innodb_fts/t/ngram_debug.test

#
#   InnoDB FULLTEXT SEARCH: CJK support - ngram parser
#
#

--source include/have_innodb.inc

# Must have debug code to use SET DEBUG
--source include/have_debug.inc
# Embedded server does not support restarting
--source include/not_embedded.inc

# Enable diag print to print query parse tree
SET GLOBAL innodb_ft_enable_diag_print = 1;

-- echo # Test Case 1: Test English with Default(latin1) Charset
# Ignore ngram token check for latin1.
SET DEBUG="+d,fts_instrument_ignore_ngram_check";

CREATE TABLE articles (
	id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
	title VARCHAR(200)
	) ENGINE=InnoDB;

ALTER TABLE articles ADD FULLTEXT INDEX (title) WITH PARSER ngram;

INSERT INTO articles (title) VALUES
  ('what is M'),
  ('Myname is wonder'),
  ('My SQL is good'),
  ('MySQL database'),
  ('Mind a good thing'),
  ('Mind good idea');

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('m*' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('my*' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('+mind -mysql' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('+(mind thing) -idea' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('mysql');

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('mysql' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('"mysql"' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title) AGAINST('"my sql"' IN BOOLEAN MODE);

SET DEBUG="-d,fts_instrument_ignore_ngram_check";

DROP TABLE articles;

SET NAMES utf8;

-- echo # Test Case 2: Test Stopwords
# Compare with results without stopword check.
SET DEBUG="+d,fts_instrument_ignore_ngram_check";

# Without stopwords
CREATE TABLE articles (
	id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
	title VARCHAR(200),
	body TEXT
	) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci;

CREATE FULLTEXT INDEX ft_index ON articles(title, body) WITH PARSER ngram;

INSERT INTO articles (title, body) VALUES
  ('从简单购物清单到画展','或企业网络中的海量数据'),
  ('要想将数据添加到数据库','访问、处理计算机数据库中保存的数据'),
  ('从简单的购物清单到画展','或企业网络中的海量数据'),
  ('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'),
  ('数据库是数据的结构化集合','它可以是任何东西');

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('的* 或* 东西' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('简单的' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('或访问' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('"数据的结构化"' IN BOOLEAN MODE);

SET DEBUG="-d,fts_instrument_ignore_ngram_check";

DROP TABLE articles;

# With Stopwords
CREATE TABLE articles (
	id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY,
	title VARCHAR(200),
	body TEXT
	) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci;

# Define a user stopword table and set to it
CREATE TABLE user_stopword(value varchar(30)) ENGINE = InnoDB
	DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci;

INSERT INTO user_stopword VALUES('的'),('或'),('东西'),('ab');

SET GLOBAL innodb_ft_server_stopword_table = 'test/user_stopword';

INSERT INTO articles (title, body) VALUES
  ('从简单购物清单到画展','或企业网络中的海量数据'),
  ('要想将数据添加到数据库','访问、处理计算机数据库中保存的数据');

CREATE FULLTEXT INDEX ft_index ON articles(title, body) WITH PARSER ngram;

INSERT INTO articles (title, body) VALUES
  ('从简单的购物清单到画展','或企业网络中的海量数据'),
  ('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'),
  ('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据 abc xyz'),
  ('数据库是数据的结构化集合','它可以是任何东西');

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('的* 或* 东西' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('简单的' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('或访问' IN BOOLEAN MODE);

SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('"数据的结构化"' IN BOOLEAN MODE);

SELECT * FROM user_stopword WHERE value like 'ab' ;
SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('"ab"' );
SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('数据 +ab' IN BOOLEAN MODE);
SELECT * FROM articles WHERE
	MATCH(title, body) AGAINST('数据 (+xyz -bc)' IN BOOLEAN MODE);

DROP TABLE user_stopword;
DROP TABLE articles;
SET GLOBAL innodb_ft_server_stopword_table=default;

-- echo # Test Case 3: Check with FTS debug points
CREATE TABLE articles (
        seq INT AUTO_INCREMENT NOT NULL PRIMARY KEY,
        c1 VARCHAR(200) CHARACTER SET big5 COLLATE big5_chinese_ci,
        c2 CHAR(200) CHARACTER SET gbk COLLATE gbk_chinese_ci,
        c3 CHAR(200) CHARACTER SET utf8,
        j1 TEXT CHARACTER SET ujis COLLATE ujis_japanese_ci,
        j2 VARCHAR(200) CHARACTER SET sjis COLLATE sjis_japanese_ci,
        j3 VARCHAR(200) CHARACTER SET ujis,
        k1 CHAR(200) CHARACTER SET euckr COLLATE euckr_korean_ci,
        k2 CHAR(200) CHARACTER SET utf8,
        e1 CHAR(200) CHARACTER SET latin1,
        FULLTEXT KEY `con1` (c1),
        FULLTEXT KEY `con2` (j1,j3),
        FULLTEXT KEY `con3` (k2) WITH PARSER ngram
        ) ENGINE=InnoDB ;
SET DEBUG='+d,alter_table_rollback_new_index';
-- error ER_UNKNOWN_ERROR
ALTER TABLE articles ADD FULLTEXT INDEX (c2(64));
-- error ER_UNKNOWN_ERROR
ALTER TABLE articles ADD FULLTEXT INDEX (j1,j3);
-- error ER_UNKNOWN_ERROR
ALTER TABLE articles ADD FULLTEXT INDEX (k1);
-- error ER_UNKNOWN_ERROR
ALTER TABLE articles ADD FULLTEXT INDEX (e1);
SET DEBUG='-d,alter_table_rollback_new_index';

# Abort the operation in dict_create_index_step by setting
# return status of dict_create_index_tree_step() to DB_OUT_OF_MEMORY
# The newly create dict_index_t should be removed from fts cache
SET DEBUG="+d,ib_dict_create_index_tree_fail";
--error ER_OUT_OF_RESOURCES
ALTER TABLE articles ADD FULLTEXT INDEX (c2(64));
--error ER_OUT_OF_RESOURCES
ALTER TABLE articles ADD FULLTEXT INDEX (j1,j3);
--error ER_OUT_OF_RESOURCES
ALTER TABLE articles ADD FULLTEXT INDEX (k1);
--error ER_OUT_OF_RESOURCES
ALTER TABLE articles ADD FULLTEXT INDEX (e1);
SET DEBUG="-d,ib_dict_create_index_tree_fail";

--error ER_INNODB_FT_LIMIT
ALTER TABLE articles ADD FULLTEXT INDEX (e1), ADD FULLTEXT INDEX (k1);
DROP TABLE articles;

SET GLOBAL innodb_ft_enable_diag_print = default;