Applied Microsoft Analysis Services 2005 and Microsoft Busin [Englisch] [Taschenbuch]

Applied Microsoft Analysis Services 2005, at a first glance, appears to be THE book to get on SSAS 2005. Unfortunately, the beauty is only cover-deep.

While the author clearly has "encyclopaedic" knowledge of the product suite, examples I came across were incomplete and quite frankly frustrating. For example, in Chapter 6, the author introduces us to the concept of using dimension definition templates for creating our Inventory cube "top-down" (meaning, without a data source)... while the instructions are clear on how to use a template to satisfy our Product dimension requirement (using the "Product" template), the author provides absolutely NO indication on which template we should use to create our Location template-- all the author writes here is "...Follow similar steps to create the Location dimension". That's it, nothing else (Chapter 6, Page 200)... no explanation is provided as to whether we reuse the Product template or not.

This book has many other examples of such dead-ends (half-truths and other misleading information). For instance, in Chapter 5, Page 188, on the topic of optimizing performance by caching linked object queries, the author writes: "...both linked dimension and measure groups expose RefreshPolicy and RefreshInterval properties which can be set at design time....". However, when I went through the associated example, I found that only linked measure groups expose said properties in support of linked object caching.

Once again, I do share the enthusiasm of the other reviewers, and have (to be fair) found some solid value in this book in explaining some of the new concepts related to the UDM, and the attribute-centric approach to creating dimension hierarchies (among other things). I also find the greatest value in a technical book to be the balance between clear explanations of new and/or thorny concepts (which this book does rather well) and clear examples for the reader to work through (which this book often does not). Nevertheless, I question the integrity of any reviewer quick to give a book 5 stars without actually reading through, actually trying each and every example, and rating the work objectively. Granted: no first version of any technical book can be perfect, but to quickly rush to awarding a book 5 stars is a dodgy practice at best-- perhaps indeed luring us to purchase a book we may later regret.

Caveat empor. 2.5 stars for this book....

This was the first book published on Analysis Services 2005. It was obviously rushed to press. At first I started making notes of all the grammatical errors, but then found so many I gave up and quit.

The technical content if very good. If you are doing a lot of Analysis Services work then you will definitely want this book. At the same time, you will want to pick up some of the other texts as they're released over the coming months. For one, the coverage of MDX syntax in this book is quite poor. You are also not going to find the "whys" of dimension modeling covered here or a set of recommended best practices.

Also, the book starts with a sample database and data warehouse and builds upon those samples throughout the book. If you are not looking for the kind of book where you type in the examples in each section, then this one may not be for you.

SQL Server 2005's Analysis Services (herein SSAS 2005) is challenging enough to merit it's own technical guide. Therefore, other popular books which attempt to cover SSAS 2005 along with other topics -- like Integration Services and Reporting Services, let along the DB engine itself, must fall short of what Teo Lachev has accomplished here.

The capability in SSAS 2005's new Unified Dimensional Model (UDM) to build a useful cube from complex, highly-relational source data, or even semi-structured data (vs. only from star or snowflake dimensional schemas) looks like a real step forward for OLAP, but the task is far from simple. In fact, some may wish that they had performed better ETL on source data. Along that line, I don't believe another book yet exists that adequately addresses real-world complexities requiring SSAS's member properties, referenced relationships, fact relationships, null properties, and other topics.

Why does this book not deserve 5 stars? Because it's sample database code is disjointed and does not flow well with the text. Usually, I had to read the text and then experiment on my own cube to resolve a challenge. Addditionally, many gramatical errors temporarily hold the reader back.

OK, then why do I rate it more than 3 stars? Because it provides understandable, authoritative coverage of pretty complex material. I repeatedly found that, when SQL Books Online did not provide a needed answer, Teo Lachev's book succeeded.

Although the author's explanations of MDX are good, they provide insufficient context for real skill-building. For MDX learners, I recommend "Fast Track to MDX (Version 2)", even though that book refers to SQL 2000. Once you get MDX (and you want to get MDX), you'll quickly see how you'll want to apply it in SSAS 2005. In closing, read Teo Lachev's book first for an understanding of the platform, and then the Fast Track book to learn how you'll perfect your access to cube data.