What issues can creating 200+ views and avoiding MDX - when there are only 2 true fact tables – cause down the line?
How would you convince your client that their current path is not in their best interests?
I’ve inherited an SSAS 2012 development from a previous contractor.
The cube models a sophisticated patient administration system; one would expect complex measures in healthcare.
Documentation: non-existent, Kimball nor Inmon are nowhere to be found, 200+ measure groups confuse the situation, next to no aggregations or partitions, and host of other elements.
Most worrying is that the prior contractor did not write a single line of MDX by employing the following strategy:
Create 200+ views based on two fact tables in SQL Server.
Ensure each view contains filters to drive a measure.
Add all 200 views are to the cube as a new data source view, and sum or count a column from said view.
Despite spending 3 weeks trying to explain the issues to business and technical users with no multidimensional, data warehousing experience, the client believes there is nothing wrong with the current cube. Never had this challenge before!
Have you come across a situation like this before?
Looking forward to the responses!