Re: data warehousing

In article <3btg71$ldh_at_newsbf01.news.aol.com>, jgreene_at_aol.com (Jgreene) writes:

<Other than that, I would suggest a good
<normalized table design, splitting indexes and data tables

Joe, you're running a 135Gb database, so you must know what you're talking about. However, I doubt that you are running a _data warehouse_ in the flavor that is commonly discussed. These databases serve Decision Support - type queries, which are known to be a strain resistant to normalized design.

<and getting a hold of a good set of monitoring/tuning scripts or tools
(scripts are
<published in such places as the Compuserve form or sometimes your Oracle
<tech rep can get you copies).

That may work if you're on a tight budget. I would suggest the use of a real configuration management/performance monitoring tool like EcoTools, Patrol, etc.

<Tuning and the formation of good queries
<can turn 30 minute queries into 30 second ones (we have seen this with
our
<large instances).

Now I'm convinced this isn't a Data Warehouse. Because of the ad hoc nature of Decision Support Systems, you just don't have the opportunity to tune queries. I would say that your advice is useful for large databases, but not really focused on the Data Warehouse issue.

Neil Raden
Envirometrics, Inc. Received on Sun Dec 11 1994 - 07:50:12 CET