为什么这么慢的mysql查询执行？

Question

select 
    'Owner' as 'Type',
    count(s.shipwynum) as Total,
    o.ownwynum as WYNum,
    o.ownshortnam as 'Short name',
    o.ownnam as 'Full name',
    cio.weburl as 'Web URL',
    count(if(s.statuscod = 'O',1,null)) as 'number of orders',
    count(if(s.statuscod = 'S',1,null)) as 'number of ships' 
from owner o 
left join ship s on o.ownwynum = s.benownwynum and s.deleted = 'N' and
      s.statuscod in ('O','S') and
      s.benownwynum != '' and s.benownwynum !='0' and s.benownwynum is not null 
left join companyinfo cio on cio.objwynum = s.benownwynum and
      cio.deleted = 'N' and cio.comflag = 'OW' where o.deleted = 'N'     
group by o.ownnam

该查询在9分钟55秒内运行，并检索10,106条记录。

我的问题是为什么要花费大量时间来获取数据，而减少执行时间又有什么问题呢？

但有趣的是，我已经运行了此查询：=

select * from ship 

执行过程需要3分12秒，将获取75,672条记录。

有优化的主意吗？

Answer 1

除非您真的有意避免s.shipwynum的空值，否则请尝试以下操作-替换：

count(s.shipwynum) as Total

与

count(*) as Total

通常，COUNT（*）的运行速度比COUNT（expression）的运行速度快。

---

另一个实验，替换为：

count(if(s.statuscod = 'O',1,null)) as 'number of orders'
count(if(s.statuscod = 'S',1,null)) as 'number of ships' 

与

sum(if(s.statuscod = 'O',1,0)) as 'number of orders'
sum(if(s.statuscod = 'S',1,0)) as 'number of ships'