Re: One sequence or not?

From: M Hashim <m.a.n.hashim_at_sympatico.ca>
Date: Tue, 5 Nov 2002 15:59:18 -0800
Message-ID: <IeWx9.3408$w07.687870@news20.bellglobal.com>

It's good on the maintenance side, for not having to maintain numerous sequences. In an object oriented environment, it's usually a single identifier, OID. It's simple and effiecient from a coding perspective.

The negative, should you require continuous record identifier, having a single sequence will NOT give you continuous numbering. It will create gaps. In some financial institutions, having gaps may be a problem Think about it, from an auditing perspective.

"Richard Kuhler" <noone_at_nowhere.com> wrote in message news:IEUx9.63856$X9.22712785_at_twister.socal.rr.com...
> I'm working on a project where they've used ONE Oracle sequence to
> generate surrogate keys for all rows in all tables. This database will
> be used for ad hoc queries and reports. Since an ID will only appear in
> one table, the reasoning is that any incorrect joins will result in no
> data being returned rather than incorrect data.
>
> Opinions? Experiences?
>
>
> Thanks,
> Richard Kuhler
>
Received on Tue Nov 05 2002 - 17:59:18 CST