mariadb/mysql-test/suite/innodb_fts/t/innodb_fts_proximity.test
Marko Mäkelä 228b7e4db5 MDEV-13626 Merge InnoDB test cases from MySQL 5.7
This imports and adapts a number of MySQL 5.7 test cases that are
applicable to MariaDB.

Some tests for old bug fixes are not that relevant because the code
has been refactored since then (especially starting with
MariaDB Server 10.6), and the tests would not reproduce the
original bug if the fix was reverted.

In the test innodb_fts.opt, there are many duplicate MATCH ranks, which
would make the results nondeterministic. The test was stabilized by
changing some LIMIT clauses or by adding sorted_result in those cases
where the purpose of a test was to show that no sorting took place
in the server.

In the test innodb_fts.phrase, MySQL 5.7 would generate FTS_DOC_ID that
are 1 larger than in MariaDB. In innodb_fts.index_table the difference is 2.
This is because in MariaDB, fts_get_next_doc_id() post-increments
cache->next_doc_id, while MySQL 5.7 pre-increments it.

Reviewed by: Thirunarayanan Balathandayuthapani
2023-11-08 12:17:14 +02:00

263 lines
7.9 KiB
Text

--source include/have_innodb.inc
--source include/maybe_versioning.inc
# This is the DDL function tests for innodb FTS
# Functional testing with FTS proximity search using '@'
# and try search default words
--disable_warnings
drop table if exists t1;
--enable_warnings
--disable_query_log
let $innodb_file_per_table_orig = `select @@innodb_file_per_table`;
--enable_query_log
# Create FTS table
CREATE TABLE t1 (
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
a VARCHAR(200),
b TEXT
) ENGINE= InnoDB;
# Create the FTS index again
CREATE FULLTEXT INDEX idx on t1 (a,b);
# Insert rows
INSERT INTO t1 (a,b) VALUES
('MySQL from Tutorial','DBMS stands for DataBase ...') ,
('when To Use MySQL Well','After that you went through a ...'),
('where will Optimizing MySQL','what In this tutorial we will show ...');
# Try to Search default stopword from innodb, "where", "will", "what"
# and "when" are all stopwords
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST ("where will");
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST ("when");
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST ("what" WITH QUERY EXPANSION);
# boolean No result expected
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST("whe*" IN BOOLEAN MODE);
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST("+what +will" IN BOOLEAN MODE);
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST("+from" IN BOOLEAN MODE);
SELECT * FROM t1 WHERE MATCH(a,b) AGAINST("+where +(show what)" IN BOOLEAN MODE);
# no result expected. Words are filtered out as stopwords
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"where will"@6' IN BOOLEAN MODE);
# no result expected
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"where will"@9' IN BOOLEAN MODE);
# insert record with @ character which is used in proximity search
INSERT INTO t1 (a,b) VALUES
('MySQL Tutorial','request docteam@oraclehelp.com ...') ,
('Trial version','query performace @1255 minute on 2.1Hz Memory 2GB...') ,
('when To Use MySQL Well','for free faq mail@xyz.com ...');
# proximity search with @ charcter
# We don't need more than one word in proximity search. Single word
# treated as single word search
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"request"@10' IN BOOLEAN MODE);
# If the distance is 0, it is treated as "phrase search"
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"Trial version"@0' IN BOOLEAN MODE);
# @ is word seperator
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"request docteam@oraclehelp.com"@10' IN BOOLEAN MODE);
# This should not return any document
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"1255 minute"@1' IN BOOLEAN MODE);
# This should return the first document. That is "1255" and "minutes" are
# in a two-word range (adjacent)
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"1255 minute"@2' IN BOOLEAN MODE);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"1255"@10' IN BOOLEAN MODE);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('1255' WITH QUERY EXPANSION);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"request docteam"@2' IN BOOLEAN MODE);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"1255 minute"' IN BOOLEAN MODE);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('request docteam@oraclehelp.com');
# Test across fields search
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"MySQL request"@3' IN BOOLEAN MODE);
# Two words are in 10 words range
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"Trial memory"@10' IN BOOLEAN MODE);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"Trial memory"@9' IN BOOLEAN MODE);
DROP TABLE t1;
# test on utf8 encoded proximity search
CREATE TABLE t1 (
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
a VARCHAR(200),
b TEXT
) CHARACTER SET = UTF8, ENGINE= InnoDB;
INSERT INTO t1 (a,b) VALUES
('MySQL from Tutorial','DBMS stands for DataBase ...') ,
('when To Use MySQL Well','After that you went through a ...'),
('where will Optimizing MySQL','what In this tutorial we will show ...');
CREATE FULLTEXT INDEX idx on t1 (a,b);
INSERT INTO t1 (a,b) VALUES
('MySQL Tutorial','request docteam@oraclehelp.com ...') ,
('Trial version','query performace @1255 minute on 2.1Hz Memory 2GB...'),
('when To Use MySQL Well','for free faq mail@xyz.com ...');
# Should have 2 rows. Note proximity search does require words in order
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"mysql use"@2' IN BOOLEAN MODE);
# Should return 0 row
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"mysql use"@1' IN BOOLEAN MODE);
INSERT INTO t1 (a,b) VALUES ('XYZ, long blob', repeat("a", 9000));
INSERT IGNORE INTO t1 (a,b) VALUES (repeat("b", 9000), 'XYZ, long blob');
# 2 rows match
SELECT count(*) FROM t1
WHERE MATCH (a,b)
AGAINST ('"xyz blob"@3' IN BOOLEAN MODE);
DROP TABLE t1;
set global innodb_file_per_table=1;
set names utf8;
--character_set utf8
# Test fts with externally stored long column
CREATE TABLE t1 (
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
a TEXT,
b TEXT,
c TEXT
) CHARACTER SET = UTF8, ROW_FORMAT=DYNAMIC, ENGINE= InnoDB;
INSERT INTO t1 (a,b,c) VALUES (repeat("b", 19000), 'XYZ, long text', 'very long blob');
INSERT INTO t1 (a,b,c) VALUES (repeat("b", 19000), 'XYZ, very little long blob very much blob', 'very long blob');
# Note 租车 is count as one word
INSERT INTO t1 (a,b,c) VALUES (repeat("b", 19000),"very 租车 供 blob","new 供需分析information");
CREATE FULLTEXT INDEX idx on t1 (a,b,c);
INSERT INTO t1 (a,b,c) VALUES (repeat("x", 19000), 'new, long text', 'very new blob');
INSERT INTO t1 (a,b,c) VALUES ('interesting, long text', repeat("x", 19000), 'very very good new blob');
# 3 rows should match
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very blob"@3' IN BOOLEAN MODE);
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very long blob"@0' IN BOOLEAN MODE);
# 4 rows should match
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very blob"@4' IN BOOLEAN MODE);
# 1 row should match
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"interesting blob"@9' IN BOOLEAN MODE);
# should have 3 rows
SELECT COUNT(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"interesting blob"@9 "very long blob"@0' IN BOOLEAN MODE);
# should have 3 rows
SELECT COUNT(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very blob"@4 - "interesting blob"@9' IN BOOLEAN MODE);
DROP TABLE t1;
CREATE TABLE t1 (
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
a VARCHAR(200),
b TEXT
) CHARACTER SET = UTF8, ENGINE= InnoDB;
# Space and special characters are not counted as word
INSERT INTO t1 (a,b) VALUES
('MySQL from Tutorial','DBMS stands for + DataBase ...');
CREATE FULLTEXT INDEX idx on t1 (a,b);
SELECT * FROM t1
WHERE MATCH (a,b)
AGAINST ('"stands database"@3' IN BOOLEAN MODE);
DROP TABLE t1;
# Test fts with externally stored long column
CREATE TABLE t1 (
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
a TEXT,
b TEXT,
c TEXT
) CHARACTER SET = UTF8, ROW_FORMAT=DYNAMIC, ENGINE= InnoDB;
INSERT INTO t1 (a,b,c) VALUES (repeat("b", 19000), 'XYZ, long text', 'very long blob');
INSERT INTO t1 (a,b,c) VALUES ('XYZ, 租车 very little long blob very much blob', repeat("b", 19000), 'very long but smaller blob');
CREATE FULLTEXT INDEX idx on t1 (a,b,c);
DELETE FROM t1;
INSERT INTO t1 (a,b,c) VALUES (repeat("b", 19000), 'XYZ, long text', 'very long blob');
INSERT INTO t1 (a,b,c) VALUES ('XYZ, 租车 very little long blob is a very much longer blob', repeat("b", 19000), 'this is very long but smaller blob');
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very blob"@4' IN BOOLEAN MODE);
SELECT count(*) FROM t1
WHERE MATCH (a,b,c)
AGAINST ('"very blob"@3' IN BOOLEAN MODE);
DROP TABLE t1;
eval SET GLOBAL innodb_file_per_table=$innodb_file_per_table_orig;