SNOWFLAKE AND DATA WAREHOUSE

SNOWFLAKE AND DATA WAREHOUSE

By W H Inmon

It was with great interest that I saw where Snowflake just advertised their support for the book on data warehouse for dummies, 3rd edition. I decided to take a look at the book and here is my assessment – my opinion - of the book.

If I were a teacher and this was a project done by my student I would give the student a D-. I would consider an F but there were a few interesting points in the book. There are so many things lacking in the book and things that were wrong with the book that will mislead people that it is hard to know where to start.

1)????? Integration of data. Nowhere in the book could I find the subject of integration of data mentioned or treated with any emphasis. The very essence of data warehousing is integration of data. Vendors and consultants have always hated integration because it is a messy subject. But data warehouse is not data warehouse without integration. Integration makes up 90% of the effort of a real data warehouse project but Snowflake and the book do not mention it at all. To use an analogy, data warehouse without integration is like a car without tires. It just is pointless to buy a car and never put tires on it. But that is what Snowflake and the book they promulgate have done. One wonders if Snowflake and the author of the book even understand what a data warehouse is and the purpose served by integrating data in a data warehouse.

A real book on data warehouse would have at least 50% of the book addressing integration and the issues of integration and techniques and technologies for integration. Shortchanging the subject of integration in discussing data warehouse is like cheating at solitaire. It casts doubt on anything else the book and the vendor say.

2)????? ?There were only a few sentences mentioning text in a data warehouse in the book. Text belongs in a modern data warehouse even more than structured data. But Snowflake and the author do not have the slightest understanding of how to get text into a data warehouse in a usable manner, how to store the text in a usable manner, and what to do with text in the data warehouse once it is there. This is a major and glaring omission from the book. To use an analogy, a modern data warehouse without text is like an ocean explorer that has only been to the Mediterranean thinking that he/she has explored the world. The explorer that knows only the Mediterranean knows nothing about the Atlantic, Pacific, Indian, or Arctic oceans. Textual data is as essential to a modern data warehouse as a horse is at a rodeo. You aren’t going to have a rodeo without horses. And Snowflake and the book will tell you that they know about data warehousing when they don’t even understand the rudiments of acquiring and managing textual data in the data warehouse.

This omission of text is even greater than the omission of integration. But both omissions are truly great and hugely significant and cast great suspicion as to whether the author and Snowflake have an understanding of data warehouse.

The book promulgated by Snowflake is misleading to a fault. It leads the reader to think that Snowflake and the author know about data warehouse. But Snowflake and the author have an enormous amount to learn before they parade as leaders.

And that is my opinion.

?

Bill Inmon lives in Denver with his wife and his two Scotty dogs – Jeb and Lena. Jeb is getting older now but he still has his good days and his bad days. It seems his good days are sunny and warm, not rainy.