Filter to show a range of TOP members of a dimension

Question

Hi there,
I am trying to create a TOP Filter that will display the following Values:
TOP 25, TOP 50 and TOP 100
I have a Location Dimension who has a attribute called Rank. This Rank is based on a column of the Location table which has a fixed value per location that was assigned manually in the database.
So when I select the TOP 25 member of the TOP Filter, I will be showing the top 25 locations based on Rank column.
What I did was sorting the Locations in the dimension using the Rank attribute and then building dynamics sets using TOPCOUNT mdx expression. The problem is the performance is very bad. 
Could anyone advice me on how would be the best way to achieve the above with a good performance?
Thanks and best regards,
Joss

Answer 1

Hi Joss,
your issue became interest for me and i have done some tests. i have created dimension table and filled by 5 000 000 elements having the different ranks, created dimension that contains three attributes: ID, Rank and Name. Then i defined calculation in cube
script:
CREATE MEMBER CURRENTCUBE.[Measures].[Rank_1]
AS [Product].[Rank].member_value, VISIBLE = 1;
CREATE DYNAMIC SET [Rank_25_1]
AS TOPCOUNT([Product].[Product].[Product],25,[Measures].[Rank_1]);
and run query 
WITH MEMBER [NewMember] as NULL
SELECT
[Measures].[NewMember] on 0 ,[Rank_25_1] on 1
FROM [Cube]
The command duration is about 22 sec. Then created real Measure [Rank_2] based on dimension attribute and add new dynamic set:
CREATE DYNAMIC SET [Rank_25_2]
AS TOPCOUNT([Product].[Product].[Product],25,[Measures].[Rank_2]);
the below query was running about 8 sec.
WITH MEMBER [NewMember] as NULL
SELECT
[Measures].[NewMember] on 0 ,[Rank_25_2] on 1
FROM [Cube]
So, performance is not so bad. I suppose you used in query some complex calculated members or used a lot of dimensions in axises.
Best regards.