Re: do triggers slow updates down severly

On 10 Oct 2002 07:41:04 -0700, jack.beukering_at_services.fujitsu.com (jack) wrote:

>I want to use update trigger to record the timestamp, so I know the
>mutation dates of records. Inserts can be handled by a default
>declaration of the timestamp column.
>My question is, does the performance of batch-jobs suffers serverly at
>e.g. 10.000 records or do you start to notice it at 100.000 records?
>
>Doe somebody has any experience with such a case?
>
>Jack

Of all the ways to timestamp a record, among many other things, a trigger is the most efficient method. A trigger isn't free, but I doubt that you'll ever notice 'severe' performance degradation. IMHO, such a timestamp is a minimal requirement in a good database design. You might check out Oracle's auditing features for ways to maintain a more complete change history.

rww
Russ Wolfe
wolfer_at_peak.org Received on Thu Oct 10 2002 - 20:18:26 CEST