February 15, 2007 at 3:53 pm
We are setting up a DataWarehouse. We will have many tables of the same name coming from different sources (.e.g. SourceA.table1, SourceA.table2,.... SourceB.table1, SourcerB.table2,....). Despite the same tables names there may be significant differences between the tables preventing them from being readily combined.
It has been suggested that we keep all tables together in one database, using the Schema as a "seperation" device:
Database A
SourceA.table1
SourceA.table2
....
SourceB.table1
SourceB.table2
I think two databases may be a better approach.
Database A
SourceA.table1
SourceA.table2
Database B
SourceB.table1
SourceB.table2
What approach would you use and why?
TIA,
Bill
P.S The app that uses these databases is a third-party ETL tool which does not make use of stored procs!
February 19, 2007 at 8:00 am
This was removed by the editor as SPAM
February 20, 2007 at 7:00 am
Ralph Kimbal in 2004 published a 38 points ETL(extraction transformation and loading) checklist it is a formal structured way to move from OLTP to OLAP. Hope this helps.
Kind regards,
Gift Peddie
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply