Friday, March 23, 2012

How to make a datawarehouse?

Hello, as far as I know for making cubes I need forst a Datawarehouse or I can even make a cube based on an OLTP database, but I think this option is not recommended because its the transaction database.

The Analisys Services Projects let me create cubes very easy based on a datawarehouse or a relational database.

The question is How do I make the datawarehouse based on the relational database?
How does the datawarehouse refreshed from the new transactional data?

I think that I will have to make some Iintegration Services Packages to convert relational data to datawarehouse but I have no idea in how to make it in theory and obviously in practice.

Thanks for your helpRalph Kimball is considered to be the "inventor" of data warehouse theory so check out his series of articles on the subject: http://www.rkimball.com/html/articles.html

He also has a couple of books which are generally highly recommended:
http://www.amazon.co.uk/exec/obidos/ASIN/0471200247/qid=1122200276/sr=8-1/ref=sr_8_xs_ap_i1_xgl/026-0530835-6201264
http://www.amazon.co.uk/exec/obidos/ASIN/B0006UEZO6/qid=1122200287/sr=2-1/ref=sr_2_11_1/026-0530835-6201264

A good textbook with worked examples is worth its weight in gold. The one I used is out of date now and I'm not sure which is the best textbook around at the moment. Perhaps someone here has some ideas?

-Jamie|||Your second amazon link is broken. Could you please provide the name ant the author?
Many thanks!
|||I would assume Jamie meant to link you to several of Ralph Kimball's books.
The two I have on my desk are the Data Warehouse Toolkit and the Data Warehouse ETL Toolkit. Both are primarily written by Ralph Kimball. Ralph also has another book called The Datawarehouse Lifecycle toolkit, but I haven't read it yet.
There is a philosophical schism in the area Data Warehouse design, architecture and implementation. Kimball is by far the more popular and is the camp that I tend to follow. The other camp is lead by Bill Inmon. You can find some more information at his website (http://www.inmongif.com).
Larry
sql

No comments:

Post a Comment