Expert Cube Development with Microsoft SQL Server 2008 Analysis Services [Englisch] [Taschenbuch]

If you work with SSAS - this is a MUST READ. Period.

I give it 5 stars - well written, strong content, and worth far more than the price.

A few comments:

- You will find viewpoints (valuable, credible viewpoints) in this book that you won't find in most SSAS books. That is part of what sets this book apart from other SSAS titles.

- On the word "expert" in the title: Expert books are not common - perhaps they should be considered a rare event. I would classify this as a very strong intermediate title. This is not a criticism in any way - a seasoned SSAS developer should read this book - but I view this more as a strong intermediate-advanced book.

- I'm an SSAS developer/instructor: people often ask me what SSAS books to buy. I instruct them to buy two...this book, and Teo Lachev's SSAS book. Sure, other SSAS books are good....but if you haven't purchased any yet and want to build a library, get this book and Teo's book. They are the Crème de la Crème.

Guys, great job, you have created a masterpiece. Well done.

Kevin S. Goff
Microsoft MVP
Business Intelligence Developer/Author/Instructor

First, just to explain my position.. I've had the experience of teaching hundreds of students Microsoft official courseware on Business Intelligence. I've also served as lead BI architect during my career for a number of large companies. So, I think that I have a unique perspective on the matter.

Who this book is for:
This book is for the intermediate SSAS developer/practitioner. One should not buy/read this book without having worked with Analysis Services for at least a year and gained knowledge up to an intermediate level. There are 2 books that I would recommend that one read before tackling this one: MDX Step by Step, and SQL Server Analysis Services Step by Step.. Both of those would be requisite knowledge before really understanding this book.

Now, let's review the book:

Chapter 1 & 2: Designing the Data Warehouse for Analysis Services. This was a great little introductory chapter that took the principles of Kimball and Inmon and applied them to Analysis Services. HOWEVER, this is not even close to enough to actually know how to properly design a data warehouse. In my opinion, this was the only introductory chapter in the book -- by that, I mean a chapter that someone with one year of experience would know inside and out as the lessons are fairly common place. I almost stopped reading the book after this chapter because I thought that I had another cube 101 book.

Chapter 3: Designing More Complex Dimensions. Holy cow, this chapter BLEW MY MIND! This is the only published work that I could find anywhere that properly deals with Type 2 slowly changing dimensions and cubes. The grouping and banding section is also wonderful. The parent/child hierarchies section was somewhat controversial. Several people that I have talked to took it to mean that the authors were saying don't use parent/child hierarchies. I don't think that is what they meant. I just think that the authors meant don't use them if you don't have to do so because there are several financial and HR scenarios where Parent/child hierarchies are the only things that will work.

Chapter 4: Measure and Measure Groups
This chapter starts out slow with basic beginner knowledge about Measures and Measure Groups. But then, the authors begin to shine when they talk about how to handle different dimensionality and non-aggregatable measures. Absolutely amazing discussion that I haven't seen in any publication..

Chapter 5: dealing with Drillthrough
If any chapter should have been expanded, this should have been the one. The authors threw out some eye candy, but the subject was dealt with too briefly. I'll call this the most disappointing chapter within the book because drillthrough is so important -- more should have been said about drillthrough security, perspectives and drill-through headaches, ssrs/sharepoint headaches when passing parameters as an alternative to drill-through ect.. Any data architect knows that drillthrough can make/break a successful implementation. Furthermore, the methods of doing it past the GUI are not readily obvious. A better discussion of drillthrough can be found here:
[...]

Chapter 6: MDX Calculations
If you've got MDX down, this was a basic chapter. Breeze your way through it. The calculation dimension was very nice. It's a tactic that everyone has to learn quickly after deploying a cube.

Chapter 7: Adding Currency Conversion
A nice chapter. Fairly basic, and I wouldn't consider this advanced. I think that Scott Cameron did a better job explaining this in MDX Step by Step.

Chapter 8: Query Performance Tuning
Read this chapter and every reference that the authors provide 3 times! This was an excellent piece of work. The only thing that I would have liked to see is more time spent explaining how to use the BIDS aggregation designer because I've never really found any place that properly described the art of designing aggregations(without the wizard).

Chapter 9: Security
Very good.. somewhat high level, but the authors do a good job of explaining dynamic security.

Chapter 10: Productionization
This is good. But a far better discussion on processing can be found in Brian Knight's Problem, Design, and Solution where he deals with how to automate processing on a cube -- especially when you have multiple partitions that can be updated.

Chapter 11: Monitoring
Excellent Chapter. I have used it to make checklists when I'm on sites.

Overall, I give the book 5 stars because it is the best out there when dealing with complicated subjects. This is not sufficient to become an advanced cube administrator but it will definitely make one a strong intermediate admin. I'd recommend this plus the SSAS wiki that I referred to earlier to reach an advanced cube level.

This book is truely unique among 'cube' books in that it presents the 'heart of the matter' or essential topics that one uses in actual practice.

I really appreciate its emphasis on modeling and design of the data structures (data warehoue, dimensions, facts) because it's the complexities of designing these structures that allow you to handle the difficult 'real-world' problems. It's not all that difficult to write the 'Cube 101' book but it is very difficult to write the 'Professional Cube'. Moreover, it's done in a way that is accessible or very easy to read. In short, great ideas presented with great communication skills!

Finally, the book is relatively comphrensive in scope. It handles many difficult concepts such as Many-to-Many relationships and Data Mining dimensional relationships. It addresses most of the crucial aspects of Analysis Services cubes without going over 350 pages.

I highly recommend this book to anyone who is working in the 'real world' as this book will pay for itself many times over.