After 3 months of trying to learn MDX, I can now see why some people refer to it as Mind Destroying eXpressions! In my 20 years of programming (school, university, work), I have successfully learnt some tough languages (Assembler, Eiffel, Prolog, C++) but, for me, none have been as incomprehensible as MDX.
My market reasearch of MDX demand here in the UK (using two very respectable job sites) showed negligible demand for MDX. I have therefore decided not to pour any more time into MDX for our DW/BI solution. I reasoned that the job market confirms my conclusion about the applicablility of MDX. We have been using SQL and "Utility" tables very effectively to add value to the "raw" data in our data warehouse and delivered lots of successful BI to our business.
So, my specific questions are:
1.What will we be missing out on by not using MDX?
2.Even if the SQL is thougher to write (which hasn't been so far), I would like to know for completness what MDX gives you that you can not do in your ETL or Data Warehouse (using SQL)?
3.Is there a tool that helps generate MDX for you? At the moment, I am relying on Excel (connected to our Cube): I create the report in a work book and then use the
OLAP PivotTable Extension tool to capture the MDX Excel is sending to SSAS. The only problem is that the MDX output in Management Studio looks completely different to that in Excel.
Amir