计算度量中的MDX LAG命令返回错误值

Question

I have following query

 with member priorSellOut as ([Date].[Year Month].CurrentMember.LAG(1),MEASURES.[Sell Out]) Select { MEASURES.[Sell Out] ,PriorSellOut } ON COLUMNS, [Date].[Year Month].[Year Month].Members ON ROWS From Sales 

the result of this query is:

  Sell Out Prior Sell Out 2018.01 34329 (null) 2018.02 37752 34329 2018.03 21798 37752 2018.04 41477 21798 2018.05 50125 41477 2018.06 6363 50125 2018.07 11511 6363 2018.08 7444 11511 2018.09 13989 7444 2018.10 1936 13989 

I want to have last 3 month ago. and use [3 Month Ago] named set.following query:

 with member priorSellOut as ([Date].[Year Month].CurrentMember.LAG(1),MEASURES.[Sell Out]) Select { MEASURES.[Sell Out] ,PriorSellOut } ON COLUMNS, [Date].[Year Month].[Year Month].Members ON ROWS From Sales Where ([3 Month Ago]) 

the result of this query is:

  Sell Out Prior Sell Out 2018.08 7444 (null) 2018.09 13989 7444 2018.10 1936 13989 

I want to have following result. Prior Sell Out Column for 2018.08 must have value.

  Sell Out Prior Sell Out 2018.08 7444 11511 2018.09 13989 7444 2018.10 1936 13989 

Thanks in advance

Answer 1

Your problem is that instead of filtering on the list of rows to display, you instead filter out all data that is not in the last 3 months.

To filter out the rows - just apply a limited set expression to the axis. We want the last 3 months to be displayed, so let's take the last three members of the months list like so: TAIL([Date].[Year Month].[Year Month].Members,3) . Now we can safely remove the WHERE clause. The result is:

with member priorSellOut as
  ([Date].[Year Month].CurrentMember.LAG(1),MEASURES.[Sell Out])
Select 
  {
      MEASURES.[Sell Out]
      ,PriorSellOut
  } ON COLUMNS,
  TAIL([Date].[Year Month].[Year Month].Members,3) ON ROWS
From Sales

NOTE: This assumes that the [Date].[Year Month].[Year Month] dimension IS PROPERLY ORDERED

EDIT: If you really necessary need to use the [3 Month Ago] named set - just put it on the axis instead of the TAIL(...) formula I used.