SQL Server 2008 R2中的查询优化

Question

我知道这个问题经常在这里被问到，但我有一些不同的要求。

我想得到这个输出。我的逻辑也是如此，但这可以通过更简单和优化的方式完成。更快地运行此查询。

这是我的问题：

select 
    ISNULL('PortFolio Code: '+a.CPORTFOLIOCODE,'Grand Total') as [PortFolio Code],
    COUNT(SZCUSTOMERNO) as [Accounts],
    CAST(COUNT(SZCUSTOMERNO) * 100 / (Select COUNT(SZCUSTOMERNO) 
                                      from dbo.COL_TRN_AGREEMENT) 
                                      as nvarchar(50))+' %' as [%],
    sum(case when a.SZBUCKETCODE =1 then a.FOSAMT else 0 end ) as [Bucket :1],
    sum(case when a.SZBUCKETCODE =2 then a.FOSAMT else 0 end ) as [Bucket :2],
    sum(case when a.SZBUCKETCODE =3 then a.FOSAMT else 0 end ) as [Bucket :3],
    sum(a.FOSAMT) as [All Buckets] 
from 
    dbo.COL_TRN_AGREEMENT a 
group by 
    a.CPORTFOLIOCODE with rollup

我收到了这个输出

这可以使用简单的逻辑以更快的执行来完成，或者这是最简单的方法。

Answer 1

这是一种方法，可能不会更快，但您仍然可以选择进行比较。

declare @cnt int
Select @cnt = COUNT(SZCUSTOMERNO) from dbo.COL_TRN_AGREEMENT

;with 
matrix (SZBUCKETCODE,Bucket_1,Bucket_2,Bucket_3) as 
       (select 1,1,0,0 union
        select 2,0,1,0 union
        select 3,0,0,1)
select isnull('PortFolio Code: '+a.CPORTFOLIOCODE,'Grand Total') as [PortFolio Code],
       COUNT(SZCUSTOMERNO) as [Accounts],
       cast(COUNT(SZCUSTOMERNO)*100/@cnt as nvarchar(50))+' %' as [%],
       sum(a.FOSAMT*m.Bucket_1) as [Bucket :1],
       sum(a.FOSAMT*m.Bucket_2) as [Bucket :2],
       sum(a.FOSAMT*m.Bucket_3) as [Bucket :3],
       sum(a.FOSAMT) as [All Buckets] 
  from dbo.COL_TRN_AGREEMENT a 
  join matrix m
    on m.SZBUCKETCODE = a.SZBUCKETCODE 
 group by a.CPORTFOLIOCODE with rollup