February 2, 2013 at 5:17 pm
I am attempting run the following MDX query to determine the distinct amount of transaction numbers for a set including 2 dimension attributes. (POS and Concierge) When I run the query, I seem to be getting the distinct amounts for each added to together. If the same transaction number exists in both, they are both counted. However if I attempt to get the same result using an Excel Pivot table, I get the desired transaction number count. Could anyone help me with getting this result in the MDX query?
WITH
SET [Week]AS StrToMember("[Date].[Hierarchy - Fiscal].[Fiscal Week].&[2012]&[48]")
SET [Store]AS {[Financial Transaction Channel].[Transaction Channel Desc].&[POS],
[Financial Transaction Channel].[Transaction Channel Desc].&[CONCIERGE]}
MEMBER [Store Transactions With Returns] AS SUM([Week] * [Store], [Measures].[Operational Transactions Count])
SELECT
{
[Store Transactions With Returns]
} ON COLUMNS,
{
[Week]
} ON ROWS
FROM EDW
Results: Store Transactions With Returns
Wk. 48 - 2012 78,636
The same attempt using an Excel Pivot table yields this:
Row LabelsOperational Transactions Count
Year 2012 78,368
Half 2 - 2012 78,368
Qtr. 4 - 2012 78,368
December - 2012 78,368
Wk. 48 - 2012 78,368 Result of addition: 78,636
1 US STORES 78,368
CONCIERGE 3,668
POS 74,968
Grand Total 78,368
If you add the POS and Concierge amounts, you get 78,636. Which is the same result I am getting from my MDX query, but after I went through the transactions line by line, I was able to determine that 78,368 is the correct amount of distinct transaction numbers.
Any help on this would be greatly appreciated.
February 8, 2013 at 5:38 am
Try this and let me know if it works
SELECT
{[Measures].[Operational Transactions Count]} ON COLUMNS,
{[Date].[Hierarchy - Fiscal].[Fiscal Week].&[2012]&[48]} ON ROWS
FROM
(SELECT {[Financial Transaction Channel].[Transaction Channel Desc].&[POS],
[Financial Transaction Channel].[Transaction Channel Desc].&[CONCIERGE]} ON COLUMNS
FROM EDW)
Mack
February 8, 2013 at 7:20 am
I ended up creating a new dataset and basically using what you describe above. When the dimension attributes are placed in the WHERE clause, it works.
Thanks for your feedback.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply