# 2011 January 19
#
# 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 tests for SQLite library.  The focus of the tests
# in this file is the use of the sqlite_stat2 histogram data on tables
# with many repeated values and only a few distinct values.
#

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

ifcapable !stat2 {
  finish_test
  return
}

set testprefix analyze5

proc eqp {sql {db db}} {
  uplevel execsql [list "EXPLAIN QUERY PLAN $sql"] $db
}

do_test analyze5-1.0 {
  execsql { CREATE TABLE t1(x INTEGER PRIMARY KEY, y, z) }
  for {set i 0} {$i < 1000} {incr i} {
    set j [expr {$i>=25 && $i<=50}]
    set k [expr {($i>=400) + ($i>=700) + ($i>=875)}]
    execsql { INSERT INTO t1 VALUES($i,$j,$k) }
  }
  execsql { 
    CREATE INDEX t1y ON t1(y);
    CREATE INDEX t1z ON t1(z);
    ANALYZE;
    SELECT * FROM sqlite_stat2 ORDER BY 1, 2, 3;
  }
} [list t1 t1y 0 0 \
        t1 t1y 1 0 \
        t1 t1y 2 0 \
        t1 t1y 3 0 \
        t1 t1y 4 0 \
        t1 t1y 5 0 \
        t1 t1y 6 0 \
        t1 t1y 7 0 \
        t1 t1y 8 0 \
        t1 t1y 9 0 \
        t1 t1z 0 0 \
        t1 t1z 1 0 \
        t1 t1z 2 0 \
        t1 t1z 3 0 \
        t1 t1z 4 1 \
        t1 t1z 5 1 \
        t1 t1z 6 1 \
        t1 t1z 7 2 \
        t1 t1z 8 2 \
        t1 t1z 9 3]

# Verify that range queries generate the correct row count estimates
#
foreach {testid where rows} {
  1  {z>=0 AND z<=0}     400
  2  {z>=1 AND z<=1}     300
  3  {z>=2 AND z<=2}     200
  4  {z>=3 AND z<=3}     100
  5  {z>=4 AND z<=4}      50
  6  {z>=-1 AND z<=-1}    50
  7  {z>1 AND z<3}       200
  8  {z>0 AND z<100}     600
  9  {z>=1 AND z<100}    600
 10  {z>1 AND z<100}     300
 11  {z>=2 AND z<100}    300
 12  {z>2 AND z<100}     100
 13  {z>=3 AND z<100}    100
 14  {z>3 AND z<100}      50
 15  {z>=4 AND z<100}     50
 16  {z>=-100 AND z<=-1}  50
 17  {z>=-100 AND z<=0}  400
 18  {z>=-100 AND z<0}    50
 19  {z>=-100 AND z<=1}  700
 20  {z>=-100 AND z<2}   700
 21  {z>=-100 AND z<=2}  900
 22  {z>=-100 AND z<3}   900

 31  {z>=0.0 AND z<=0.0}   400
 32  {z>=1.0 AND z<=1.0}   300
 33  {z>=2.0 AND z<=2.0}   200
 34  {z>=3.0 AND z<=3.0}   100
 35  {z>=4.0 AND z<=4.0}    50
 36  {z>=-1.0 AND z<=-1.0}  50
 37  {z>1.5 AND z<3.0}     200
 38  {z>0.5 AND z<100}     600
 39  {z>=1.0 AND z<100}    600
 40  {z>1.5 AND z<100}     300
 41  {z>=2.0 AND z<100}    300
 42  {z>2.1 AND z<100}     100
 43  {z>=3.0 AND z<100}    100
 44  {z>3.2 AND z<100}      50
 45  {z>=4.0 AND z<100}     50
 46  {z>=-100 AND z<=-1.0}  50
 47  {z>=-100 AND z<=0.0}  400
 48  {z>=-100 AND z<0.0}    50
 49  {z>=-100 AND z<=1.0}  700
 50  {z>=-100 AND z<2.0}   700
 51  {z>=-100 AND z<=2.0}  900
 52  {z>=-100 AND z<3.0}   900

} {
  do_test analyze5-1.$testid {
    eqp "SELECT * FROM t1 WHERE $where"
  } [format {0 0 0 {SEARCH TABLE t1 USING INDEX t1z (z>? AND z<?) (~%d rows)}} \
       $rows]
}

# Change the table values from integer to floating point and then
# repeat the same sequence of tests.  We should get the same results.
#
do_test analyze5-2.0 {
  db eval {
    UPDATE t1 SET z=z+0.0;
    ANALYZE;
    SELECT sample FROM sqlite_stat2 WHERE idx='t1z' ORDER BY sampleno;
  }
} {0.0 0.0 0.0 0.0 1.0 1.0 1.0 2.0 2.0 3.0}
foreach {testid where rows} {
  1  {z>=0 AND z<=0}     400
  2  {z>=1 AND z<=1}     300
  3  {z>=2 AND z<=2}     200
  4  {z>=3 AND z<=3}     100
  5  {z>=4 AND z<=4}      50
  6  {z>=-1 AND z<=-1}    50
  7  {z>1 AND z<3}       200
  8  {z>0 AND z<100}     600
  9  {z>=1 AND z<100}    600
 10  {z>1 AND z<100}     300
 11  {z>=2 AND z<100}    300
 12  {z>2 AND z<100}     100
 13  {z>=3 AND z<100}    100
 14  {z>3 AND z<100}      50
 15  {z>=4 AND z<100}     50
 16  {z>=-100 AND z<=-1}  50
 17  {z>=-100 AND z<=0}  400
 18  {z>=-100 AND z<0}    50
 19  {z>=-100 AND z<=1}  700
 20  {z>=-100 AND z<2}   700
 21  {z>=-100 AND z<=2}  900
 22  {z>=-100 AND z<3}   900

 31  {z>=0.0 AND z<=0.0}   400
 32  {z>=1.0 AND z<=1.0}   300
 33  {z>=2.0 AND z<=2.0}   200
 34  {z>=3.0 AND z<=3.0}   100
 35  {z>=4.0 AND z<=4.0}    50
 36  {z>=-1.0 AND z<=-1.0}  50
 37  {z>1.5 AND z<3.0}     200
 38  {z>0.5 AND z<100}     600
 39  {z>=1.0 AND z<100}    600
 40  {z>1.5 AND z<100}     300
 41  {z>=2.0 AND z<100}    300
 42  {z>2.1 AND z<100}     100
 43  {z>=3.0 AND z<100}    100
 44  {z>3.2 AND z<100}      50
 45  {z>=4.0 AND z<100}     50
 46  {z>=-100 AND z<=-1.0}  50
 47  {z>=-100 AND z<=0.0}  400
 48  {z>=-100 AND z<0.0}    50
 49  {z>=-100 AND z<=1.0}  700
 50  {z>=-100 AND z<2.0}   700
 51  {z>=-100 AND z<=2.0}  900
 52  {z>=-100 AND z<3.0}   900
} {
  do_test analyze5-2.$testid {
    eqp "SELECT * FROM t1 WHERE $where"
  } [format {0 0 0 {SEARCH TABLE t1 USING INDEX t1z (z>? AND z<?) (~%d rows)}} \
       $rows]
}

# Repeat the same range query tests using TEXT columns.
#
do_test analyze5-3.0 {
  db eval {
    UPDATE t1 SET y=CASE z WHEN 0 THEN 'alpha' WHEN 1 THEN 'bravo'
                           WHEN 2 THEN 'charlie' ELSE 'delta' END;
    ANALYZE;
    SELECT sample FROM sqlite_stat2 WHERE idx='t1y' ORDER BY sampleno;
  }
} {alpha alpha alpha alpha bravo bravo bravo charlie charlie delta}
foreach {testid where rows} {
  1  {y>='alpha' AND y<='alpha'}     400
  2  {y>='bravo' AND y<='bravo'}     300
  3  {y>='charlie' AND y<='charlie'} 200
  4  {y>='delta' AND y<='delta'}     100
  5  {y>='echo' AND y<='echo'}        50
  6  {y>='' AND y<=''}                50
  7  {y>'bravo' AND y<'delta'}       200
  8  {y>'alpha' AND y<'zzz'}         600
  9  {y>='bravo' AND y<'zzz'}        600
 10  {y>'bravo' AND y<'zzz'}         300
 11  {y>='charlie' AND y<'zzz'}      300
 12  {y>'charlie' AND y<'zzz'}       100
 13  {y>='delta' AND y<'zzz'}        100
 14  {y>'delta' AND y<'zzz'}          50
 15  {y>='echo' AND y<'zzz'}          50
 16  {y>=0 AND y<=''}                 50
 17  {y>=0 AND y<='alpha'}           400
 18  {y>=0 AND y<'alpha'}             50
 19  {y>=0 AND y<='bravo'}           700
 20  {y>=0 AND y<'charlie'}          700
 21  {y>=0 AND y<='charlie'}         900
 22  {y>=0 AND y<'delta'}            900
 23  {y>'alpha' AND y<x'00'}         600
 24  {y>='bravo' AND y<x'00'}        600
 25  {y>'bravo' AND y<x'00'}         300
 26  {y>='charlie' AND y<x'00'}      300
 27  {y>'charlie' AND y<x'00'}       100
 28  {y>='delta' AND y<x'00'}        100
 29  {y>'delta' AND y<x'00'}          50
 30  {y>='echo' AND y<x'00'}          50
} {
  do_test analyze5-3.$testid {
    eqp "SELECT * FROM t1 WHERE $where"
  } [format {0 0 0 {SEARCH TABLE t1 USING INDEX t1y (y>? AND y<?) (~%d rows)}} \
       $rows]
}


finish_test