自然语言全文本检索
缺省或者modifier被设置为in natural language mode,都是进行自然语言检索。对于表中的每一行,match()都会返回一个关联值。
mysql> CREATE TABLE articles ( -> id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, -> title VARCHAR(200), -> body TEXT, -> FULLTEXT ( title , body ) -> ) ENGINE=INNODB; Query OK, 0 rows affected (0.04 sec) mysql> INSERT INTO articles (title,body) VALUES -> ('MySQL Tutorial','DBMS stands for DataBase ...'), -> ('How To Use MySQL Well','After you went through a ...'), -> ('Optimizing MySQL','In this tutorial we will show ...'), -> ('1001 MySQL Tricks','1. Never run mysqld as root. 2. ...'), -> ('MySQL vs. YourSQL','In the following database comparison ...'), -> ('MySQL Security','When configured properly, MySQL ...'); Query OK, 6 rows affected (0.00 sec) Records: 6 Duplicates: 0 Warnings: 0 mysql> select * from articles -> where match(title,body) -> against('database' in natural language mode); +----+-------------------+------------------------------------------+ | id | title | body | +----+-------------------+------------------------------------------+ | 1 | MySQL Tutorial | DBMS stands for DataBase ... | | 5 | MySQL vs. YourSQL | In the following database comparison ... | +----+-------------------+------------------------------------------+ 2 rows in set (0.00 sec) mysql>
缺省情况下,检索是大小写不敏感的。如果要想进行大小写敏感的检索,对于索引的列要进行二进制collation。比如字符集类型为latin1的列可以修改为Latin1_bin。
当match()被作为where子句的时候,返回的行会被自动排序,根据检索关联度进行排序。
mysql> INSERT INTO articles (title,body) VALUES -> ('oracle Tutorial','DBMS stands for DataBase ...DataBase'); Query OK, 1 row affected (0.00 sec) mysql> select * from articles -> where match(title,body) -> against('database' in natural language mode); +----+-------------------+------------------------------------------+ | id | title | body | +----+-------------------+------------------------------------------+ | 7 | oracle Tutorial | DBMS stands for DataBase ...DataBase | | 1 | MySQL Tutorial | DBMS stands for DataBase ... | | 5 | MySQL vs. YourSQL | In the following database comparison ... | +----+-------------------+------------------------------------------+ 3 rows in set (0.00 sec) mysql>
可以查看一下匹配的次数:
#使用索引 mysql> SELECT -> COUNT(*) -> FROM -> articles -> WHERE -> MATCH (title , body) AGAINST ('database' IN NATURAL LANGUAGE MODE); +----------+ | COUNT(*) | +----------+ | 2 | +----------+ 1 row in set (0.00 sec) mysql> #使用全表扫描 mysql> SELECT -> COUNT(IF(MATCH (title , body) AGAINST ('database' IN NATURAL LANGUAGE MODE), -> 1, -> NULL)) AS count -> FROM -> articles; +-------+ | count | +-------+ | 3 | +-------+ 1 row in set (0.00 sec) mysql>
对于自然语言全文本检索,match()中的列名必须和全文索引中的列相同。上例中,如果想对title或body分开检索,就需要分别创建全文索引。
上面的例子中,基本展示了如何使用match()。返回的结果根据关联值的降序排列。
下面的例子,展示如何显式输出关联值得大小。返回的行不是有序的,因为select语句既不包含where也没有order by。
mysql> SELECT -> id, -> MATCH (title , body) AGAINST ('database' IN NATURAL LANGUAGE MODE) AS score -> FROM -> articles; +----+---------------------+ | id | score | +----+---------------------+ | 1 | 0.22764469683170319 | | 2 | 0 | | 3 | 0 | | 4 | 0 | | 5 | 0.22764469683170319 | | 6 | 0 | +----+---------------------+ 6 rows in set (0.00 sec) mysql>
下面的例子更复杂,查询返回关联值得具体值,同时进行降序排序。为了实现这个目的,使用了match()两次。这样的语句不会有额外的开销,优化器会注意到两次match()调用是一样的,所以只会执行全文检索一次。
mysql> SELECT -> id, -> body, -> MATCH (title , body) AGAINST ('Security implications of running MySQL as root' IN NATURAL LANGUAGE MODE) AS score -> FROM -> articles -> WHERE -> MATCH (title , body) AGAINST ('Security implications of running MySQL as root' IN NATURAL LANGUAGE MODE); +----+------------------------------------------+----------------------------+ | id | body | score | +----+------------------------------------------+----------------------------+ | 4 | 1. Never run mysqld as root. 2. ... | 0.6055193543434143 | | 6 | When configured properly, MySQL ... | 0.6055193543434143 | | 1 | DBMS stands for DataBase ... | 0.000000001885928302414186 | | 2 | After you went through a ... | 0.000000001885928302414186 | | 3 | In this tutorial we will show ... | 0.000000001885928302414186 | | 5 | In the following database comparison ... | 0.000000001885928302414186 | +----+------------------------------------------+----------------------------+ 6 rows in set (0.00 sec) mysql>
用双引号引起来的词组,检索匹配的结果只能是双引号中的字面值。全文检索会将双引号中的内容分解成单词,然后执行检索匹配。非单词字符是不需要匹配的,只是按照其中的单词顺序进行匹配,比如"test phrase"和"test, phrase"是匹配的。
全文检索会将字母、数字、下划线的组合当成一个word。但是也会将'当成一个word序列,不过一行只能有一个',比如会将aaa'bbb当成一个单词,但是aaa''bbb就不是一个单词了,而是两个。如果'放在开头或者结果,会被丢弃。
内嵌的文本解释器决定单词的开头和结尾,根据delimiter符号进行判断,比如逗号、空格、点号。如果不是根据delimiter分割的,比如中文,解释器就无法判断出单词的开头和结尾了。
所以,用户必须使用某些delimiter对文本进行处理后再检索。在5.7.6中可以使用插件ngram解释器来实现对中文、日文、韩文的支持,或者使用MeCab解释器来支持日文。
也可以自己编写插件解释器。示例代码位于plugin/fulltext目录。
在全文检索中,有些单词是被忽略的:
--太短的单词。默认最小长度是3个字符(innodb)、4个字符(myisam)。可以设置innodb中的innodb_ft_min_token_size、myisam中的ft_min_word_len
--stopword中的单词会被忽略。stopword是指那些类似the、some一样太普通以致被认为是没有什么语义值得单词。有一个内嵌stopword列表。也可以重新定义。
每个正确的单词在查询时都被会加权,根据其在集合和查询中的重要性。所以出现频率越高,权重就越低。单词的权重会被用来计算行的关联值。
全文检索如果本生行数就比较少,可能检索不出正确的结果。