As I am primarily a developer, but as a JOAT (jack of all trades) I have been tasked with developing a demographics data cube, which I have and it works except now when data is being analyzed I am being hit with SCD issues. First I need to say that I already have read up on changing the fact table to include current date and a IsCurrent column. Except what can I do at the Cube level as I do not live in the DB world here to be changing the fact tables?Here is a sample and example and you can see with the counts the issue I have. Client Id Distinct Count All 7208 <--GOOD NUMBER Adult DD 767 Adult MI 4137 Adult SA 309 Child DD 150 Child MI 1267 Child SA 1 Unknown 778The total of 7208 for the fiscal year ending is correct and should be 7208, however if you add the values of the categories it comes up to 7409 and when I analyzed the data from the lowest client level, 201 clients did change a category sometime during the year. We do not want to count the "older" category, but only the newest based on another dimension table which is Fiscal Dates (Year, Quarter, Month attributes).This is the query I use to get the above dataselect {[Measures].[Client Id Distinct Count]} on 0, NON EMPTY {[Population].[Population Name].MEMBERS} on 1from DemographicsCubeWhere [Dates].[Fiscal Dates].[Fiscal Year].&[2015]Is there a way just to count the current data and ignore the historical data for a client who has "changed" values in a dimension and do it at the cube level. Even if it is with a calculated member I am good with that. I just need some direction.Thanks!
↧