在自定义Wordpress搜索中“使用临时”

Question

我目前使用一个自定义的wordpress MySQL查询来为我的一些购物网站获取相关产品。我刚刚分析了我的页面，我注意到这个查询大约需要3秒，但我不确定如何优化它。查询如下：

explain select 
    p . *,
    unix_timestamp(p.post_modified) as post_modified_ut,
    unix_timestamp(p.post_date) as post_date_ut,
    (((2.3 * (MATCH (p.post_title) AGAINST ('Motorola+MBP+36+Digital+Video+Monitor' IN BOOLEAN MODE)))) + (0.6 * (MATCH (p.post_content) AGAINST ('Motorola+MBP+36+Digital+Video+Monitor' IN BOOLEAN MODE)))) AS relevance
from
    wp_posts as p,
    wp_terms as t,
    wp_term_taxonomy as tt,
    wp_term_relationships as tr
where
    (MATCH (p.post_title , p.post_content) AGAINST ('Motorola+MBP+36+Digital+Video+Monitor' IN BOOLEAN MODE))
        and tr.object_id = p.ID
        and tr.term_taxonomy_id = tt.term_taxonomy_id
        and tt.term_id = t.term_id
        and p.post_type = 'post'
        and p.post_status in ('inherit' , 'publish')        
group by p.ID , p.post_title
order by relevance desc
limit 5;

解释的结果是：

+----+-------------+-------+--------+-------------------------------------------------+------------------+---------+------------------------------------+------+---------------------------------+
| id | select_type | table | type   | possible_keys                                   | key              | key_len | ref                                | rows | Extra                           |
+----+-------------+-------+--------+-------------------------------------------------+------------------+---------+------------------------------------+------+---------------------------------+
|  1 | SIMPLE      | tt    | ALL    | PRIMARY,term_id_taxonomy                        | NULL             | NULL    | NULL                               | 2822 | Using temporary; Using filesort |
|  1 | SIMPLE      | t     | eq_ref | PRIMARY                                         | PRIMARY          | 8       | reviewexplorer.tt.term_id          |    1 | Using index                     |
|  1 | SIMPLE      | tr    | ref    | PRIMARY,term_taxonomy_id                        | term_taxonomy_id | 8       | reviewexplorer.tt.term_taxonomy_id |    5 |                                 |
|  1 | SIMPLE      | p     | eq_ref | PRIMARY,type_status_date,searches,searches_more | PRIMARY          | 8       | reviewexplorer.tr.object_id        |    1 | Using where                     |
+----+-------------+-------+--------+-------------------------------------------------+------------------+---------+------------------------------------+------+---------------------------------+

正如您所看到的，我正在使用临时表，我不想这样做，我想创建一个索引来加快查询速度，但我不能完全理解explain告诉我什么。

Answer 1

显然，MySQL不能在wp_term_taxonomy表上使用任何索引，必须在临时表中examine+sort所有2822行。也就是说，2822行并不是很大，它可以解释这么长的查询时间。是数据库负载高还是磁盘负载高？不管怎样..。

通过更仔细地查看您的查询，它是非常混乱的。显然你有一个wp_term_taxonomy.term_taxonomy_id和wp_term_relationships.term_taxonomy_id，但是一个可能的索引在wp_term_taxonomy.term_id_taxonomy上。

看到区别了吗？term_taxonomy_id term_id_taxonomy≠。

这是打字错误吗？一个错误？一个“特征”？