Re: duplicating ORACLE GL database for quick report...

From: Dan Clamage <clamage_at_mime.dw.lucent.com>
Date: 1997/09/17
Message-ID: <01bcc377$3aa90050$54110b87@clamagent>#1/1

peewee_at_club-internet.fr wrote in article <874306244.20616_at_dejanews.com>...
> We're implementing Oracle GL v10.7 on an UNIX server / a 50 client
> forecast.
> Many reports will be developped under FSG and we fear an
> overload of the main database.
> Our consultant suggested us the possibility of
> duplicating easily the GL database and launch
> - in real time - FSG queries on the copy.
> Is it a valuable method ? Is it a reliable choice ?

The generalized approach is called data warehousing. You denormalize some tables and build a "duplicate" data set, preferably on another server, to handle queries against the denormalized set. This new set represents canned approach to obtaining data. The reason its denormalized is so you don't have to compute joins. The duplicate data is updated periodically using the built-in replication capability of Oracle.

There are plenty of books and tools on the subject. Oracle has also the concept of a data mart, which is a small, focused data warehouse (perhaps only looking at the Marketing related tables). - Dan Clamage Received on Wed Sep 17 1997 - 00:00:00 CDT