Search in Book Reviews

The book gets off to a rather slow start, with a somewhat unfocused chapters on basic relational concepts, business models and discovering keys, which I thought strange. If the reader needed this basic introduction, they probably are not ready to try building a data warehouse, never mind worrying about which philosophical approach was best.

Things then improve as the book settles down into the more practical steps of getting the model correct, and deciding what data needs including. It then moves into more data warehouse type areas, looking at managing hierarchies, the calendar, and transactions, all key areas for data warehousing. Most of these chapters concentrate on relational techniques. There are some sections I personally found odd - for example some explanations of when to use bit map indexes - which seemed out of place in a book that felt it needed to start by defining normal forms and then only the first three.

By the end of the book I did not really feel the authors had addressed "head-on the challenging questions raised by Kimball" as promised on the back cover. I also found it not a particularly easy read; it never really managed to make me want to pick up the book to explore further, though to be fair I've read much worse. The book seemed strongest in the chapters when it concentrated on relational based techniques for smaller data warehouses, which are being built by people who need a refresher at the beginning of the book on normal forms. Overall a worthwhile book for beginners to data warehousing who need to build smaller relational data warehouses, but unfortunately never really justifies its "mastering" title.