Hi
I would like some advice on how best to model the following scenario with only creating one dimension.
I have the following denormalised Dimension
Product > Product Franchise > Product Franchise Group
The Franchise Group ID and Desc exists on all three of the tables.
I have two fact tables:
FactActualSales is at the Product level which has a join to the Product
table
FactBudgetSales is at the Product Franchise Group level with a join straight to the Product Franchise Group table
I have created
[list]
a context for each of the fact tables
a dimension in the universe called Franchise Group that is based on the Franchise Group Desc column from the Product Franchise Group table
[/list]
As expected when I create a query with the actual sales amount and Franchise Group objects the SQL generated joins through the denormalised dim tables up to the Product>Product Franchise>Product Franchise Group.
And the budget query joins directly to the Product Franchise Group table.
My question is (finally) is there a way that I can optimise the Actual Sales query to pick up the Franchise Group Desc column from the Product table without having to join through the snowflaked tables? The goal is to only have one Franchise Group Dimension object in the universe.
I have been trying to implement this using the aggregate aware function on the Franchise Group Desc object but can not get it to work as expected.
Any suggestions would be greatly appreciated.
Thanks
Kevin
keviwilso (BOB member since 2010-07-01)