Question

我在MySQL上运行此查询。

Select Vendor, sum(Rate) as Rate
from (select case Vendor when 'NSN' then 'Nokia' else Vendor end as Vendor, Rate
from (  Select vendor ,(count(1) )*100/(Select count(id_incident)from incident where open_time between '2015-01-01'and'2015-01-30') as Rate from incident  where open_time between '2015-01-01'and'2015-01-30'group by upper (vendor) )) as y group by vendor;

并且它给出了这个错误：

错误代码1248，SQL状态42000：每个派生表必须有自己的别名“。

问题是什么？

Answer 1

您忘记给内部子查询添加别名。我选择了x

Select Vendor, sum(Rate) as Rate
from 
(
   select case Vendor when 'NSN' then 'Nokia' else Vendor end as Vendor, Rate
   from 
   (  
     Select vendor ,(count(1) )*100/(Select count(id_incident)from incident where open_time between '2015-01-01'and'2015-01-30') as Rate 
     from incident  
     where open_time between '2015-01-01'and'2015-01-30'
     group by upper (vendor) 
   ) as x
) as y 
group by vendor;

Answer 2

您的查询比以前更加复杂：

 Select (case when vendor = 'NSN' then 'NOKIA' else upper(vendor) end) as vendor,
        count(*)*100 / overall.cnt as Rate 
 from incident i cross join
      (Select count(*) as cnt
       from incident
       where open_time between '2015-01-01'and'2015-01-30'
      ) overall
 where open_time between '2015-01-01'and'2015-01-30'
 group by (case when vendor = 'NSN' then 'NOKIA' else upper(vendor) end)

评论：

子查询（在MySQL中）会产生额外的开销，用于实现中间结果。
每行调用一次select子句中的子查询。在from中，它只计算一次。
如果id_incident不为NULL，则只需使用count(*)（或count(1)）。如果您没有真正检查count()值，则在NULL中添加列会产生误导。

错误代码1248，SQL状态42000：每个派生表必须具有自己的别名

2 个答案: