performance approach PL/sql ?

From: Nan <nandagopalj_at_hotmail.com>
Date: 5 Apr 2002 07:21:43 -0800
Message-ID: <8193246.0204050721.5a844441@posting.google.com>

I have 20 oracle tables containing in excess of million rows each. I have to write a query that joins the 20 tables. The query generates summary information from all the 20 tables. I have a driving table and join its key to some of the 19 tables and their key to other tables and so on.
I have to do this for all the million records in driving table.

I plan to write a PL/SQL that basically performs a join on all the records and associates the row IDs of the 19 tables with the driving table's key and store the relationships in a new oracle table. Once this script completes, I can just run my SQL off of the relationship table and get the info with a direct row id based query for all 19 tables. This will improve performance of extracting all the 20 tables' data.

Your opinions on this approach is highly appreciated..

Nandu. Received on Fri Apr 05 2002 - 09:21:43 CST