sqlite/ext/fts5/test/fts5detail.test

# 2015 December 18
#
# 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
# focus of this script is testing the FTS5 module.
#

source [file join [file dirname [info script]] fts5_common.tcl]
set testprefix fts5detail

# If SQLITE_ENABLE_FTS5 is not defined, omit this file.
ifcapable !fts5 {
  finish_test
  return
}

fts5_aux_test_functions db

#--------------------------------------------------------------------------
# Simple tests.
#
do_execsql_test 1.0 {
  CREATE VIRTUAL TABLE t1 USING fts5(a, b, c, detail=col);
  INSERT INTO t1 VALUES('h d g', 'j b b g b', 'i e i d h g g'); -- 1
  INSERT INTO t1 VALUES('h j d', 'j h d a h', 'f d d g g f b'); -- 2
  INSERT INTO t1 VALUES('j c i', 'f f h e f', 'c j i j c h f'); -- 3
  INSERT INTO t1 VALUES('e g g', 'g e d h i', 'e d b e g d c'); -- 4
  INSERT INTO t1 VALUES('b c c', 'd i h a f', 'd i j f a b c'); -- 5
  INSERT INTO t1 VALUES('e d e', 'b c j g d', 'a i f d h b d'); -- 6
  INSERT INTO t1 VALUES('g h e', 'b c d i d', 'e f c i f i c'); -- 7
  INSERT INTO t1 VALUES('c f j', 'j j i e a', 'h a c f d h e'); -- 8
  INSERT INTO t1 VALUES('a h i', 'c i a f a', 'c f d h g d g'); -- 9
  INSERT INTO t1 VALUES('j g g', 'e f e f f', 'h j b i c g e'); -- 10
}

do_execsql_test 1.1 {
  INSERT INTO t1(t1) VALUES('integrity-check');
}

foreach {tn match res} {
  1 "a:a" {9}
  2 "b:g" {1 4 6}
  3 "c:h" {1 3 6 8 9 10}
} {
  do_execsql_test 1.2.$tn.1 {
    SELECT rowid FROM t1($match);
  } $res

  do_execsql_test 1.2.$tn.2 {
    SELECT rowid FROM t1($match || '*');
  } $res
}

do_catchsql_test 1.3.1 {
  SELECT rowid FROM t1('h + d');
} {1 {fts5: phrase queries are not supported (detail!=full)}}

do_catchsql_test 1.3.2 {
  SELECT rowid FROM t1('NEAR(h d)');
} {1 {fts5: NEAR queries are not supported (detail!=full)}}


#-------------------------------------------------------------------------
# integrity-check with both detail= and prefix= options.
#
do_execsql_test 2.0 {
  CREATE VIRTUAL TABLE t2 USING fts5(a, detail=col, prefix="1");
  INSERT INTO t2(a) VALUES('aa ab');
}

#db eval {SELECT rowid, fts5_decode(rowid, block) aS r FROM t2_data} {puts $r}

do_execsql_test 2.1 {
  INSERT INTO t2(t2) VALUES('integrity-check');
}

do_execsql_test 2.2 {
  SELECT fts5_test_poslist(t2) FROM t2('aa');
} {0.0.0}

do_execsql_test 2.3 {
  SELECT fts5_test_collist(t2) FROM t2('aa');
} {0.0}

set ::pc 0
#puts [nearset {{ax bx cx}} -pc ::pc -near 10 -- b*]
#exit

#-------------------------------------------------------------------------
# Check that the xInstCount, xInst, xPhraseFirst and xPhraseNext APIs
# work with detail=col tables.
#
set data {
  1  {abb aca aca} {aba bab aab aac caa} {abc cbc ccb bcc bab ccb aca}
  2  {bca aca acb} {ccb bcc bca aab bcc} {bab aaa aac cbb bba aca abc}
  3  {cca abc cab} {aab aba bcc cac baa} {bab cbb acb aba aab ccc cca}
  4  {ccb bcb aba} {aba bbb bcc cac bbb} {cbb aaa bca bcc aab cac aca}
  5  {bca bbc cac} {aba cbb cac cca aca} {cab acb cbc ccb cac bbb bcb}
  6  {acc bba cba} {bab bbc bbb bcb aca} {bca ccc cbb aca bac ccc ccb}
  7  {aba bab aaa} {abb bca aac bcb bcc} {bcb bbc aba aaa cba abc acc}
  8  {cab aba aaa} {ccb aca caa bbc bcc} {aaa abc ccb bbb cac cca abb}
  9  {bcb bab bac} {bcb cba cac bbb abc} {aba aca cbb acb abb ccc ccb}
  10 {aba aab ccc} {abc ccc bcc cab bbb} {aab bcc cbb ccc aaa bac baa}
  11 {bab acb cba} {aac cab cab bca cbc} {aab cbc aac baa ccb acc cac}
  12 {ccc cbb cbc} {aaa aab bcc aac bbc} {cbc cbc bac bac ccc bbc acc}
  13 {cab bbc abc} {bbb bab bba aca bab} {baa bbb aab bbb ccb bbb ccc}
  14 {bbc cab caa} {acb aac abb cba acc} {cba bba bba acb abc abb baa}
  15 {aba cca bcc} {aaa acb abc aab ccb} {cca bcb acc aaa caa cca cbc}
  16 {bcb bba aba} {cbc acb cab caa ccb} {aac aaa bbc cab cca cba abc}
  17 {caa cbb acc} {ccb bcb bca aaa bcc} {bbb aca bcb bca cbc cbc cca}
  18 {cbb bbc aac} {ccc bbc aaa aab baa} {cab cab cac cca bbc abc bbc}
  19 {ccc acc aaa} {aab cbb bca cca caa} {bcb aca aca cab acc bac bcc}
  20 {aab ccc bcb} {bbc cbb bbc aaa bcc} {cbc aab ccc aaa bcb bac cbc}
  21 {aba cab ccc} {bbc cbc cba acc bbb} {acc aab aac acb aca bca acb}
  22 {bcb bca baa} {cca bbc aca ccb cbb} {aab abc bbc aaa cab bcc bcc}
  23 {cac cbb caa} {bbc aba bbb bcc ccb} {bbc bbb cab bbc cac abb acc}
  24 {ccb acb caa} {cab bba cac bbc aac} {aac bca abc cab bca cab bcb}
  25 {bbb aca bca} {bcb acc ccc cac aca} {ccc acb acc cac cac bba bbc}
  26 {bab acc caa} {caa cab cac bac aca} {aba cac caa acc bac ccc aaa}
  27 {bca bca aaa} {ccb aca bca aaa baa} {bab acc aaa cca cba cca bac}
  28 {ccb cac cac} {bca abb bba bbc baa} {aca ccb aac cab ccc cab caa}
  29 {abc bca cab} {cac cbc cbb ccc bcc} {bcc aaa aaa acc aac cac aac}
  30 {aca acc acb} {aab aac cbb caa acb} {acb bbc bbc acc cbb bbc aac}
  31 {aba aca baa} {aca bcc cab bab acb} {bcc acb baa bcb bbc acc aba}
  32 {abb cbc caa} {cba abb bbb cbb aca} {bac aca caa cac caa ccb bbc}
  33 {bcc bcb bcb} {cca cab cbc abb bab} {caa bbc aac bbb cab cba aaa}
  34 {caa cab acc} {ccc ccc bcc acb bcc} {bac bba aca bcb bba bcb cac}
  35 {bac bcb cba} {bcc acb bbc cba bab} {abb cbb abc abc bac acc cbb}
  36 {cab bab ccb} {bca bba bab cca acc} {acc aab bcc bac acb cbb caa}
  37 {aca cbc cab} {bba aac aca aac aaa} {baa cbb cba aba cab bca bcb}
  38 {acb aab baa} {baa bab bca bbc bbb} {abc baa acc aba cab baa cac}
  39 {bcb aac cba} {bcb baa caa cac bbc} {cbc ccc bab ccb bbb caa aba}
  40 {cba ccb abc} {cbb caa cba aac bab} {cbb bbb bca bbb bac cac bca}
}

set data {
  1  {abb aca aca} {aba bab aab aac caa} {abc cbc ccb bcc bab ccb aca}
}

proc matchdata {expr {bAsc 1}} {

  set tclexpr [db one {
    SELECT fts5_expr_tcl($expr, 'nearset $cols -pc ::pc', 'x', 'y', 'z')
  }]
  set res [list]

  #puts "$expr -> $tclexpr"
  foreach {id x y z} $::data {
    set cols [list $x $y $z]
    set ::pc 0
    #set hits [lsort -command instcompare [eval $tclexpr]]
    set hits [eval $tclexpr]
    if {[llength $hits]>0} {
      lappend res [list $id $hits]
    }
  }

  if {$bAsc} {
    set res [lsort -integer -increasing -index 0 $res]
  } else {
    set res [lsort -integer -decreasing -index 0 $res]
  }

  return [concat {*}$res]
}

foreach {tn tbl} {
  1 { CREATE VIRTUAL TABLE t3 USING fts5(x, y, z, detail=col) }
  2 { CREATE VIRTUAL TABLE t3 USING fts5(x, y, z, detail=none) }
} {
  reset_db
  fts5_aux_test_functions db
  execsql $tbl
  foreach {id x y z} $data {
    execsql { INSERT INTO t3(rowid, x, y, z) VALUES($id, $x, $y, $z) }
  }
  foreach {tn2 expr} {
    1 aaa    2 ccc    3 bab    4 aac
    5 aa*    6 cc*    7 ba*    8 aa*
    9 a*     10 b*   11 c*
  } {

    set res [matchdata $expr]

    do_execsql_test 3.$tn.$tn2.1 {
      SELECT rowid, fts5_test_poslist(t3) FROM t3($expr)
    } $res

    do_execsql_test 3.$tn.$tn2.2 {
      SELECT rowid, fts5_test_poslist2(t3) FROM t3($expr)
    } $res
  }
}

#-------------------------------------------------------------------------
# Simple tests for detail=none tables.
#
do_execsql_test 4.0 {
  CREATE VIRTUAL TABLE t4 USING fts5(a, b, c, detail=none);
  INSERT INTO t4 VALUES('a b c', 'b c d', 'e f g');
  INSERT INTO t4 VALUES('1 2 3', '4 5 6', '7 8 9');
}

do_catchsql_test 4.1 {
  SELECT * FROM t4('a:a')
} {1 {fts5: column queries are not supported (detail=none)}}

do_catchsql_test 4.2 {
  SELECT * FROM t4('a:a &')
} {1 {fts5: syntax error near "&"}}

#-------------------------------------------------------------------------
# Test that for the same content detail=none uses less space than 
# detail=col, and that detail=col uses less space than detail=full
#
reset_db
do_test 5.1 {
  foreach {tbl detail} {t1 none t2 col t3 full} {
    execsql "CREATE VIRTUAL TABLE $tbl USING fts5(x, y, z, detail=$detail)"
    foreach {rowid x y z} $::data {
      execsql "INSERT INTO $tbl (rowid, x, y, z) VALUES(\$rowid, \$x, \$y, \$z)"
    }
  }
} {}

do_execsql_test 5.2 {
  SELECT 
    (SELECT sum(length(block)) from t1_data) <
    (SELECT sum(length(block)) from t2_data)
} {1}

do_execsql_test 5.3 {
  SELECT 
    (SELECT sum(length(block)) from t2_data) <
    (SELECT sum(length(block)) from t3_data)
} {1}



finish_test