Question

以下是我的MDX查询以生成老化报告。我想使用名为SET的名字在一个计算的措施内。我收到了这个错误：

Query（3,1）该函数需要1的元组集表达式论点。使用了字符串或数字表达式。

这可以解决吗？

WITH 
  SET [Cnt] AS 
    {'FILTER(
          [Cheque Detail Fact Keys].[Cheque Master ID].[Cheque Master ID]
         ,[Measures].[Paid Amt]<>0
       )'
    } 
  SET [x] AS 
    {
      ClosingPeriod
      (
        [Cal Date].[Month].[Month]
       ,[Cal Date].[Month].[All]
      )
    } 
  MEMBER [Measures].[0-30] AS 
    Sum
    (
      [x].Item(0).Lag(1) : [x].Item(0).Lag(0)
     ,Count(Cnt) //[Measures].[Master Count] 
    ) 
  MEMBER [Measures].[31-60] AS 
    Sum
    (
      [x].Item(0).Lag(2) : [x].Item(0).Lag(1)
     ,Count(Cnt) //[Measures].[Master Count] 
    ) 
  MEMBER [Measures].[>60] AS 
    Sum
    (
      NULL : [x].Item(0).Lag(4)
     ,Count(Cnt)  //[Measures].[Master Count] 
    ) 
SELECT 
  {
    [Measures].[0-30]
   ,[Measures].[31-60]
   ,[Measures].[>60]
  } ON 0
 ,{[Customer].[Name].[Name].ALLMEMBERS} ON 1
FROM [My Cube];

Answer 1

我认为这只是你脚本的第一部分。你为什么用字符串？尝试删除叛逆者，只需添加.MEMBERS以获得清晰度：

WITH 
  SET [Cnt] AS 
    {FILTER(
          [Cheque Detail Fact Keys].[Cheque Master ID].[Cheque Master ID].MEMBERS
         ,[Measures].[Paid Amt]<>0
       )
    } 
   ...
   ...

错误信息似乎有点神秘，但我相信花括号本身就是一个功能 - 它们将其中的任何成员转换为一个集合。所以你已经有效地写了这个{<some string>}，但是花括号期待的东西可以是一个集合。

运行以下AdvWrks脚本时出现相同的错误消息：

WITH 
  SET [Cnt] AS 
    {'FILTER(
          [Date].[Date].[Date].MEMBERS
         ,[Measures].[Internet Sales Amount] > 10000
       )'
    } 
SELECT 
  [Measures].[Internet Sales Amount] ON 0
 ,[Cnt] ON 1
FROM [Adventure Works];

要在以下度量中使用Cnt计数，您可以创建一个度量但我不认为这将是上下文感知的，因为该集合在其他任何事项之前被评估：

WITH 
  SET [Cnt] AS 
    {FILTER(
          [Cheque Detail Fact Keys].[Cheque Master ID].[Cheque Master ID]
         ,[Measures].[Paid Amt]<>0
       )
    } 
  MEMBER [Measures].[Master Count]  AS 
    [Cnt].count
  SET [x] AS 
    {
      ClosingPeriod
      (
        [Cal Date].[Month].[Month]
       ,[Cal Date].[Month].[All]
      )
    } 
  MEMBER [Measures].[0-30] AS 
    Sum
    (
      [x].Item(0).Lag(1) : [x].Item(0).Lag(0)
     ,[Measures].[Master Count] 
    ) 
  MEMBER [Measures].[31-60] AS 
    Sum
    (
      [x].Item(0).Lag(2) : [x].Item(0).Lag(1)
     ,[Measures].[Master Count] 
    ) 
  MEMBER [Measures].[>60] AS 
    Sum
    (
      NULL : [x].Item(0).Lag(4)
     ,[Measures].[Master Count] 
    ) 
SELECT 
  {
    [Measures].[0-30]
   ,[Measures].[31-60]
   ,[Measures].[>60]
  } ON 0
 ,{[Customer].[Name].[Name].ALLMEMBERS} ON 1
FROM [My Cube];

Answer 2

问题是您正在尝试将字符串传递给集合。您需要首先使用StrToSet函数将字符串转换为集合。另外正如@whytheq所说，你不需要大括号。

尝试替换

WITH 
  SET [Cnt] AS 
    {'FILTER(
          [Cheque Detail Fact Keys].[Cheque Master ID].[Cheque Master ID]
         ,[Measures].[Paid Amt]<>0
       )'
    }

用这个 - ＆gt;

WITH 
  SET [Cnt] AS 
    StrToSet('FILTER(
          [Cheque Detail Fact Keys].[Cheque Master ID].[Cheque Master ID]
         ,[Measures].[Paid Amt]<>0
       )'
     )

使用COUNT功能的正确方法是<<SomeSet>>.COUNT，不 COUNT(<<SomeSet>>)

Answer 3

[Cnt]集看起来很贵......是基于事实表吗？如果是这样，粒度是什么？

也许您可以在DSV中为度量值组的基表添加计算字段（例如，IIF（付费Amt＆lt;＆gt; 0,1,0）），然后在该字段上创建基于SUM的度量。然后你的查询变成......

WITH  
  SET [x] AS 
    {
      ClosingPeriod
      (
        [Cal Date].[Month].[Month]
       ,[Cal Date].[Month].[All]
      )
    } 
  MEMBER [Measures].[0-30] AS 
    Sum
    (
      [x].Item(0).Lag(1) : [x].Item(0).Lag(0)
     ,[Measures].[NonZero Check Count] 
    ) 
  MEMBER [Measures].[31-60] AS 
    Sum
    (
      [x].Item(0).Lag(2) : [x].Item(0).Lag(1)
     ,[Measures].[NonZero Check Count] 
    ) 
  MEMBER [Measures].[>60] AS 
    Sum
    (
      NULL : [x].Item(0).Lag(4)
     ,[Measures].[NonZero Check Count] 
    ) 
SELECT 
  {
    [Measures].[0-30]
   ,[Measures].[31-60]
   ,[Measures].[>60]
  } ON 0
 ,{[Customer].[Name].[Name].ALLMEMBERS} ON 1
FROM [My Cube];

老化报告MDX查询问题？

3 个答案: