mysql全文搜索失败

一尘不染

mysql全文搜索失败

mysql

这是我测试过的。

mysql> select * from product;
+------------+---------+---------------+
| Id         | Product | ProductIdType |
+------------+---------+---------------+
| B00005N5PF | one pen | ASIN          |
| B000J5XS3C |         | ASIN          |
+------------+---------+---------------+
2 rows in set (0.00 sec)

mysql> select * from product p where match (p.Product) against ('pen' in boolean mode);
Empty set (0.00 sec)

mysql> select * from product p where match (p.Product) against ('one pen' in boolean mode);
Empty set (0.00 sec)

这是create语句。

CREATE TABLE product
(
    Id                VARCHAR(16),
    Product           VARCHAR(128),
    ProductIdType     VARCHAR(8),
  PRIMARY KEY (Id),
  FULLTEXT (Product)
) ENGINE=MyISAM;

等号和“赞”效果很好。所以为什么？

阅读 321

2020-05-17

共1个答案

一尘不染

-- drop table testproduct;
CREATE TABLE testproduct
(
    Id                VARCHAR(16),
    prod_name           TEXT,
    ProductIdType     VARCHAR(8),
  PRIMARY KEY (Id),
  FULLTEXT (prod_name)
) ENGINE=MyISAM;

insert into testproduct (id,prod_name,productidtype) values ('B00005N5PF','one pen and a good price for a pen','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('B570J5XS3C',null,'ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C00ZZ5N5PF','let us get rid of some noise','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('D00LL5N5PA','four score and seven years ago our fore...','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('EEEZZ5N5PF','he has a harpoon','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C01ZZ5N5PF','and then we','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('B00ZZ5N5PF','he has a pen in his pocket not a banana','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C02ZZ5N5PF','went to the store','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C03ZZ5N5PF','and decided that we should buy some','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C04ZZ5N5PF','fruit cups or fruit or berries or pebbles','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C037Z5N5PF','then he and her she and it','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('C04K95N5PF','threw some daggers and a harpoon','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('D88895N5PF','more noise and some of this','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('D34595N5PF','this article about harpoons really drills into the throwing of harpoon or harpoons to those that deserve a harpoon','ASIN');
insert into testproduct (id,prod_name,productidtype) values ('D12395N5PF','and there we go','ASIN');

全文搜索需要多种多样的方式来消除重复的“噪音”。用最少的数据进行测试将产生较差的结果。将您的整个收藏集投入其中，以获取有意义的内容。如以下某些链接所示，甚至可以尝试搜索最小字数的设置。

停用词

有各种语言的停用词
MySql 列表，这些词表示在搜索过程中忽略的无关紧要的词。该列表已编译到服务器中，但可以覆盖，如本手册页和文本所示：

要覆盖默认停用词列表，请设置ft_stopword_file系统变量。（请参见第5.1.4节“服务器系统变量”。）变量值应为包含停用词列表的文件的路径名，或为禁用停用词过滤的空字符串。除非给出绝对路径名以指定其他目录，否则服务器将在数据目录中查找文件。更改此变量的值或停用词文件的内容后，重新启动服务器并重建FULLTEXT索引。

一些样本查询

-- select * from testproduct
SELECT * FROM testproduct WHERE MATCH(prod_name) AGAINST('score' IN BOOLEAN MODE);
SELECT * FROM testproduct WHERE MATCH(prod_name) AGAINST('harpoon' IN BOOLEAN MODE);
SELECT * FROM testproduct WHERE MATCH(prod_name) AGAINST('banana' IN BOOLEAN MODE);
SELECT * FROM testproduct WHERE MATCH(prod_name) AGAINST('years' IN BOOLEAN MODE);

获得多个单词匹配：

SELECT id,prod_name, match( prod_name )
AGAINST ( '+harpoon +article' IN BOOLEAN MODE ) AS relevance
FROM testproduct 
ORDER BY relevance DESC

在relevance列中给出实际权重：

SELECT id,prod_name, match( prod_name )
AGAINST ( '+harpoon +article' IN NATURAL LANGUAGE MODE) AS relevance
FROM testproduct 
ORDER BY relevance DESC
+------------+--------------------------------------------------------------------------------------------------------------------+--------------------+
| id         | prod_name                                                                                                          | relevance          |
+------------+--------------------------------------------------------------------------------------------------------------------+--------------------+
| D34595N5PF | this article about harpoons really drills into the throwing of harpoon or harpoons to those that deserve a harpoon | 3.6207125186920166 |
| EEEZZ5N5PF | he has a harpoon                                                                                                   | 1.2845110893249512 |
| C04K95N5PF | threw some daggers and a harpoon                                                                                   | 1.2559525966644287 |
|------------+--------------------------------------------------------------------------------------------------------------------+--------------------+

从这里取消了多个单词部分。谢谢斯宾塞

2020-05-17