通过分组、内部查询和计数优化查询。

Question

我想获取相对的标签。

我在标签和帖子表之间有很多关系。

例如，对于 "爱 "这个标签，我试图获取他们发布的所有标签。Love 标签。

这是我的查询（标签67是 "爱 "的意思）。

SELECT hashtag_id, count(hashtag_id) as count
from post_hashtag 
where 
   # where posts has hashtag '67' 
   post_id in ( SELECT post_id FROM post_hashtag WHERE hashtag_id = 67 ) 
   # remove hashtag 67 from result
   and hashtag_id != 67
# group them and sort by count, so the must repeated hashtag is the best relative hashtag
GROUP by hashtag_id
ORDER by count desc
limit 4

我试着优化我的查询，但我不能更多的优化（目前需要2-12秒的时间，根据帖子的数量）。

有什么办法可以优化吗？

解释查询

+----+-------------+-----------------+------------+--------+---------------+---------------------------------------------------+---------+-------------------------------------------+---------+----------+-------------+
| id | select_type | table           | partitions | type   | possible_keys | key                                               | key_len | ref                                       | rows    | filtered | Extra       |
+----+-------------+-----------------+------------+--------+---------------+---------------------------------------------------+---------+-------------------------------------------+---------+----------+-------------+
|  1 | SIMPLE      | post_hashtag    | NULL       | index  | NULL          | fk_np_account_post_has_np_hashtag_np_hashtag1_idx | 4       | NULL                                      | 4623584 |   100.00 | Using index |
|  1 | SIMPLE      | hashtag         | NULL       | eq_ref | PRIMARY       | PRIMARY                                           | 4       | graphicj_novin.np_post_hashtag.hashtag_id |       1 |   100.00 | NULL        |
+----+-------------+-----------------+------------+--------+---------------+---------------------------------------------------+---------+-------------------------------------------+---------+----------+-------------+

post_hashtag 有这些领域

post_id,hashtag_id

两个字段都是外键

Answer 1

MySQL经常优化 WHERE IN (SELECT ...) 差。用一个 JOIN 而不是。

SELECT p1.hashtag_id, count(*) as count
from post_hashtag AS p1
JOIN post_hashtag AS p2 ON p1.post_id = p2.post_id
WHERE p1.hashtag_id != 67
AND p2.hashtag_id = 67
GROUP by p1.hashtag_id
ORDER by count desc
limit 4