Day 1: 1:58am – Eureka
I went back to the drawing board with my dimensions, as they were the ones causing me all the trouble.
To be more specific, I took out the elephant gun and started them from scratch.
Because I’m working from an OLTP datasource, and not a data warehouse that might be a bit cleaner, I was getting into trouble. Trouble to the tune of missing values, duplicate values, and the like, in my dimension attributes.
So, I wrote some quick and dirty TSQL UPDATE statements against my base OLTP tables, and then wrote some views in order to accomplish a bit of horizontal and vertical filtering.
Back in SSDT, I defined some hierarchies, and processed the dimensions one at a time, so that I could review the errors and warnings as they came up, and not be crushed by 29 errors at once.
I was rewarded for my patience:
Those three words – “Deployment Completed Successfully” – were music to my ears.
I flipped over to the Cube Browser tab in my newly-minted Cube and started to experiment. The Cube Browser is where the whole OLAP/Analysis Services experience really started to come to life. It was exciting to see the hierarchies and measures in action – a visual confirmation that my data was starting to take its proper shape. I was able to drag and drop the Dimension attributes and Measures quickly around the interface and was giddy at how quickly the results were updated. The filtering was straight-forward and intuitive, and allowed me to quickly spot trends in the data.
I got through 200 pages of my Wrox book, and managed to get a very rudimentary cube up and running today, so I’ll sleep well tonight!