Hi,

I am trying to design a cube and have hit a snag with the design.

I have a facttable that contains transactions and it is linked to a few dimensions, date, trantype etc.

I know need to model in groupings of these transactions. 

Problem is for each group, I need want to take the earliest date for the underlying transactions and use that as the date for the group.

So this ruled out using a fact dimension.

If I add a new dimension table with the group key and start date, would I be able to query it using the dimension start date instead of the fact table date and retrieve the measures from the fact table or would I need to create a new fact table for the groups with its own measures.

Hopefully this makes some sense.

Thanks

Hi J,

   My idea is to create another column in say Date (or Time) Dimension and name it as GroupStartDate and populate it with Earliest date. I mean for different dates already exist in Date Dimension there is only one start date for particular period.

And create another Dimension with EarliestDate as key and GroupName. And in the Dimension usage add this dimension as referenced Dimension.

Hope this helps

Regards,