diff options
Diffstat (limited to 'test/fts2p.test')
| -rw-r--r-- | test/fts2p.test | 357 |
1 files changed, 357 insertions, 0 deletions
diff --git a/test/fts2p.test b/test/fts2p.test new file mode 100644 index 0000000..38a8079 --- /dev/null +++ b/test/fts2p.test @@ -0,0 +1,357 @@ +# 2008 June 26 +# +# 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 exercises some new testing functions in the FTS2 module, +# and then uses them to do some basic tests that FTS2 is internally +# working as expected. +# +# $Id: fts2p.test,v 1.1 2008/07/22 23:32:28 shess Exp $ +# + +set testdir [file dirname $argv0] +source $testdir/tester.tcl + +# If SQLITE_ENABLE_FTS2 is not defined, omit this file. +ifcapable !fts2 { + finish_test + return +} + +#************************************************************************* +# Probe to see if support for these functions is compiled in. +# TODO(shess): Change main.mk to do the right thing and remove this test. +db eval { + DROP TABLE IF EXISTS t1; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) VALUES (1, 'x'); +} + +set s {SELECT dump_terms(t1, 1) FROM t1 LIMIT 1} +set r {1 {unable to use function dump_terms in the requested context}} +if {[catchsql $s]==$r} { + finish_test + return +} + +#************************************************************************* +# Test that the new functions give appropriate errors. +do_test fts2p-0.0 { + catchsql { + SELECT dump_terms(t1, 1) FROM t1 LIMIT 1; + } +} {1 {dump_terms: incorrect arguments}} + +do_test fts2p-0.1 { + catchsql { + SELECT dump_terms(t1, 0, 0, 0) FROM t1 LIMIT 1; + } +} {1 {dump_terms: incorrect arguments}} + +do_test fts2p-0.2 { + catchsql { + SELECT dump_terms(1, t1) FROM t1 LIMIT 1; + } +} {1 {unable to use function dump_terms in the requested context}} + +do_test fts2p-0.3 { + catchsql { + SELECT dump_terms(t1, 16, 16) FROM t1 LIMIT 1; + } +} {1 {dump_terms: segment not found}} + +do_test fts2p-0.4 { + catchsql { + SELECT dump_doclist(t1) FROM t1 LIMIT 1; + } +} {1 {dump_doclist: incorrect arguments}} + +do_test fts2p-0.5 { + catchsql { + SELECT dump_doclist(t1, NULL) FROM t1 LIMIT 1; + } +} {1 {dump_doclist: empty second argument}} + +do_test fts2p-0.6 { + catchsql { + SELECT dump_doclist(t1, '') FROM t1 LIMIT 1; + } +} {1 {dump_doclist: empty second argument}} + +do_test fts2p-0.7 { + catchsql { + SELECT dump_doclist(t1, 'a', 0) FROM t1 LIMIT 1; + } +} {1 {dump_doclist: incorrect arguments}} + +do_test fts2p-0.8 { + catchsql { + SELECT dump_doclist(t1, 'a', 0, 0, 0) FROM t1 LIMIT 1; + } +} {1 {dump_doclist: incorrect arguments}} + +do_test fts2p-0.9 { + catchsql { + SELECT dump_doclist(t1, 'a', 16, 16) FROM t1 LIMIT 1; + } +} {1 {dump_doclist: segment not found}} + +#************************************************************************* +# Utility function to check for the expected terms in the segment +# level/index. _all version does same but for entire index. +proc check_terms {test level index terms} { + # TODO(shess): Figure out why uplevel in do_test can't catch + # $level and $index directly. + set ::level $level + set ::index $index + do_test $test.terms { + execsql { + SELECT dump_terms(t1, $::level, $::index) FROM t1 LIMIT 1; + } + } [list $terms] +} +proc check_terms_all {test terms} { + do_test $test.terms { + execsql { + SELECT dump_terms(t1) FROM t1 LIMIT 1; + } + } [list $terms] +} + +# Utility function to check for the expected doclist for the term in +# segment level/index. _all version does same for entire index. +proc check_doclist {test level index term doclist} { + # TODO(shess): Again, why can't the non-:: versions work? + set ::term $term + set ::level $level + set ::index $index + do_test $test { + execsql { + SELECT dump_doclist(t1, $::term, $::level, $::index) FROM t1 LIMIT 1; + } + } [list $doclist] +} +proc check_doclist_all {test term doclist} { + set ::term $term + do_test $test { + execsql { + SELECT dump_doclist(t1, $::term) FROM t1 LIMIT 1; + } + } [list $doclist] +} + +#************************************************************************* +# Test the segments resulting from straight-forward inserts. +db eval { + DROP TABLE IF EXISTS t1; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) VALUES (1, 'This is a test'); + INSERT INTO t1 (rowid, c) VALUES (2, 'That was a test'); + INSERT INTO t1 (rowid, c) VALUES (3, 'This is a test'); +} + +# Check for expected segments and expected matches. +do_test fts2p-1.0.segments { + execsql { + SELECT level, idx FROM t1_segdir ORDER BY level, idx; + } +} {0 0 0 1 0 2} +do_test fts2p-1.0.matches { + execsql { + SELECT OFFSETS(t1) FROM t1 + WHERE t1 MATCH 'this OR that OR was OR a OR is OR test' ORDER BY rowid; + } +} [list {0 0 0 4 0 4 5 2 0 3 8 1 0 5 10 4} \ + {0 1 0 4 0 2 5 3 0 3 9 1 0 5 11 4} \ + {0 0 0 4 0 4 5 2 0 3 8 1 0 5 10 4}] + +# Check the specifics of the segments constructed. +# Logical view of entire index. +check_terms_all fts2p-1.0.1 {a is test that this was} +check_doclist_all fts2p-1.0.1.1 a {[1 0[2]] [2 0[2]] [3 0[2]]} +check_doclist_all fts2p-1.0.1.2 is {[1 0[1]] [3 0[1]]} +check_doclist_all fts2p-1.0.1.3 test {[1 0[3]] [2 0[3]] [3 0[3]]} +check_doclist_all fts2p-1.0.1.4 that {[2 0[0]]} +check_doclist_all fts2p-1.0.1.5 this {[1 0[0]] [3 0[0]]} +check_doclist_all fts2p-1.0.1.6 was {[2 0[1]]} + +# Segment 0,0 +check_terms fts2p-1.0.2 0 0 {a is test this} +check_doclist fts2p-1.0.2.1 0 0 a {[1 0[2]]} +check_doclist fts2p-1.0.2.2 0 0 is {[1 0[1]]} +check_doclist fts2p-1.0.2.3 0 0 test {[1 0[3]]} +check_doclist fts2p-1.0.2.4 0 0 this {[1 0[0]]} + +# Segment 0,1 +check_terms fts2p-1.0.3 0 1 {a test that was} +check_doclist fts2p-1.0.3.1 0 1 a {[2 0[2]]} +check_doclist fts2p-1.0.3.2 0 1 test {[2 0[3]]} +check_doclist fts2p-1.0.3.3 0 1 that {[2 0[0]]} +check_doclist fts2p-1.0.3.4 0 1 was {[2 0[1]]} + +# Segment 0,2 +check_terms fts2p-1.0.4 0 2 {a is test this} +check_doclist fts2p-1.0.4.1 0 2 a {[3 0[2]]} +check_doclist fts2p-1.0.4.2 0 2 is {[3 0[1]]} +check_doclist fts2p-1.0.4.3 0 2 test {[3 0[3]]} +check_doclist fts2p-1.0.4.4 0 2 this {[3 0[0]]} + +#************************************************************************* +# Test the segments resulting from inserts followed by a delete. +db eval { + DROP TABLE IF EXISTS t1; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) VALUES (1, 'This is a test'); + INSERT INTO t1 (rowid, c) VALUES (2, 'That was a test'); + INSERT INTO t1 (rowid, c) VALUES (3, 'This is a test'); + DELETE FROM t1 WHERE rowid = 1; +} + +do_test fts2p-1.1.segments { + execsql { + SELECT level, idx FROM t1_segdir ORDER BY level, idx; + } +} {0 0 0 1 0 2 0 3} +do_test fts2p-1.1.matches { + execsql { + SELECT OFFSETS(t1) FROM t1 + WHERE t1 MATCH 'this OR that OR was OR a OR is OR test' ORDER BY rowid; + } +} {{0 1 0 4 0 2 5 3 0 3 9 1 0 5 11 4} {0 0 0 4 0 4 5 2 0 3 8 1 0 5 10 4}} + +check_terms_all fts2p-1.1.1 {a is test that this was} +check_doclist_all fts2p-1.1.1.1 a {[2 0[2]] [3 0[2]]} +check_doclist_all fts2p-1.1.1.2 is {[3 0[1]]} +check_doclist_all fts2p-1.1.1.3 test {[2 0[3]] [3 0[3]]} +check_doclist_all fts2p-1.1.1.4 that {[2 0[0]]} +check_doclist_all fts2p-1.1.1.5 this {[3 0[0]]} +check_doclist_all fts2p-1.1.1.6 was {[2 0[1]]} + +check_terms fts2p-1.1.2 0 0 {a is test this} +check_doclist fts2p-1.1.2.1 0 0 a {[1 0[2]]} +check_doclist fts2p-1.1.2.2 0 0 is {[1 0[1]]} +check_doclist fts2p-1.1.2.3 0 0 test {[1 0[3]]} +check_doclist fts2p-1.1.2.4 0 0 this {[1 0[0]]} + +check_terms fts2p-1.1.3 0 1 {a test that was} +check_doclist fts2p-1.1.3.1 0 1 a {[2 0[2]]} +check_doclist fts2p-1.1.3.2 0 1 test {[2 0[3]]} +check_doclist fts2p-1.1.3.3 0 1 that {[2 0[0]]} +check_doclist fts2p-1.1.3.4 0 1 was {[2 0[1]]} + +check_terms fts2p-1.1.4 0 2 {a is test this} +check_doclist fts2p-1.1.4.1 0 2 a {[3 0[2]]} +check_doclist fts2p-1.1.4.2 0 2 is {[3 0[1]]} +check_doclist fts2p-1.1.4.3 0 2 test {[3 0[3]]} +check_doclist fts2p-1.1.4.4 0 2 this {[3 0[0]]} + +check_terms fts2p-1.1.5 0 3 {a is test this} +check_doclist fts2p-1.1.5.1 0 3 a {[1]} +check_doclist fts2p-1.1.5.2 0 3 is {[1]} +check_doclist fts2p-1.1.5.3 0 3 test {[1]} +check_doclist fts2p-1.1.5.4 0 3 this {[1]} + +#************************************************************************* +# Test results when all references to certain tokens are deleted. +db eval { + DROP TABLE IF EXISTS t1; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) VALUES (1, 'This is a test'); + INSERT INTO t1 (rowid, c) VALUES (2, 'That was a test'); + INSERT INTO t1 (rowid, c) VALUES (3, 'This is a test'); + DELETE FROM t1 WHERE rowid IN (1,3); +} + +# Still 4 segments because 0,3 will contain deletes for rowid 1 and 3. +do_test fts2p-1.2.segments { + execsql { + SELECT level, idx FROM t1_segdir ORDER BY level, idx; + } +} {0 0 0 1 0 2 0 3} +do_test fts2p-1.2.matches { + execsql { + SELECT OFFSETS(t1) FROM t1 + WHERE t1 MATCH 'this OR that OR was OR a OR is OR test' ORDER BY rowid; + } +} {{0 1 0 4 0 2 5 3 0 3 9 1 0 5 11 4}} + +check_terms_all fts2p-1.2.1 {a is test that this was} +check_doclist_all fts2p-1.2.1.1 a {[2 0[2]]} +check_doclist_all fts2p-1.2.1.2 is {} +check_doclist_all fts2p-1.2.1.3 test {[2 0[3]]} +check_doclist_all fts2p-1.2.1.4 that {[2 0[0]]} +check_doclist_all fts2p-1.2.1.5 this {} +check_doclist_all fts2p-1.2.1.6 was {[2 0[1]]} + +check_terms fts2p-1.2.2 0 0 {a is test this} +check_doclist fts2p-1.2.2.1 0 0 a {[1 0[2]]} +check_doclist fts2p-1.2.2.2 0 0 is {[1 0[1]]} +check_doclist fts2p-1.2.2.3 0 0 test {[1 0[3]]} +check_doclist fts2p-1.2.2.4 0 0 this {[1 0[0]]} + +check_terms fts2p-1.2.3 0 1 {a test that was} +check_doclist fts2p-1.2.3.1 0 1 a {[2 0[2]]} +check_doclist fts2p-1.2.3.2 0 1 test {[2 0[3]]} +check_doclist fts2p-1.2.3.3 0 1 that {[2 0[0]]} +check_doclist fts2p-1.2.3.4 0 1 was {[2 0[1]]} + +check_terms fts2p-1.2.4 0 2 {a is test this} +check_doclist fts2p-1.2.4.1 0 2 a {[3 0[2]]} +check_doclist fts2p-1.2.4.2 0 2 is {[3 0[1]]} +check_doclist fts2p-1.2.4.3 0 2 test {[3 0[3]]} +check_doclist fts2p-1.2.4.4 0 2 this {[3 0[0]]} + +check_terms fts2p-1.2.5 0 3 {a is test this} +check_doclist fts2p-1.2.5.1 0 3 a {[1] [3]} +check_doclist fts2p-1.2.5.2 0 3 is {[1] [3]} +check_doclist fts2p-1.2.5.3 0 3 test {[1] [3]} +check_doclist fts2p-1.2.5.4 0 3 this {[1] [3]} + +#************************************************************************* +# Test results when everything is optimized manually. +db eval { + DROP TABLE IF EXISTS t1; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) VALUES (1, 'This is a test'); + INSERT INTO t1 (rowid, c) VALUES (2, 'That was a test'); + INSERT INTO t1 (rowid, c) VALUES (3, 'This is a test'); + DELETE FROM t1 WHERE rowid IN (1,3); + DROP TABLE IF EXISTS t1old; + ALTER TABLE t1 RENAME TO t1old; + CREATE VIRTUAL TABLE t1 USING fts2(c); + INSERT INTO t1 (rowid, c) SELECT rowid, c FROM t1old; + DROP TABLE t1old; +} + +# Should be a single optimal segment with the same logical results. +do_test fts2p-1.3.segments { + execsql { + SELECT level, idx FROM t1_segdir ORDER BY level, idx; + } +} {0 0} +do_test fts2p-1.3.matches { + execsql { + SELECT OFFSETS(t1) FROM t1 + WHERE t1 MATCH 'this OR that OR was OR a OR is OR test' ORDER BY rowid; + } +} {{0 1 0 4 0 2 5 3 0 3 9 1 0 5 11 4}} + +check_terms_all fts2p-1.3.1 {a test that was} +check_doclist_all fts2p-1.3.1.1 a {[2 0[2]]} +check_doclist_all fts2p-1.3.1.2 test {[2 0[3]]} +check_doclist_all fts2p-1.3.1.3 that {[2 0[0]]} +check_doclist_all fts2p-1.3.1.4 was {[2 0[1]]} + +check_terms fts2p-1.3.2 0 0 {a test that was} +check_doclist fts2p-1.3.2.1 0 0 a {[2 0[2]]} +check_doclist fts2p-1.3.2.2 0 0 test {[2 0[3]]} +check_doclist fts2p-1.3.2.3 0 0 that {[2 0[0]]} +check_doclist fts2p-1.3.2.4 0 0 was {[2 0[1]]} + +finish_test |