It's very appalling to see writers that claim to have worked on the product but could not give clear and accurate examples of any topics discussed. Waste of money from my perspective.

I have thousands of hours of experience with SSAS 2000 and 2005. I think that SSAS is one of the best (if unappreciated) Microsoft products.

I bought this book for one reason - it was the first one available. If you need one now, then get it. But be warned that it's not particularly good.

The key problem is that the authors are fascinated with XML. They use raw XML to explain a wide variety of concepts and tasks. For the authors and publishers, it has the benefit of wasting a lot of space. This fattens the book and makes it look like you are getting more for your money.

Unfortunately for the readers, the book is difficult to read and completely misses the point of the SSAS interface. XML is the underlying metadata structure of SSAS. That is the last place you should look to understand cube and dimensional structure, or for modifying how the cube works.

For a professional programmer (me), time is money and productivity is everything. First you should should use the graphical and tabular representation of metadata to manipulate the cubes and dimension. THEN, you write MDX functions when necessary. If all else fails, mess with the XML.

If you can wait a few weeks, there are two new books coming out for AS 2008. I believe either would a much better alternative to this book.

Don't waste your money buying this book like I did.

Regards,

Dan.

First, this book is not a tutorial. If you don't know SSAS and want to learn it, look elsewhere first and return to this book later.

What this book provides is an in-depth view of how Analysis Services really works. For those of us who spend most of our time working with Analysis Services, this book is invaluable for understanding how the engine behaves and why. You'll find details on processing, aggregations, attribute relationships, and virtually every other aspect of SSAS. The level of detail is exactly what you would expect from members of the SSAS team at Microsoft, which is to say it is very detailed and technical in nature.

There are five chapters on MDX that explain the subtleties of the various functions and how they perform. There are also chapters that delve into performance tuning, security, and administrative tasks. I highly recommend this book to anyone already familiar with SSAS because it contains valuable information not found anywhere else.

If this is the best SSAS book for the hardcore technologist, then I shudder to think what the worst one is.

I bought this book having no knowledge of the SSAS product or MDX. However my preference is for a theory-heavy book rather than a step-by-step introductory tutorial, so I would rather go straight for the jugular with the expert level book. To their credit the authors being deeply involved in the development of SSAS seem to know the product quite well, but their ability to explain concepts effectively and give you the big picture, enabling you to best construct an OLAP system with their product, is a bit lacking.

First, I'm puzzled how a book can go to mass production without someone at least running the text through the grammar checker--there are numerous grammatical errors throughout this book. The index also doesn't seem very accurate. I also found myself questioning the correctness of a couple code samples. These are but the first signs that the book was hastily written & published in an effort to be the first book to market on the 2008 version.

My second problem is with the ability of the authors to explain concepts of a technology that is foreign to people coming from the RDBMS mindset. MDX is an odd language, but many of their explanations & code samples often make the language even more confusing than is probably necessary. Let's take an example: in the cube-based mdx script chapter, they introduce the concept of static vs. named sets. They essentially explain it as such: "dynamic named sets are different than static named sets. Without explaining what a dynamic named set is, we'll just give you a code sample showing you the difference and hope you figure it out." While it is possible to then figure out the concept, I found myself very slowly reading the MDX chapters to try to both grasp the odd syntax of the language while understanding the subtleties of their cryptic code samples. Having gone through the MDX chapters I still don't think I know how to best use all the features of the language, and will probably only learn it by writing it.

Third, while it might not necessarily be the fault of this book, reading it makes me want to strangle whoever created MDX. Where is Anders Helsberg when you need him? This language is a mess. For example, we have a NON EMPTY operator and NonEmpty function being deceptively similar, yet different. Why not give them different names, just to make it a little less confusing? Also if the WHERE clause is so different from a SQL WHERE clause, why not call it something else? I get the feeling this language has gone through a few iterations to make it more SQL-like, and in so doing the original vision has been hijacked with "improvements" that only muddle it. My frustration with this book is it hasn't successfully enabled me to "get" the method to the madness of this language; i'm often distracted by its peculiarities.

One positive note, if you read carefully you will pick up some best practices, recommendations for optimizing performance, etc. It would've been nice to have a chapter dedicated to best practices, so if you don't have the time and need to skim, you can get the information quickly. However its there if you dig.

OK, if you are a technologist and are inclined to write a book on a specific technology then read this most excellent book and see how it should be done. This explains everything about SSAS in finite details. I am writing a course on Microsoft BI and this is my reference of choice.