1

我的查询有很大的问题，执行时间超过 10 秒

许多子查询和左连接

这是我的查询

SELECT   tickets.tic_id,
         tickets.tic_title,
         tic_statue,
         tic_priority,
         tickets.tic_msg,
         customers.cus_name,
         employee.emp_name,
         count(comments.com_id),
         customers.cus_id ,
         (
            SELECT   comments.com_msg
            from     comments
            WHERE    tickets.tic_id = comments.tic_id
            order by com_id DESC
            limit    1
          ) AS last_comment,
          (
              SELECT customers.cus_name
              from   customers INNER JOIN comments on (customers.cus_id = comments.tic_uid)
              where  comments.com_msg = last_comment
              LIMIT  1
          )AS by_customer,
          (
            SELECT employee.emp_name
            from   employee INNER JOIN comments on (employee.emp_id = comments.tic_emp)
            where  comments.com_msg = last_comment
            LIMIT  1)AS by_employee,
         tic_date,
         (SELECT   comments.com_statue
          from     comments
          WHERE    tickets.tic_id = comments.tic_id
          order by com_id DESC
          limit    1)
           AS last_comment_color
FROM     tickets
         LEFT JOIN employee
           on (tickets.tic_emp = employee.emp_id)
         LEFT JOIN customers
           on (tickets.tic_cus = customers.cus_id)
         LEFT join cats
           on (tickets.tic_cat = cats.cat_id)
         LEFT JOIN comments
           on (tickets.tic_id = comments.tic_id)
GROUP by tickets.tic_id
ORDER BY tickets.tic_statue ASC, comments.com_time DESC
limit    50

结果工作正常，但时间长

索引：tickets.tic_id comments.com_id comments.tic_id customers.cus_id

谢谢 ：）

4

2 回答 2

2

试试这个

SELECT   tickets.tic_id,
         tickets.tic_title,
         tic_statue,
         tic_priority,
         tickets.tic_msg,
         customers.cus_name,
         employee.emp_name,
         count(comments.com_id),
         customers.cus_id ,
         @cm := coalesce( last_comment.COM_MSG, 000000 ) as COM_MSG,
         @cn := coalesce( by_customer.CUS_NAME, 000000 ) as CUS_NAME,
         @be := coalesce( by_employee.EMP, 000000 ) as EMP,
         @cs := coalesce( last_comment_color.COM_STATUE, 000000 ) as COM_STATUE

        FROM
        ( select
          @cm := 0,
          @cn := 0,
          @be := 0,
          @cs := 0
        ) sqlvars,

        tickets

         LEFT JOIN (
            SELECT   com_msg AS COM_MSG
            from     comments
            order by com_id DESC
            limit 1
          ) AS last_comment
          ON tickets.tic_id = last_comment.tic_id

         LEFT JOIN (
              SELECT com_msg,cus_name AS CUS_NAME
              from   customers INNER JOIN comments on (cus_id = comments.tic_uid)
              LIMIT  1
          )AS by_customer,
          ON by_customer.com_msg = last_comment

          LEFT JOIN (
            SELECT com_msg,emp_name AS EMP
            from   employee INNER JOIN comments on (emp_id = comments.tic_emp)
            LIMIT  1)AS by_employee
            ON by_employee.com_msg = last_comment

         LEFT JOIN (SELECT tic_id ,com_statue AS COM_STATUE
          from     comments
          order by com_id DESC
          limit    1)
           AS last_comment_color   
         ON  tickets.tic_id = last_comment_color.tic_id


         LEFT JOIN employee
           on (tickets.tic_emp = employee.emp_id)
         LEFT JOIN customers
           on (tickets.tic_cus = customers.cus_id)
         LEFT join cats
           on (tickets.tic_cat = cats.cat_id)
         LEFT JOIN comments
           on (tickets.tic_id = comments.tic_id)
GROUP by tickets.tic_id
ORDER BY tickets.tic_statue ASC, comments.com_time DESC
limit    50
于 2012-05-17T19:15:03.260 回答
1

我之前用类似的表创建了一个大型票务系统。tickets我发现用一列更新主表要容易十倍last_comment_id，这样您就可以在没有子选择的情况下获得所有信息。目前，您必须扫描您的comments表格五次，因此取决于它的大小，这将需要一段时间。

于 2012-05-17T18:52:26.110 回答