I have put together a cube for auto loan analysis. It is populated via monthly snapshots. There are many actions taken on these loans that have a date stamp associated with them (repossess/resell/refinance, etc.). These fields are on each row and only populated
if utilized, else NULL.
In the design of the cube, all of these fields were specified as being associated with date dimensions. Now that I am actually building this thing out, my date dimension is aliased (role playing) 15 times! This seems ridiculous to me as many of these date
fields are not useful for analysis, per se. Primarily the analysts need the ability to filter based off of them. Ie. Display measures [x],[x1] in a pivot table where reposession date > '[Jan 2005], etc.
They use excel 2007 for analysis. Is there another way to handle all of these auxiliary/sparsely populated dates besides creating dimensions for them?
Thanks for your help!
View Complete Post