Re: Data Warehouse question

From: Hans-Friedrich Pfeiffer <Pfeiffer.Hans-Friedrich_at_t-online.de>
Date: Mon, 02 Aug 1999 01:21:17 +0200
Message-ID: <37A4D66D.AF52989F@t-online.de>

michael_bialik_at_my-deja.com wrote:
>
> Hi,
>
> What is the better way :
> 1. To keep a "fact" table and a number of "dimension"
> tables separate ( perform a star join for each
> query ), or
> 2. Denormalize the tables and avoid the joins.
>
> Fact table - 3M - 4M rows.
> Dimension tables - 20 - 30K rows max.
>
> Thanx. Michael.
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.

Hi,

when you already use fact tables, I think your fact table is denormalized.
Using facts and dimensions has the great advantage that you can query your facts
by your dimensions. Queries or later demands of queries are much simplier then.

I use a star schema for a data warehouse database. For the performance :
think about using mmaterialized views in Oracle 8i. I use them frequently
and the performance is GREAT ! I have tables of 70.000.000 entries ( fact )
and in general 3 to 5 dimension tables - you image the join ? Using materialized views is a clever answer to performance questions.

Greetings

Hans Pfeiffer Received on Sun Aug 01 1999 - 18:21:17 CDT