlite_stat4;
} {
  {x 211} {x 423} {x 635} {x 847}
  {x 1590} {x 3710} {x 5830} {x 7950}
}


#-------------------------------------------------------------------------
# The following would cause a crash at one point.
#
reset_db
do_execsql_test 5.1 {
  PRAGMA encoding = 'utf-16';
  CREATE TABLE t0(v);
  ANALYZE;
}

#-------------------------------------------------------------------------
# This was also crashing (corrupt sqlite_stat4 table).
#
reset_db
do_execsql_test 6.1 {
  CREATE TABLE t1(a, b);
  CREATE INDEX i1 ON t1(a);
  CREATE INDEX i2 ON t1(b);
  INSERT INTO t1 VALUES(1, 1);
  INSERT INTO t1 VALUES(2, 2);
  INSERT INTO t1 VALUES(3, 3);
  INSERT INTO t1 VALUES(4, 4);
  INSERT INTO t1 VALUES(5, 5);
  ANALYZE;
  PRAGMA writable_schema = 1;
  CREATE TEMP TABLE x1 AS
    SELECT tbl,idx,neq,nlt,ndlt,sample FROM sqlite_stat4
    ORDER BY (rowid%5), rowid;
  DELETE FROM sqlite_stat4;
  INSERT INTO sqlite_stat4 SELECT * FROM x1;
  PRAGMA writable_schema = 0;
  ANALYZE sqlite_master;
}
do_execsql_test 6.2 {
  SELECT * FROM t1 WHERE a = 'abc';
}

#-------------------------------------------------------------------------
# The following tests experiment with adding corrupted records to the
# 'sample' column of the sqlite_stat4 table.
#
reset_db
sqlite3_db_config_lookaside db 0 0 0

database_may_be_corrupt
do_execsql_test 7.1 {
  CREATE TABLE t1(a, b);
  CREATE INDEX i1 ON t1(a, b);
  INSERT INTO t1 VALUES(1, 1);
  INSERT INTO t1 VALUES(2, 2);
  INSERT INTO t1 VALUES(3, 3);
  INSERT INTO t1 VALUES(4, 4);
  INSERT INTO t1 VALUES(5, 5);
  ANALYZE;
  UPDATE sqlite_stat4 SET sample = X'' WHERE rowid = 1;
  ANALYZE sqlite_master;
}

do_execsql_test 7.2 {
  UPDATE sqlite_stat4 SET sample = X'FFFF';
  ANALYZE sqlite_master;
  SELECT * FROM t1 WHERE a = 1;
} {1 1}

do_execsql_test 7.3 {
  ANALYZE;
  UPDATE sqlite_stat4 SET neq = '0 0 0';
  ANALYZE sqlite_master;
  SELECT * FROM t1 WHERE a = 1;
} {1 1}

do_execsql_test 7.4 {
  ANALYZE;
  UPDATE sqlite_stat4 SET ndlt = '0 0 0';
  ANALYZE sqlite_master;
  SELECT * FROM t1 WHERE a = 3;
} {3 3}

do_execsql_test 7.5 {
  ANALYZE;
  UPDATE sqlite_stat4 SET nlt = '0 0 0';
  ANALYZE sqlite_master;
  SELECT * FROM t1 WHERE a = 5;
} {5 5}

database_never_corrupt

#-------------------------------------------------------------------------
#
reset_db
do_execsql_test 8.1 {
  CREATE TABLE t1(x TEXT);
  CREATE INDEX i1 ON t1(x);
  INSERT INTO t1 VALUES('1');
  INSERT INTO t1 VALUES('2');
  INSERT INTO t1 VALUES('3');
  INSERT INTO t1 VALUES('4');
  ANALYZE;
}
do_execsql_test 8.2 {
  SELECT * FROM t1 WHERE x = 3;
} {3}

#-------------------------------------------------------------------------
# Check that the bug fixed by [91733bc485] really is fixed.
#
reset_db
do_execsql_test 9.1 {
  CREATE TABLE t1(a, b, c, d, e);
  CREATE INDEX i1 ON t1(a, b, c, d);
  CREATE INDEX i2 ON t1(e);
}
do_test 9.2 {
  execsql BEGIN;
  for {set i 0} {$i < 100} {incr i} {
    execsql "INSERT INTO t1 VALUES('x', 'y', 'z', $i, [expr $i/2])"
  }
  for {set i 0} {$i < 21} {incr i} {
    execsql "INSERT INTO t1 VALUES('x', 'y', 'z', 101, $i)"
  }
  for {set i 102} {$i < 200} {incr i} {
    execsql "INSERT INTO t1 VALUES('x', 'y', 'z', $i, [expr $i/2])"
  }
  execsql COMMIT
  execsql ANALYZE
} {}

do_eqp_test 9.3.1 {
  SELECT * FROM t1 WHERE a='x' AND b='y' AND c='z' AND d=101 AND e=5;
} {/t1 USING INDEX i2/}
do_eqp_test 9.3.2 {
  SELECT * FROM t1 WHERE a='x' AND b='y' AND c='z' AND d=99 AND e=5;
} {/t1 USING INDEX i1/}

set value_d [expr 101]
do_eqp_test 9.4.1 {
  SELECT * FROM t1 WHERE a='x' AND b='y' AND c='z' AND d=$value_d AND e=5
} {/t1 USING INDEX i2/}
set value_d [expr 99]
do_eqp_test 9.4.2 {
  SELECT * FROM t1 WHERE a='x' AND b='y' AND c='z' AND d=$value_d AND e=5
} {/t1 USING INDEX i1/}

#-------------------------------------------------------------------------
# Check that the planner takes stat4 data into account when considering
# "IS NULL" and "IS NOT NULL" constraints.
#
do_execsql_test 10.1.1 {
  DROP TABLE IF EXISTS t3;
  CREATE TABLE t3(a, b);
  CREATE INDEX t3a ON t3(a);
  CREATE INDEX t3b ON t3(b);
}
do_test 10.1.2 {
  for {set i 1} {$i < 100} {incr i} {
    if {$i>90} { set a $i } else { set a NULL }
    set b [expr $i % 5]
    execsql "INSERT INTO t3 VALUES($a, $b)"
  }
  execsql ANALYZE
} {}
do_eqp_test 10.1.3 {
  SELECT * FROM t3 WHERE a IS NULL AND b = 2
} {/t3 USING INDEX t3b/}
do_eqp_test 10.1.4 {
  SELECT * FROM t3 WHERE a IS NOT NULL AND b = 2
} {/t3 USING INDEX t3a/}

do_execsql_test 10.2.1 {
  DROP TABLE IF EXISTS t3;
  CREATE TABLE t3(x, a, b);
  CREATE INDEX t3a ON t3(x, a);
  CREATE INDEX t3b ON t3(x, b);
}
do_test 10.2.2 {
  for {set i 1} {$i < 100} {incr i} {
    if {$i>90} { set a $i } else { set a NULL }
    set b [expr $i % 5]
    execsql "INSERT INTO t3 VALUES('xyz', $a, $b)"
  }
  execsql ANALYZE
} {}
do_eqp_test 10.2.3 {
  SELECT * FROM t3 WHERE x = 'xyz' AND a IS NULL AND b = 2
} {/t3 USING INDEX t3b/}
do_eqp_test 10.2.4 {
  SELECT * FROM t3 WHERE x = 'xyz' AND a IS NOT NULL AND b = 2
} {/t3 USING INDEX t3a/}

#-------------------------------------------------------------------------
# Check that stat4 data is used correctly with non-default collation
# sequences.
#
foreach {tn schema} {
  1 {
    CREATE TABLE t4(a COLLATE nocase, b);
    CREATE INDEX t4a ON t4(a);
    CREATE INDEX t4b ON t4(b);
  }
  2 {
    CREATE TABLE t4(a, b);
    CREATE INDEX t4a ON t4(a COLLATE nocase);
    CREATE INDEX t4b ON t4(b);
  }
} {
  drop_all_tables
  do_test 11.$tn.1 { execsql $schema } {}

  do_test 11.$tn.2 {
    for {set i 0} {$i < 100} {incr i} {
      if { ($i % 10)==0 } { set a ABC } else { set a DEF }
      set b [expr $i % 5]
        execsql { INSERT INTO t4 VALUES($a, $b) }
    }
    execsql ANALYZE
  } {}

  do_eqp_test 11.$tn.3 {
    SELECT * FROM t4 WHERE a = 'def' AND b = 3;
  } {/t4 USING INDEX t4b/}

  if {$tn==1} {
    set sql "SELECT * FROM t4 WHERE a = 'abc' AND b = 3;"
    do_eqp_test 11.$tn.4 $sql {/t4 USING INDEX t4a/}
  } else {

    set sql "SELECT * FROM t4 WHERE a = 'abc' COLLATE nocase AND b = 3;"
    do_eqp_test 11.$tn.5 $sql {/t4 USING INDEX t4a/}

    set sql "SELECT * FROM t4 WHERE a COLLATE nocase = 'abc' AND b = 3;"
    do_eqp_test 11.$tn.6 $sql {/t4 USING INDEX t4a/}
  }
}

foreach {tn schema} {
  1 {
    CREATE TABLE t4(x, a COLLATE nocase, b);
    CREATE INDEX t4a ON t4(x, a);
    CREATE INDEX t4b ON t4(x, b);
  }
  2 {
    CREATE TABLE t4(x, a, b);
    CREATE INDEX t4a ON t4(x, a COLLATE nocase);
    CREATE INDEX t4b ON t4(x, b);
  }
} {
  drop_all_tables
  do_test 12.$tn.1 { execsql $schema } {}

  do_test 12.$tn.2 {
    for {set i 0} {$i < 100} {incr i} {
      if { ($i % 10)==0 } { set a ABC } else { set a DEF }
      set b [expr $i % 5]
        execsql { INSERT INTO t4 VALUES(X'abcdef', $a, $b) }
    }
    execsql ANALYZE
  } {}

  do_eqp_test 12.$tn.3 {
    SELECT * FROM t4 WHERE x=X'abcdef' AND a = 'def' AND b = 3;
  } {/t4 USING INDEX t4b/}

  if {$tn==1} {
    set sql "SELECT * FROM t4 WHERE x=X'abcdef' AND a = 'abc' AND b = 3;"
    do_eqp_test 12.$tn.4 $sql {/t4 USING INDEX t4a/}
  } else {
    set sql {
      SELECT * FROM t4 WHERE x=X'abcdef' AND a = 'abc' COLLATE nocase AND b = 3
    }
    do_eqp_test 12.$tn.5 $sql {/t4 USING INDEX t4a/}
    set sql {
      SELECT * FROM t4 WHERE x=X'abcdef' AND a COLLATE nocase = 'abc' AND b = 3
    }
    do_eqp_test 12.$tn.6 $sql {/t4 USING INDEX t4a/}
  }
}

#-------------------------------------------------------------------------
# Check that affinities are taken into account when using stat4 data to
# estimate the number of rows scanned by a rowid constraint.
#
drop_all_tables
do_test 13.1 {
  execsql {
    CREATE TABLE t1(a, b, c, d);
    CREATE INDEX i1 ON t1(a);
    CREATE INDEX i2 ON t1(b, c);
  }
  for {set i 0} {$i<100} {incr i} {
    if {$i %2} {set a abc} else {set a def}
    execsql { INSERT INTO t1(rowid, a, b, c) VALUES($i, $a, $i, $i) }
  }
  execsql ANALYZE
} {}
do_eqp_test 13.2.1 {
  SELECT * FROM t1 WHERE a='abc' AND rowid<15 AND b<12
} {/SEARCH t1 USING INDEX i1/}
do_eqp_test 13.2.2 {
  SELECT * FROM t1 WHERE a='abc' AND rowid<'15' AND b<12
} {/SEARCH t1 USING INDEX i1/}
do_eqp_test 13.3.1 {
  SELECT * FROM t1 WHERE a='abc' AND rowid<100 AND b<12
} {/SEARCH t1 USING INDEX i2/}
do_eqp_test 13.3.2 {
  SELECT * FROM t1 WHERE a='abc' AND rowid<'100' AND b<12
} {/SEARCH t1 USING INDEX i2/}

#-------------------------------------------------------------------------
# Check also that affinities are taken into account when using stat4 data
# to estimate the number of rows scanned by any other constraint on a
# column other than the leftmost.
#
drop_all_tables
do_test 14.1 {
  execsql { CREATE TABLE t1(a, b INTEGER, c) }
  for {set i 0} {$i<100} {incr i} {
    set c [expr $i % 3]
    execsql { INSERT INTO t1 VALUES('ott', $i, $c) }
  }
  execsql {
    CREATE INDEX i1 ON t1(a, b);
    CREATE INDEX i2 ON t1(c);
    ANALYZE;
  }
} {}
do_eqp_test 13.2.1 {
  SELECT * FROM t1 WHERE a='ott' AND b<10 AND c=1
} {/SEARCH t1 USING INDEX i1/}
do_eqp_test 13.2.2 {
  SELECT * FROM t1 WHERE a='ott' AND b<'10' AND c=1
} {/SEARCH t1 USING INDEX i1/}

#-------------------------------------------------------------------------
# By default, 16 non-periodic samples are collected for the stat4 table.
# The following tests attempt to verify that the most common keys are
# being collected.
#
proc check_stat4 {tn} {
  db eval ANALYZE
  db eval {SELECT a, b, c, d FROM t1} {
    incr k($a)
    incr k([list $a $b])
    incr k([list $a $b $c])
    if { [info exists k([list $a $b $c $d])]==0 } { incr nRow }
    incr k([list $a $b $c $d])
  }

  set L [list]
  foreach key [array names k] {
    lappend L [list $k($key) $key]
  }

  set nSample $nRow
  if {$nSample>16} {set nSample 16}

  set nThreshold [lindex [lsort -decr -integer -index 0 $L] [expr $nSample-1] 0]
  foreach key [array names k] {
    if {$k($key)>$nThreshold} {
      set expect($key) 1
    }
    if {$k($key)==$nThreshold} {
      set possible($key) 1
    }
  }


  set nPossible [expr $nSample - [llength [array names expect]]]

  #puts "EXPECT: [array names expect]"
  #puts "POSSIBLE($nPossible/[array size possible]): [array names possible]"
  #puts "HAVE: [db eval {SELECT test_decode(sample) FROM sqlite_stat4 WHERE idx='i1'}]"

  db eval {SELECT test_decode(sample) AS s FROM sqlite_stat4 WHERE idx='i1'} {
    set seen 0
    for {set i 0} {$i<4} {incr i} {
      unset -nocomplain expect([lrange $s 0 $i])
      if {[info exists possible([lrange $s 0 $i])]} {
        set seen 1
        unset -nocomplain possible([lrange $s 0 $i])
      }
    }
    if {$seen} {incr nPossible -1}
  }
  if {$nPossible<0} {set nPossible 0}

  set res [list [llength [array names expect]] $nPossible]
  uplevel [list do_test $tn [list set {} $res] {0 0}]
}

drop_all_tables
do_test 14.1.1 {
  execsql {
    CREATE TABLE t1(a,b,c,d);
    CREATE INDEX i1 ON t1(a,b,c,d);
  }
  for {set i 0} {$i < 160} {incr i} {
    execsql { INSERT INTO t1 VALUES($i,$i,$i,$i) }
    if {($i % 10)==0} { execsql { INSERT INTO t1 VALUES($i,$i,$i,$i) } }
  }
} {}
check_stat4 14.1.2

do_test 14.2.1 {
  execsql { DELETE FROM t1 }
  for {set i 0} {$i < 1600} {incr i} {
    execsql { INSERT INTO t1 VALUES($i/10,$i/17,$i/27,$i/37) }
  }
} {}
check_stat4 14.2.2

do_test 14.3.1 {
  for {set i 0} {$i < 10} {incr i} {
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
    execsql { INSERT INTO t1 VALUES($i*50,$i*50,$i*50,$i*50) }
  }
} {}
check_stat4 14.3.2

do_test 14.4.1 {
  execsql {DELETE FROM t1}
  for {set i 1} {$i < 160} {incr i} {
    set b [expr $i % 10]
    if {$b==0 || $b==2} {set b 1}
    execsql { INSERT INTO t1 VALUES($i/10,$b,$i,$i) }
  }
} {}
check_stat4 14.4.2
db func lrange lrange
db func lindex lindex
do_execsql_test 14.4.3 {
  SELECT lrange(test_decode(sample), 0, 1) AS s FROM sqlite_stat4
  WHERE lindex(s, 1)=='1' ORDER BY rowid
} {
  {0 1} {1 1} {2 1} {3 1}
  {4 1} {5 1} {6 1} {7 1}
  {8 1} {9 1} {10 1} {11 1}
  {12 1} {13 1} {14 1} {15 1}
}

#-------------------------------------------------------------------------
# Test that nothing untoward happens if the stat4 table contains entries
# for indexes that do not exist. Or NULL values in the idx column.
# Or NULL values in any of the other columns.
#
drop_all_tables
do_execsql_test 15.1 {
  CREATE TABLE x1(a, b, UNIQUE(a, b));
  INSERT INTO x1 VALUES(1, 2);
  INSERT INTO x1 VALUES(3, 4);
  INSERT INTO x1 VALUES(5, 6);
  ANALYZE;
  INSERT INTO sqlite_stat4 VALUES(NULL, NULL, NULL, NULL, NULL, NULL);
}
db close
sqlite3 db test.db
do_execsql_test 15.2 { SELECT * FROM x1 } {1 2 3 4 5 6}

do_execsql_test 15.3 {
  INSERT INTO sqlite_stat4 VALUES(42, 42, 42, 42, 42, 42);
}
db close
sqlite3 db test.db
do_execsql_test 15.4 { SELECT * FROM x1 } {1 2 3 4 5 6}

do_execsql_test 15.5 {
  UPDATE sqlite_stat1 SET stat = NULL;
}
db close
sqlite3 db test.db
do_execsql_test 15.6 { SELECT * FROM x1 } {1 2 3 4 5 6}

do_execsql_test 15.7 {
  ANALYZE;
  UPDATE sqlite_stat1 SET tbl = 'no such tbl';
}
db close
sqlite3 db test.db
do_execsql_test 15.8 { SELECT * FROM x1 } {1 2 3 4 5 6}

do_execsql_test 15.9 {
  ANALYZE;
  UPDATE sqlite_stat4 SET neq = NULL, nlt=NULL, ndlt=NULL;
}
db close
sqlite3 db test.db
do_execsql_test 15.10 { SELECT * FROM x1 } {1 2 3 4 5 6}

# This is just for coverage....
do_execsql_test 15.11 {
  ANALYZE;
  UPDATE sqlite_stat1 SET stat = stat || ' unordered';
}
db close
sqlite3 db test.db
do_execsql_test 15.12 { SELECT * FROM x1 } {1 2 3 4 5 6}

#-------------------------------------------------------------------------
# Test that allocations used for sqlite_stat4 samples are included in
# the quantity returned by SQLITE_DBSTATUS_SCHEMA_USED.
#
set one [string repeat x 1000]
set two [string repeat x 2000]
do_test 16.1 {
  reset_db
  execsql {
    CREATE TABLE t1(a, UNIQUE(a));
    INSERT INTO t1 VALUES($one);
    ANALYZE;
  }
  set nByte [lindex [sqlite3_db_status db SCHEMA_USED 0] 1]

  reset_db
  execsql {
    CREATE TABLE t1(a, UNIQUE(a));
    INSERT INTO t1 VALUES($two);
    ANALYZE;
  }
  set nByte2 [lindex [sqlite3_db_status db SCHEMA_USED 0] 1]
  puts -nonewline " (nByte=$nByte nByte2=$nByte2)"

  expr {$nByte2 > $nByte+900 && $nByte2 < $nByte+1100}
} {1}

#-------------------------------------------------------------------------
# Test that stat4 data may be used with partial indexes.
#
do_test 17.1 {
  reset_db
  execsql {
    CREATE TABLE t1(a, b, c, d);
    CREATE INDEX i1 ON t1(a, b) WHERE d IS NOT NULL;
    INSERT INTO t1 VALUES(-1, -1, -1, NULL);
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
    INSERT INTO t1 SELECT 2*a,2*b,2*c,d FROM t1;
  }

  for {set i 0} {$i < 32} {incr i} {
    if {$i<8} {set b 0} else { set b $i }
    execsql { INSERT INTO t1 VALUES($i%2, $b, $i/2, 'abc') }
  }
  execsql {ANALYZE main.t1}
} {}

do_catchsql_test 17.1.2 {
  ANALYZE temp.t1;
} {1 {no such table: temp.t1}}

do_eqp_test 17.2 {
  SELECT * FROM t1 WHERE d IS NOT NULL AND a=0 AND b=10 AND c=10;
} {/USING INDEX i1/}
do_eqp_test 17.3 {
  SELECT * FROM t1 WHERE d IS NOT NULL AND a=0 AND b=0 AND c=10;
} {/USING INDEX i1/}

do_execsql_test 17.4 {
  CREATE INDEX i2 ON t1(c, d);
  ANALYZE main.i2;
}
do_eqp_test 17.5 {
  SELECT * FROM t1 WHERE d IS NOT NULL AND a=0 AND b=10 AND c=10;
} {/USING INDEX i1/}
do_eqp_test 17.6 {
  SELECT * FROM t1 WHERE d IS NOT NULL AND a=0 AND b=0 AND c=10;
} {/USING INDEX i2/}

#-------------------------------------------------------------------------
#
do_test 18.1 {
  reset_db
  execsql {
    CREATE TABLE t1(a, b);
    CREATE INDEX i1 ON t1(a, b);
  }
  for {set i 0} {$i < 9} {incr i} {
    execsql {
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
      INSERT INTO t1 VALUES($i, 0);
    }
  }
  execsql ANALYZE
  execsql { SELECT count(*) FROM sqlite_stat4 }
} {9}

#-------------------------------------------------------------------------
# For coverage.
#
ifcapable view {
  do_test 19.1 {
    reset_db
    execsql {
      CREATE TABLE t1(x, y);
      CREATE INDEX i1 ON t1(x, y);
      CREATE VIEW v1 AS SELECT * FROM t1;
      ANALYZE;
    }
  } {}
}
ifcapable auth {
  proc authproc {op args} {
    if {$op == "SQLITE_ANALYZE"} { return "SQLITE_DENY" }
    return "SQLITE_OK"
  }
  do_test 19.2 {
    reset_db
    db auth authproc
    execsql {
      CREATE TABLE t1(x, y);
      CREATE VIEW v1 AS SELECT * FROM t1;
    }
    catchsql ANALYZE
  } {1 {not authorized}}
}

#-------------------------------------------------------------------------
#
reset_db
proc r {args} { expr rand() }
db func r r
db func lrange lrange
do_test 20.1 {
  execsql {
    CREATE TABLE t1(a,b,c,d);
    CREATE INDEX i1 ON t1(a,b,c,d);
  }
  for {set i 0} {$i < 16} {incr i} {
    execsql {
      INSERT INTO t1 VALUES($i, r(), r(), r());
      INSERT INTO t1 VALUES($i, $i,  r(), r());
      INSERT INTO t1 VALUES($i, $i,  $i,  r());
      INSERT INTO t1 VALUES($i, $i,  $i,  $i);
      INSERT INTO t1 VALUES($i, $i,  $i,  $i);
      INSERT INTO t1 VALUES($i, $i,  $i,  r());
      INSERT INTO t1 VALUES($i, $i,  r(), r());
      INSERT INTO t1 VALUES($i, r(), r(), r());
    }
  }
} {}
do_execsql_test 20.2 { ANALYZE }
for {set i 0} {$i<16} {incr i} {
    set val "$i $i $i $i"
    do_execsql_test 20.3.$i {
      SELECT count(*) FROM sqlite_stat4
      WHERE lrange(test_decode(sample), 0, 3)=$val
    } {1}
}

#-------------------------------------------------------------------------
#
reset_db

do_execsql_test 21.0 {
  CREATE TABLE t2(a, b);
  CREATE INDEX i2 ON t2(a);
}

do_test 21.1 {
  for {set i 1} {$i < 100} {incr i} {
    execsql {
      INSERT INTO t2 VALUES(CASE WHEN $i < 80 THEN 'one' ELSE 'two' END, $i)
    }
  }
  execsql ANALYZE
} {}

# Condition (a='one') matches 80% of the table. (rowid<10) reduces this to
# 10%, but (rowid<50) only reduces it to 50%. So in the first case below
# the index is used. In the second, it is not.
#
do_eqp_test 21.2 {
  SELECT * FROM t2 WHERE a='one' AND rowid < 10
} {/*USING INDEX i2 (a=? AND rowid<?)*/}
do_eqp_test 21.3 {
  SELECT * FROM t2 WHERE a='one' AND rowid < 50
} {/*USING INTEGER PRIMARY KEY*/}

#-------------------------------------------------------------------------
#
reset_db
do_execsql_test 22.0 {
  CREATE TABLE t3(a, b, c, d, PRIMARY KEY(a, b)) WITHOUT ROWID;
  SELECT * FROM t3;
} {}
do_execsql_test 22.1 {
  WITH r(x) AS (
    SELECT 1
    UNION ALL
    SELECT x+1 FROM r WHERE x<=100
  )

  INSERT INTO t3 SELECT
    CASE WHEN (x>45 AND x<96) THEN 'B' ELSE 'A' END,  /* Column "a" */
    x,                                                /* Column "b" */
    CASE WHEN (x<51) THEN 'one' ELSE 'two' END,       /* Column "c" */
    x                                                 /* Column "d" */
  FROM r;

  CREATE INDEX i3 ON t3(c);
  CREATE INDEX i4 ON t3(d);
  ANALYZE;
}

# Expression (c='one' AND a='B') matches 5 table rows. But (c='one' AND a=A')
# matches 45. Expression (d<?) matches 20. Neither index is a covering index.
#
# Therefore, with stat4 data, SQLite prefers (c='one' AND a='B') over (d<20),
# and (d<20) over (c='one' AND a='A').
foreach {tn where res} {
  1 "c='one' AND a='B' AND d < 20"   {/*INDEX i3 (c=? AND a=?)*/}
  2 "c='one' AND a='A' AND d < 20"   {/*INDEX i4 (d<?)*/}
} {
  do_eqp_test 22.2.$tn "SELECT * FROM t3 WHERE $where" $res
}

proc int_to_char {i} {
  set ret ""
  set char [list a b c d e f g h i j]
  foreach {div} {1000 100 10 1} {
    append ret [lindex $char [expr ($i / $div) % 10]]
  }
  set ret
}
db func int_to_char int_to_char

do_execsql_test 23.0 {
  CREATE TABLE t4(
    a COLLATE nocase, b, c,
    d, e, f,
    PRIMARY KEY(c, b, a)
  ) WITHOUT ROWID;
  CREATE INDEX i41 ON t4(e);
  CREATE INDEX i42 ON t4(f);

  WITH data(a, b, c, d, e, f) AS (
    SELECT int_to_char(0), 'xyz', 'zyx', '*', 0, 0
    UNION ALL
    SELECT
      int_to_char(f+1), b, c, d, (e+1) % 2, f+1
    FROM data WHERE f<1024
  )
  INSERT INTO t4 SELECT a, b, c, d, e, f FROM data;
  ANALYZE;
} {}

do_eqp_test 23.1 {
  SELECT * FROM t4 WHERE
    (e=1 AND b='xyz' AND c='zyx' AND a<'AEA') AND f<300
  -- Formerly used index i41.  But i41 is not a covering index whereas
  -- the PRIMARY KEY is a covering index, and so as of 2017-10-15, the
  -- PRIMARY KEY is preferred.
} {SEARCH t4 USING PRIMARY KEY (c=? AND b=? AND a<?)}
do_eqp_test 23.2 {
  SELECT * FROM t4 WHERE
    (e=1 AND b='xyz' AND c='zyx' AND a<'JJJ') AND f<300
} {SEARCH t4 USING INDEX i42 (f<?)}

do_execsql_test 24.0 {
  CREATE TABLE t5(c, d, b, e, a, PRIMARY KEY(a, b, c)) WITHOUT ROWID;
  WITH data(a, b, c, d, e) AS (
    SELECT 'z', 'y', 0, 0, 0
    UNION ALL
    SELECT
      a, CASE WHEN b='y' THEN 'n' ELSE 'y' END, c+1, e/250, e+1
    FROM data
    WHERE e<1000
  )
  INSERT INTO t5(a, b, c, d, e) SELECT * FROM data;
  CREATE INDEX t5d ON t5(d);
  CREATE INDEX t5e ON t5(e);
  ANALYZE;
}

foreach {tn where eqp} {
  1 "d=0 AND a='z' AND b='n' AND e<200" {/*t5d (d=? AND a=? AND b=?)*/}
  2 "d=0 AND a='z' AND b='n' AND e<100" {/*t5e (e<?)*/}

  3 "d=0 AND e<300"                     {/*t5d (d=?)*/}
  4 "d=0 AND e<200"                     {/*t5e (e<?)*/}
} {
  do_eqp_test 24.$tn "SeLeCt * FROM t5 WHERE $where" $eqp
}

#-------------------------------------------------------------------------
# Test that if stat4 data is available but cannot be used because the
# rhs of a range constraint is a complex expression, the default estimates
# are used instead.
ifcapable stat4&&cte {
  do_execsql_test 25.1 {
    CREATE TABLE t6(a, b);
    WITH ints(i,j) AS (
      SELECT 1,1 UNION ALL SELECT i+1,j+1 FROM ints WHERE i<100
    ) INSERT INTO t6 SELECT * FROM ints;
    CREATE INDEX aa ON t6(a);
    CREATE INDEX bb ON t6(b);
    ANALYZE;
  }

  # Term (b<?) is estimated at 25%. Better than (a<30) but not as
  # good as (a<20).
  do_eqp_test 25.2.1 { SELECT * FROM t6 WHERE a<30 AND b<? } \
    {SEARCH t6 USING INDEX bb (b<?)}
  do_eqp_test 25.2.2 { SELECT * FROM t6 WHERE a<20 AND b<? } \
    {SEARCH t6 USING INDEX aa (a<?)}

  # Term (b BETWEEN ? AND ?) is estimated at 1/64.
  do_eqp_test 25.3.1 {
    SELECT * FROM t6 WHERE a BETWEEN 5 AND 10 AND b BETWEEN ? AND ?
  } {SEARCH t6 USING INDEX bb (b>? AND b<?)}

  # Term (b BETWEEN ? AND 60) is estimated to return roughly 15 rows -
  # 60 from (b<=60) multiplied by 0.25 for the b>=? term. Better than
  # (a<20) but not as good as (a<10).
  do_eqp_test 25.4.1 {
    SELECT * FROM t6 WHERE a < 10 AND (b BETWEEN ? AND 60)
  } {SEARCH t6 USING INDEX aa (a<?)}

  do_eqp_test 25.4.2 {
    SELECT * FROM t6 WHERE a < 20 AND (b BETWEEN ? AND 60)
  } {SEARCH t6 USING INDEX bb (b>? AND b<?)}
}

#-------------------------------------------------------------------------
# Check that a problem in they way stat4 data is used has been
# resolved (see below).
#
reset_db
do_test 26.1.1 {
  db transaction {
    execsql {
      CREATE TABLE t1(x, y, z);
      CREATE INDEX t1xy ON t1(x, y);
      CREATE INDEX t1z ON t1(z);
    }
    for {set i 0} {$i < 10000} {incr i} {
      execsql { INSERT INTO t1(x, y) VALUES($i, $i) }
    }
    for {set i 0} {$i < 10} {incr i} {
      execsql {
        WITH cnt(x) AS (SELECT 1 UNION ALL SELECT x+1 FROM cnt WHERE x<100)
        INSERT INTO t1(x, y) SELECT 10000+$i, x FROM cnt;
        INSERT INTO t1(x, y) SELECT 10000+$i, 100;
      }
    }
    execsql {
      UPDATE t1 SET z = rowid / 20;
      ANALYZE;
    }
  }
} {}

do_execsql_test 26.1.2 {
  SELECT count(*) FROM t1 WHERE x = 10000 AND y < 50;
} {49}
do_execsql_test 26.1.3 {
  SELECT count(*) FROM t1 WHERE z = 444;
} {20}

# The analyzer knows that any (z=?) expression matches 20 rows. So it
# will use index "t1z" if the estimate of hits for (x=10000 AND y<50)
# is greater than 20 rows.
#
# And it should be. The analyzer has a stat4 sample as follows:
#
#   sample=(x=10000, y=100) nLt=(10000 10099)
#
# There should be no other samples that start with (x=10000). So it knows
# that (x=10000 AND y<50) must match somewhere between 0 and 99 rows, but
# no more than that. Guessing less than 20 is therefore unreasonable.
#
# At one point though, due to a problem in whereKeyStats(), the planner was
# estimating that (x=10000 AND y<50) would match only 2 rows.
#
do_eqp_test 26.1.4 {
  SELECT * FROM t1 WHERE x = 10000 AND y < 50 AND z = 444;
} {SEARCH t1 USING INDEX t1z (z=?)}


# This test - 26.2.* - tests that another manifestation of the same problem
# is no longer present in the library. Assuming:
#
#   CREATE INDEX t1xy ON t1(x, y)
#
# and that have samples for index t1xy as follows:
#
#
#   sample=('A', 70)        nEq=(100, 2)        nLt=(900, 970)
#   sample=('B', 70)        nEq=(100, 2)        nLt=(1000, 1070)
#
# the planner should estimate that (x = 'B' AND y > 25) matches 76 rows
# (70 * 2/3 + 30). Before, due to the problem, the planner was estimating
# that this matched 100 rows.
#
reset_db
do_execsql_test 26.2.1 {
  BEGIN;
    CREATE TABLE t1(x, y, z);
    CREATE INDEX i1 ON t1(x, y);
    CREATE INDEX i2 ON t1(z);

    WITH
    cnt(y) AS (SELECT 0 UNION ALL SELECT y+1 FROM cnt WHERE y<99),
    letters(x) AS (
      SELECT 'A' UNION SELECT 'B' UNION SELECT 'C' UNION SELECT 'D'
    )
    INSERT INTO t1(x, y) SELECT x, y FROM letters, cnt;

    WITH
    letters(x) AS (
      SELECT 'A' UNION SELECT 'B' UNION SELECT 'C' UNION SELECT 'D'
    )
    INSERT INTO t1(x, y) SELECT x, 70 FROM letters;

    WITH
    cnt(i) AS (SELECT 0 UNION ALL SELECT i+1 FROM cnt WHERE i<9999)
    INSERT INTO t1(x, y) SELECT i, i FROM cnt;

    UPDATE t1 SET z = (rowid / 95);
    ANALYZE;
  COMMIT;
}

do_eqp_test 26.2.2 {
  SELECT * FROM t1 WHERE x='B' AND y>25 AND z=?;
} {SEARCH t1 USING INDEX i1 (x=? AND y>?)}


finish_test