Question

您好，我有800万行数据，需要优化Mysql查询以从该数据中获取一行。我正在使用以下查询，但是其服务器响应时间过长，导致页面加载速度出现问题

SELECT q.id
     , q.title
     , q.question
     , q.price
     , q.created
     , q.userid
     , q.duedate
     , q.tags
     , s.id subjectid
     , sc.id subcategoryid
     , s.title subject
     , sc.title subcategory
     , q.statusid
     , (SELECT  COUNT(id) FROM tbl_answers a WHERE a.questionid = q.id AND a.statusid = 1 AND a.deleted = 'N') q_num_answers
     , u.username
     , u.image
     , u.gender
     , (SELECT  COUNT(id) FROM tbl_answers a WHERE a.userid = q.userid AND a.statusid = 1 AND a.deleted = 'N') num_answers
     , (SELECT  COUNT(id) FROM tbl_questions WHERE userid = q.userid AND statusid = 1 AND deleted = 'N') AS num_questions
     , 0 amt_earned
     , 0 amt_spent
     , 0 num_sold
     , (SELECT  COUNT(ur.id) FROM tbl_users_ratings ur WHERE ur.userid = q.userid AND ur.deleted = 'N') u_num_ratings
     , (SELECT  COALESCE(SUM(ur.rating), 0) FROM tbl_users_ratings ur WHERE ur.userid = q.userid AND ur.deleted = 'N') u_score
  FROM tbl_questions q
  JOIN tbl_subjects s 
    ON q.subject = s.id
  JOIN tbl_subjects sc 
    ON q.subcategory = sc.id
  LEFT 
  JOIN tbl_users u 
    ON u.id = q.userid
 WHERE q.deleted = '$show_deleted' 
   AND q.id = ? 
 LIMIT 1

Answer 1

这些索引可能有帮助。（我假设id是每个表的PRIMARY KEY。）

ur:  (deleted, userid, rating)
a:  (deleted, statusid, userid)
a:  (deleted, statusid, questionid)

请提供EXPLAIN SELECT ...。

除非您需要检查COUNT(id)是否不是id，否则不要使用NULL。编写它的通常方法是COUNT(*)。

在一个地方，您正在检查deleted的提供值。在另一种方法中，您需要对其进行硬编码。也许错了？

AND  ur.deleted = 'N'

如果PRIMARY KEY的{{1}}为q，则将导致1行或无行。我想念什么？

id

（可能会有更多提示。请更改我建议的内容并遵循其他人的建议。然后让我们再看一遍。）

带有子查询和联接的MYSQL查询优化

1 个答案: