我试图在包含“The Zen Circus”的字段中找到“禅”字符串。 我有一个FULLTEXT索引。
select url,name,
, MATCH(name) AGAINST ( 'zen*' IN BOOLEAN MODE) as A
, MATCH(name) AGAINST ( '"the zen*"' IN BOOLEAN MODE) as B
, MATCH(name) AGAINST ('>the* zen' IN BOOLEAN MODE) as C
, MATCH(name) AGAINST ('thezen*' IN BOOLEAN MODE) as D
, MATCH(name) AGAINST ('cir*' IN BOOLEAN MODE) as E
, MATCH(name) AGAINST ('circus*' IN BOOLEAN MODE) as F
from pages where url='thezencircus'
我得到了这个结果:
url = thezencircus
name = The Zen Circus
A = 0 (why?)
B = 0 (why?)
C = 0 (why?)
D = 0 (ok)
E = 1 (ok)
F = 1 (ok)
我还在msyql配置文件中输入了ft_min_word_len = 2.
有什么想法吗?
答案 0 :(得分:10)
请参阅此回答MySQL full text search for words with three or less letters
的[mysqld]
的ft_min_word_len = 3
然后,您必须重新启动服务器并重建FULLTEXT索引。
记住重新启动并重建索引。
<小时/> 的修改show variables like 'ft_%';
以确认字长与您设置的字长相匹配。
答案 1 :(得分:2)
Kao是对的,但这完全没有帮助我-为什么? InnoDB表使用不同的设置-尝试更新此设置并使用以下命令进行检查:
show variables like 'innodb_ft_min_token_size';