Question

我有一个简单的事实表，其中包含AverageBalance列，TimeKey（始终为月末日期），AccountKey和AccountClassKey。

我添加了计算所得的成员，该成员将计算不同时间层次上的平均余额。

CREATE MEMBER CURRENTCUBE.[Measures].[Average Calc] AS 
    Avg(
      EXISTING([Time].[Processing Date].[Processing Date].MEMBERS) ,
      [Measures].[Avg Bal]
    ) ;

它工作正常。

但是我有一个在计算中添加以下逻辑的问题：根据AccountClassKey，应该从平均余额的其余部分中添加或减去平均余额。

因此，我尝试了类似的操作（首先我添加了“帐户类别密钥”作为衡量标准）：

CREATE MEMBER CURRENTCUBE.[Measures].[AvgBalMult] AS
[Measures].[Avg Bal] * (
       CASE 
       WHEN [Measures].[Account Class Key] = 1 THEN 1  
       WHEN [Measures].[Account Class Key] = 2 THEN -1
       else 0
       END);

但是，随着时间的推移，我在报表中总是得到零，我意识到这是因为“帐户类别”度量值首先被汇总，因此除非我查看帐户，否则它永远不会具有1或2的值。级别。

最后，我想得到这样的东西：

CREATE MEMBER CURRENTCUBE.[Measures].[Average Calc] AS 
    Avg(
      EXISTING([Time].[Processing Date].[Processing Date].MEMBERS) ,
      [Measures].[Avg Bal] * (
       CASE 
       WHEN [Measures].[Account Class Key] = 1 THEN 1  
       WHEN [Measures].[Account Class Key] = 2 THEN -1
       else 0
       END)
    ) ;

在帐户级别，它将给我“平均余额”而没有符号，但是在进行汇总时，应该与上述逻辑进行汇总。

Answer 1

您可以尝试MeasureExpression属性。它在聚合之前起作用。

https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/configure-measure-properties?view=sql-server-2017

https://social.msdn.microsoft.com/Forums/sqlserver/en-US/0ba16255-55b8-411c-a6ff-027701156f2c/ssas-measureexpression-influence-on-usage-of-aggregations?forum=sqlanalysisservices

https://blog.crossjoin.co.uk/2005/06/30/measure-expressions/

Answer 2

尝试用下面的代码替换前两个代码段。我以为“ AccountClassKey”实际上是一个维度属性，您已经为该计算进行了度量。基于此，我将其命名为“ [DimAccount]。[AccountClass]。[Account Class Key]”

CREATE MEMBER CURRENTCUBE.[Measures].[AccountClassKeyMultipiler] AS
(
CASE 
WHEN [DimAccount].[AccountClass].[Account Class Key].currentmember.Properties ("Member_Value",TYPED) = "1" THEN 1  
WHEN [DimAccount].[AccountClass].[Account Class Key].currentmember.Properties ("Member_Value",TYPED) = "2" THEN -1  
else 0
END);

CREATE MEMBER CURRENTCUBE.[Measures].[AvgBalMult] AS
[Measures].[Avg Bal] * [Measures].[AccountClassKeyMultipiler]

编辑在这里，我试图解决冒险作品中的类似情况

我们列出产品和销售月份

select 
{[Measures].[Internet Sales Amount]
}
on columns,
(
[Product].[Category].[Category],
[Date].[Month of Year].[Month of Year]
)
on rows 
from 
[Adventure Works]
where [Date].[Calendar Year].&[2013]

接下来，我们将上面的查询转换为基于月数的每种产品的平均销售额

with member 
[Measures].[Internet Sales AmountAvg]
as
Avg(EXISTING([Date].[Month of Year].[Month of Year]) ,[Measures].[Internet Sales Amount]) 
select 
{
[Measures].[Internet Sales Amount]
,
[Measures].[Internet Sales AmountAvg]
}
on columns,
(
[Product].[Category].[Category]
)
on rows 
from 
[Adventure Works]
where [Date].[Calendar Year].&[2013]

现在我们添加一个基于案例的乘数。看看最后一列

with
member [Measures].[Internet Sales AmountAvg]
as
Avg(EXISTING([Date].[Month of Year].[Month of Year]) ,[Measures].[Internet Sales Amount]) 

member [Measures].[Multipiler]
as
CASE 
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Accessory' THEN 1  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Mountain' THEN 2  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Road' THEN 3  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Touring' THEN 4 
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Components' THEN 5  
else 0
END

member [Measures].[Internet Sales AmountAvg2]
as
[Measures].[Internet Sales AmountAvg]*[Measures].[Multipiler]

select 
{
[Measures].[Internet Sales Amount]
,
[Measures].[Internet Sales AmountAvg]
,
[Measures].[Multipiler]
,
[Measures].[Internet Sales AmountAvg2]
}
on columns,
(
[Product].[Category].[Category],[Product].[Product Line].[Product Line]
)
on rows 
from 
[Adventure Works]
where [Date].[Calendar Year].&[2013]

现在让我们看看如果从行中删除[Product]。[Product Line]。[Product Line]会发生什么情况

with
member [Measures].[Internet Sales AmountAvg]
as
Avg(EXISTING([Date].[Month of Year].[Month of Year]) ,[Measures].[Internet Sales Amount]) 

member [Measures].[Multipiler]
as
CASE 
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Accessory' THEN 1  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Mountain' THEN 2  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Road' THEN 3  
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Touring' THEN 4 
WHEN [Product].[Product Line].currentmember.Properties ("Member_Value",TYPED) = 'Components' THEN 5  
else 0
END

member [Measures].[Internet Sales AmountAvg2]
as
[Measures].[Internet Sales AmountAvg]*[Measures].[Multipiler]

select 
{
[Measures].[Internet Sales Amount]
,
[Measures].[Internet Sales AmountAvg]
,
[Measures].[Multipiler]
,
[Measures].[Internet Sales AmountAvg2]
}
on columns,
(
[Product].[Category].[Category]
)
on rows 
from 
[Adventure Works]
where [Date].[Calendar Year].&[2013]

最重要的是，在这种情况下，度量的行为取决于维度属性值，您需要在一个轴上具有该维度属性。如果不存在，则使用默认值。如果逻辑中未处理默认值，则会导致错误结果

计算成员的麻烦始终等于零

2 个答案: