Hi,I'm very new to SSAS, and cubes in general, and have a question. We are working on the early stages of a reporting solution for Dynamics AX, and a key piece of that is allowing certain users to connect to cubes from Excel and create their own analyses. I currently have a product dimension that has 10-20 attributes in it, which is used in the sales cube. So far everything has gone smoothly. Now I have a requirement to allow for reporting against product master data from within Excel, meaning the user would like to be able to include anything from the product master record in their analyses. This is a much larger set of columns... probably at least 100+.I'm wondering what the best practice would be for handling this requirement in a cube. On the Oracle side of our business, we don't use OLAP, so simply enhancing the product dimension to contain all the attributes would be an easy choice. But I'm curious if I'm setting us up for future performance issues doing this in a cube?My initial ideas were:1.) Enhance the product dimension to add all necessary attributes. Since it exists in the sales cube and elsewhere, I'm questioning if space and performance could be an issue.-or-2.) Create a secondary product master dimension to house the attributes only needed for simple master data reporting. That could then be put in a new master data cube. My thinking there is that we then don't have all of the "extra" columns in other cubes like Sales which will likely become large over time. But doing this is creating a silo that could cause issues in the future when a user wants an attribute from the secondary dimension in a cube that only has the primary dimension.I'd lean towards #1 assuming performance would be adequate.Is there a best practice around how you experts would handle this requirement?Am I right to assume that every additional column I add to the dimension will increase the size of and slow down the processing time of a cube?I appreciate any advice you can give.Thanks,Tony
↧
