Question

我希望过滤以下结果集，以便仅显示维度A值1的计数为1的结果，而不考虑维度A值2的计数值

           Dimension A Value 1     Dimension A Value 2
Entity ID   Count                Count
11              1   
78          1   
90          1   
101         1   
114                                1
118         1   
125         1   
134                                    1
140         1   
161         1   
169         1   
186         1                  2

过滤后的集合看起来像

    Dimension A Value 1 Dimension A Value 2
Entity ID   Count               Count
11           1  
78           1  
90           1  
101          1  
118          1  
125          1  
140          1  
161          1  
169          1  
186          1                 2

mdx是

WITH
SET [~COLUMNS] AS
    {[Dimension A].[Dimension A].[Value 1], [Dimension A].[Dimension A].[Value 2]}
SET [~ROWS] AS
    {[Entity].[Entity].[Entity ID].Members}
SELECT
NON EMPTY CrossJoin([~COLUMNS], {[Measures].[Count]}) ON COLUMNS,
NON EMPTY [~ROWS] ON ROWS
FROM [My Cube]

我一直在使用Filter和NonEmpty，但是我是MDX的新手，我的sql大脑受伤了。我想这对于拥有许多MDX的人来说可能是微不足道的，但是我失败了。轻一点，这是我的第一个问题

Answer 1

您的查询应类似于

Select ([Dimension A].[AttributeHierarchy1].[AttributeHierarchy1],{[Measures].[Value1],[Measures].[Value2]}) on columns,

filter([Dimension B].[EntityID].[EntityID],[Measures].[Value1]=0)
on rows 
from yourcube

但是可能会有问题。例如，您的dimesnion具有两个值A和B，对于特定行A，value1 = 1但B，Value1 = 0，该行将显示为A对其进行质化，并且B被继承。

编辑

让我们以一个例子为例，我希望看到瓶罐和笼子的互联网销售收入超过150美元

我的初始查询

select 
([Product].[Subcategory].[Subcategory],[Measures].[Internet Sales Amount]
)
on columns,
[Customer].[City].[City]
on rows 
from 
[Adventure Works]

结果

现在修改查询

select 
([Product].[Subcategory].[Subcategory],[Measures].[Internet Sales Amount]
)
on columns,

filter
(
[Customer].[City].[City], [Measures].[Internet Sales Amount]>150
)
on rows 
from 
(select [Product].[Subcategory].&[28] on 0 from [Adventure Works])

结果

Answer 2

您可以尝试使用HAVING子句：

WITH
SET [~COLUMNS] AS
    {
      [Dimension A].[Dimension A].[Value 1], 
      [Dimension A].[Dimension A].[Value 2]
    } 
MEMBER [Measures].[CountValue1] AS   //<<<<this is new <<<<<<<<<<<<<<<<<<<<<<<<<<<<<
    (
       [Measures].[Count],
       [Dimension A].[Dimension A].[Value 1]
    )
SELECT
NON EMPTY 
  CrossJoin(
    [~COLUMNS]
  , {[Measures].[Count]}
  ) ON COLUMNS,
NON EMPTY 
  [Entity].[Entity].[Entity ID].MEMBERS 
  HAVING [Measures].[CountValue1] = 1    //<<CHANGED TO NEW MEASURE
  ON ROWS
FROM [My Cube];

如果您可以使用HAVING而不是FILTER，则可能会看到性能的提高-特别是当您的脚本变得更加复杂时：
https://blog.crossjoin.co.uk/2006/01/04/the-having-clause/

请完成以下较慢的FILTER版本：

WITH
SET [~COLUMNS] AS
    {
      [Dimension A].[Dimension A].[Value 1], 
      [Dimension A].[Dimension A].[Value 2]
    } 
//>>>>>> following is new >>>>>>>>>>>>>>>>>>>>>
MEMBER [Measures].[CountValueNEW] AS   
    (
       [Measures].[Count],
       [Dimension A].[Dimension A].[Value 1]
    )
SELECT
NON EMPTY 
    [~COLUMNS]
   *{[Measures].[Count]}
ON 0,
NON EMPTY 
FILTER(
  [Entity].[Entity].[Entity ID].MEMBERS,
  [Measures].[CountValueNEW] = 1 
)  
ON 1
FROM [My Cube];

我需要过滤此MDX结果集

2 个答案: