Re: One sequence per table?

From: DFS <nospam_at_dfs_.com>
Date: Fri, 19 Aug 2005 09:04:11 -0400
Message-ID: <dNkNe.321$WO2.63@fe06.lga>

sybrandb_at_yahoo.com wrote:
> If the tables are closely related (say there is one table with
> incidents, and one table with changes, the incident/change number
> being communicated to customer) having separate sequences is a major
> pain.

Couldn't it be resolved relatively easily by the application or interface logic?

> Also you need to take in account is your single sequence is
> ever going to 'wrap', and hit itself in the tail.

Probably not. This one might get up to 400,000, worst case.

> That said, I would in the majority of cases set up separate sequences.

OK. I was probably heading that way, but I remembered an Oracle DBA I respected (this was before I learned about DBAs in general :-) using the single SEQ pool.

> Twelve sequences looks like nothing for me.

Probably not. It might save 1/2 an hour if you go slow.

Thanks for the reply, Sybrand. Received on Fri Aug 19 2005 - 08:04:11 CDT