This content has been marked as final.
Show 2 replies

1. Re: Cube Union
wendress Aug 31, 2012 11:20 AM (in response to user7853353)Can you provide some insight into why you are designing this as a two cube solution? If the cubes share the same dimensions why are all the measure not in the same cube? (In most cases, two or more cubes would be created if the stored measures in each cube have different sparsity characteristics, are loaded from different tables on different schedules, have different aggregation operators, etc. This doesn't seem too be the case in your application). 
2. Re: Cube Union
user7853353 Aug 31, 2012 12:15 PM (in response to wendress)I have a dimension "symbol" where my facts for some symbols are loaded with a preadjustment, while my facts for other symbols are loaded without an adjustment.
For the facts which have been loaded with an adjustment a calculation must be applied to them before they can be viewed by the user. The input parameter for the calculation changes every few months i.e. the calculation is dynamic.
Only a certain subset of symbols will actually require this calculation to be applied to them. So the idea was to have two separate cubes and bring them together via a union.
I suppose the alternative here is to convert each measure in a single cube to a calculated measure and put in a simple condition statement such as:
if symbol_attribute eq 'gets_calc'
then return perform_calc(cube_value)
else return cube_value
I would than have to modify my fact table view to union the sources of the data (adjusted symbols and non adjusted symbols)
Possibly there is a cleaner way to this?