Re: Data size impact on Performances

If the information is to be used in a certain way (ie; for data warehousing) you may want to consider using materialized views (aka snapshots). Have these do the work of any data manipulation base your quiries and the results your looking for. It's great if you can drop/create the indexes on the MVs once in a while; or more frequently depending on how much data gets entered in at any one point in time.

We have tables with 18+ million row that we use MVs on. We also have scripts that run via cron every night that create just plan summary tables, rebuilds their indexes and performs an analyze on them. (And we do use OEM for some of this as well - what a pain!) If you do use MVs (or even just plan tables of the summary information), be sure to let those using the data know how frequently the data gets refreshed.

"V. Fraser" <vincent.fraser_at_canadair.ca> wrote in message news:5092032b.0203071420.7cb2f7ac_at_posting.google.com...
> Hi,
>
> I have an application running with Oracle 8.1.7 under AIX that support
> 250, 000 object in a single table. The table has all the proper
> indexes and was tuned by an Oracle admin.
>
> If I plan to insert new objects in this table to reach a total of 600,
> 000 documents, is their a way to evaluate the impact on a single query
> on this table using an attribute defined in the indexe of this table?
> A special formula? A chart to extrapolate?
>
> Thanks in advance,
>
> Vincent
Received on Sun Mar 17 2002 - 08:48:30 CST