Updates

Latest Tweet



What's New?

Check out for latest innovation, a computer based training video collection


Like this Page

The Data Warehouse Toolkit: The Complete Guide to Dimensional Modeling (Second Edition) Review by F. Hu

Answers the question - how should I set up the schema?

There are many aspects to setting up a data warehouse and I have not found a book which adequately covers enough of the bases to think that it gives enough information that I might actually be able to build a data warehouse from scratch. This book covers the fairly narrow domain of how should I setup the database schema for my data warehouse. So, if you're question is - how do I start setting up a database for my data warehouse, this is the book for you.

The book is unique in that it presents several case studies of how data warehouses could be setup. In some ways, these examples are so simplistic, that it hardly seems necessary to put them down. But along with each example, a different aspect of the data warehouse is explored and this is how the author has chosen to organize these topics in this book.

What appeared to be lacking was relating the new data warehouse schema to what you might already have in your production environment. If the examples were enhanced to show what the original production database schema looks like, that might give not only a better idea of how to create the schema, but how to translate it and the problems associated with that.

So much of data warehouse design is inter-related so the concentration on dimensional modeling is limiting and no generic book can ever describe the situation you are facing in real life. You might begin to suspect that the reason why data warehouse projects are so complicated and fail is because of books like this which specify a specific and non-intuitive method for constructing database. The methods described in this book are certainly not intuitive. In my experience, things like creating tables to represent dates and abandoning the natural pkeys in your prodcution schema (as is recommended by this book) tend to make things more complex and difficult to understand as a developer. So beware of implementing things from ths book that don't make sense to you, because they probably don't.