Reporting Services, MDX and Aggregated Values

Undoubtedly the best time to find out about an undocumented change in functionality in a product is halfway through a demo of said functionality to a large group of people. This happened to me last week: I was teaching my MDX course (written on the last CTP of SP2) and had just finished my speech on all the rubbish aspects of Reporting Services/Analysis Services integration and was trying to show how Reporting Services automatically filtered out all but the lowest level of granularity of data from an MDX query (see Teo Lachev’s post here: http://prologika.com/CS/blogs/blog/archive/2006/02/08/853.aspx and Reed Jacobsen’s posts here: http://sqljunkies.com/WebLog/hitachiconsulting/archive/2006/08/07/22359.aspx and here: http://sqljunkies.com/WebLog/hitachiconsulting/archive/2006/08/04/22346.aspx for details) when I found that it wasn’t doing it any more. Later on I emailed Teo to ask if he knew anything about this change – he didn’t but he asked the RS dev team and they gave him some details (he then blogged about it here: http://prologika.com/CS/blogs/blog/archive/2007/06/10/aggregate-rows-and-sp2.aspx). So, a small victory for the people – Reporting Services now no longer tries to force its own aggregation functionality on you and you always see the full results of your query, unless you’re already using the RS Aggregate function in your reports. This is apparently in response to customer demand. Just be sure to check any existing RS/AS reports you’ve got in production to make sure they’re not displaying extra rows now!

I’m now hopeful that the message will get through about the other stupid restrictions that RS places on AS data sources, such as only being able to put the measures dimension on columns. I’ve not kept up with the changes in RS2008 as much as I should (Teo again has a good overview of what’s coming here: http://prologika.com/CS/blogs/blog/archive/2007/06/09/teched-2007-us-memoirs.aspx) but I’ve not heard that these restrictions will be lifted. The problem is of course that RS expects to do all the aggregation of data itself, but the RS dev team don’t seem to understand that if I’m using AS then I will have designed all my aggregation business logic into my cubes and dimensions and I couldn’t care less about what RS can do in this area (for an example of their mindset, see the somewhat patronising comments on this posting on Connect from last year: http://connect.microsoft.com/SQLServer/feedback/ViewFeedback.aspx?FeedbackID=125414).

While we’re talking about Katmai I might as well mention why I’ve not blogged about the new AS features in it – it’s because there aren’t all that many. You’ve probably already read Mosha’s post about the attribute relationship designer (http://sqlblog.com/blogs/mosha/archive/2007/06/07/katmai-june-ctp-attribute-relationships-tab.aspx); Vidas Matelis has covered the other cube design wizard changes here: http://www.ssas-info.com/VidasMatelisBlog/?p=22#more-22. More features will come in later CTPS but overall it’s really going to be all about performance improvements and manageability, and while I know a bit about some of the planned features it’s too early to blog about them because it’s too early to say what’s going to actually get included. Suffice to say that for the serious AS developer there are going to be some very welcome improvements but there won’t be anything that excites the marketing people.