Page 1 of 1

Creating a slimmed down version of a massive cube

Posted: Tue Jun 18, 2013 4:14 pm
by ChrisF79
We have a sales cube with thousands of items, customers, etc. I want to create a slimmed down version of this cube (only 4 dimensions instead of 11, and only consolidated elements for the item numbers, customers, etc) for reporting purposes and thought about making a rule. Since our various distributors have different criteria on how their sales are reported, simply using a "one size fits all rule" won't work so I'd be writing a rule for each distributor. The thing is, the rule takes a bit to calculate. What's a best practice here? Since I'm only including prior year and current year, should I just use a TI process or is a rule truly the way to go? Any advice?

Re: Creating a slimmed down version of a massive cube

Posted: Tue Jun 18, 2013 4:53 pm
by jim wood
I would create the cube then have 2 process. One to update the dimensions and one to copy the data. Then run this after each major data run. For planning data entered during the day run an end of day process. It's a fall back option when the rules are to slow.

Re: Creating a slimmed down version of a massive cube

Posted: Tue Jun 18, 2013 5:35 pm
by ChrisF79
Thanks Jim. That's what I'll do then.