Re: A row locking problem that baffles all...

On Thu, 14 Jan 1999, Paul S wrote:
>
> "Pessimistic" locking is far safer. Optimistic locking is fraught
> with the problems that you mention, e.g. when several users post to
> the same record at around the same time, Data integrity can be
> compromised. Many of our clients are established RDBMS users with
> onsite DBAs/DB advisors who know the various locking mechanisms
> available. Because of the sensitivity of their data most have
> expressed that they cannot afford the corrupting of data that
> "optimistic" locking situations can produce, hence the safer
> "Pessimistic" approach.

In the "optimistic" approach, you check that the row has not been updated while you were working on it, prior to issuing your own update. That way you don't get the integrity problems to which you allude. Thw only drawback with this approach is that some unlucky users have to start over under certain circumstances. Still, in your situation it seems far better than having the user locked out while an irate customer calls their lawyer.

--
Jeremiah Wilton http://www.wolfenet.com/~jeremiah Received on Thu Jan 14 1999 - 10:32:57 CST