在我的sql语句中,我为同一个参数调用了两次sum,是否重复了这些努力？

Question

考虑下面我的SQL查询; 它为同一个参数调用两次.这是重复服务器完成的工作.有一个更好的方法吗？

SELECT      Status_Detail_Code, count(*) as 
   [Number of times assigned], round(sum(Duration)/60,2) as [total duration Hr]

 FROM         dbo.V_TIMELINE

 WHERE     (CADATE > N'20080101')  

group by Status_Detail_Code order by sum(Duration) desc

Answer 1

不,SQL Server重用聚合.

实际上,如果构建查询计划,您将看到SUM聚合运算符(如Stream Aggregate)表示为类似的结果集Expr****.

此表达式的值稍后将用作其他运算符的输入.

这是示例查询:

SELECT  ROUND(SUM(id), -1)
FROM    master
GROUP BY
        name
ORDER BY
        SUM(id) DESC

它的计划是:

  |--Compute Scalar(DEFINE:([Expr1004]=round([Expr1003],(-1))))
       |--Sort(ORDER BY:([Expr1003] DESC))
            |--Stream Aggregate(GROUP BY:([test].[dbo].[master].[name]) DEFINE:([Expr1003]=SUM([test].[dbo].[master].[id])))
                 |--Index Scan(OBJECT:([test].[dbo].[master].[ix_name_desc]), ORDERED BACKWARD)

如您所见,聚合完成一次并存储在其中Expr1003.

Expr1003然后在Sort运算符(处理ORDER BY)和Compute Scalar(处理ROUND)中重用