THE SQL Server Blog Spot on the Web

Welcome to - The SQL Server blog spot on the web Sign in | |
in Search

SQLBI - Marco Russo

Please visit the same blog at - comments are now disabled here but you can comment on the new blog (all blog posts available here are available there, too).
You can follow me on Twitter: @marcorus

TechEd North America 2012-Day 4 #msTechEd #teched

I hadn’t time yesterday to write a blog post before of the beginning of the day and I recover now that I’m already back to Europe. I spent many hours at the Microsoft booth meeting people who asked questions about Tabular, Power View and PowerPivot. I have to say that I’m really glad so many people started using PowerPivot and of the large interest around Tabular.

During my BISM: Multidimensional vs. Tabular I’ve been helped by Alberto during the demo (we joked about he was the demo monkey) and the feedback received are good. I tried to compare the strength and weakness of the two modeling options without spending time describing the area in which they are similar. During the days before I had many discussions about scenarios based on snapshots and I added a few slides to the presentation in order to cover this area that I thought was marginal but seems to be a very common one. To know more, watch the presentation when it will be available or wait for some article I’ll write in the future describing these patterns.

In less than 10 days, I’ll be at TechEd Europe and I’m really looking forward to meet other Tabular developers and PowerPivot users there!

Published Friday, June 15, 2012 7:21 PM by Marco Russo (SQLBI)
Filed under: , ,



M2M user said:

Very informative and well explained.

At a certain point, you said you don't expect MS to implement M2M in the model, but more likely provide a wizard to write the standard DAX code required.

Why shouldn't they go the way they did with SSAS 2005?

Or even like some in-memory competitors (like QlikView 9+) where filters are propagated to every table, even along M2M relations.

June 16, 2012 12:23 PM

Marco Russo (SQLBI) said:

The reason is side effects. Looking at technology today, the easiest way to implement M2M in data model is automating the cross-filter propagation described here:

However, my experience with M2M in UDM says that as soon as you have multiple M2M in your data model, you might end up creating multiple role dimension just because there are calculations in which you want to use M2M and other in which you don't. This is expecially true whenever there are multiple M2M paths connecting two dimensions.

I already proposed a simpler DAX syntax to handle cascading M2M because of that. But I understand the pain of having to create multiple formulas for any M2M measure. Anyway, for that I would prefer to see an approach that also solves the DateTool problem (YTD and other formulas applied to multiple measures) that could be easily applied to the M2M too, without moving M2M to the model.

Anyway, it's a question of priorities. Solving DateTool is more important than M2M, in my opinion.

June 17, 2012 6:43 AM

Pranav said:

I went thru the session recording & it was very informative.

Will it be possible for you to post the presentation used for Tabular versus Multi-Dimensional TechEd session in some location - Channel9 / MSDN link is not working for slides. It only works for session recording.

June 20, 2012 6:16 AM

Marco Russo (SQLBI) said:

For other downloads you have to register and login to, then look in the catalog of the sessions and opening the breakout session desired you can also download the slides.

June 20, 2012 7:21 AM

Pranav said:

Thanks for the guidance - I was able to download the session material from TechEd site :-)

June 21, 2012 7:46 AM

Pranav said:

Hi Marco,

We are working on a Tabular Model Solution that implements two M:M relationships & have been encountering performance issues in DAX measures.

While reviewing our solution with SQL SMEs, we have got recommendation that we should move to Multidimensional Model to improve performance. As per them, M:M implementation has better performance in Multidimensional compared to Tabular.

However, while going thru TechEd DB413 Session (Many to Many in DAX) Presentation, in conclusion it mentions that "M:M is Very fast in Tabular when compared with Multidimensional"

Will it be possible for you to share some additional reference points that can be used to make decision whether to continue with Tabular Model or move to Multidimensional Model to improve performance of M:M implementation?



August 1, 2012 6:30 AM

Marco Russo (SQLBI) said:


I can grant you that a proper model and DAX formulas in Tabular are faster than Multidimensional. Orders of magnitude.

Please write me at to discuss your case.


August 2, 2012 1:29 AM
New Comments to this post are disabled

About Marco Russo (SQLBI)

Marco Russo is a consultant, writer and trainer specialized in Business Intelligence with Microsoft technologies. He runs the SQLBI.COM website, which is dedicated to distribute resources useful for BI developers, like Integration Services components, Analysis Services models, tools, technical information and so on. Marco is certified as MCT, MCDBA, MCSD.NET, MCSA, MCSE+I.

This Blog



Privacy Statement