Re: row vs row.column level locking

On Fri, 16 Dec 2005 19:22:15 +0000, NetComrade wrote:

> Why not? Let's say i have a table within a web application that has
> fields that are updatable by end-user and and fields updatable by a
> batch job. Currently I have to denormalize and separate the two
> tables, so that user won't block batch and vice versa. If I have
> multiple batch jobs running at different times I might have to do even
> more separation (if the batch jobs are taking too long)...

Or, you must do partitioning or you must organize your jobs better so that two fields in the same table are not updated in two different batches without a good reason. I would be extremely uncomfortable with the idea that two processes can write to the the same row. That would be a guaranteed recipe for corruption.

-- 
http://www.mgogala.com