Maximum number of dimensions 2

[sbharamreddi]

At my job site, I am getting a proposal to build a cube with about 250 dimensions. The maximum number dimensions I have had in a cube is about 60. I was wondering if there is a limitation on the number of dimensions I can have in a cube in Analysis Manager. Any comments or experience related my question is appreciated.

Thanks

 

[johnherman]

As an OLAP'er and DW professional, I don't think you really have 250 real dimensions. It seems way too big. I imagine you are somehow creating artificial dimensions to take advantage of certain features of the product. I am not aware of any limitation in the product. That's gonna be one honkin' big cube - that might be where you have a problem.

-------------------------
The trouble with doing something right the first time is that nobody appreciates how difficult it was - Steven Wright

 

[sbharamreddi]

Thanks John,
I actually got this weird request from a user group which said they could possibly be analyzing about 250 data elements around one measure. I still need to see what those are and I have no idea right now. Before I could proceed I just wanted to check to see if there were any limitations technically. Thanks for your help.

 

[MDXer]

I'm with John on the number of dimensions that you actually have. A single dimension can Play the role of multiple dimensions. Date for example. You can have 3 dates in your cube but it only requires a single dimension this dimension Plays the Role of all 3.

You users may also not be taking hierarchies into account. I have had beople say the need dimensions for COuntries, States and Cities they believed this would be done in 3 dimensions, when actually it is a Geographic dimension with a hierarchy.

Also remember that even if they have 3 source system each with a Customer table doesn't mean they need 3 Customer dimensions. If the data spans multiple systems then look at conforming your dimensions.

Besides that with 250 dimensions when you get more than a couple dozen involved my experience is the users get lost. You may also want to look at breaking the data out into more specific cubes.

Paul
---------------------------------------
Shoot Me! Shoot Me NOW!!!
- Daffy Duck