Re: Single vs. Multiple Sequences

From: Holger Baer <holger.baer_at_science-computing.de>
Date: Fri, 09 Sep 2005 15:37:51 +0200
Message-ID: <dfs37f$ja9$1@news.BelWue.DE>

EdStevens wrote:
> Chris McMahon wrote:
>

>>Are there performance advantages to having a sequence generator per
>>table VS a single sequence generator used across all tables?  Are there
>>locking issues that might be a problem with using a single sequence
>>generator?

>
>
> There are too many variables, and too little information provided, but
> it should be simple enough for you to set up a test. Nothing beats
> actual measurement *in your environment*.
>
> That said, I'd be surprised if there were any *significant*
> differences. So much so, that I'd think you'd let the business
> requirements drive the use of sequences rather than any minor
> performance differences. The way you state your question, it sounds as
> if you believe that every table arbitrarily requires some sort of
> sequence number be generated for the rows.
>

Read the link Stephen provided. You'll be amazed ;-)

But your point still stands. Why do people never test? Is it so hard to use that grey mass between the ears for something else than heat the blood?

Cheers
Holger Received on Fri Sep 09 2005 - 08:37:51 CDT