Re: Large set-based UPDATE

Hi ctcgag, thanks for taking the time to respond. Comments embedded below.

ctcgag_at_hotmail.com wrote in message news:<20030212161103.540$nd_at_newsreader.com>...

> What length of time would be acceptable? Any hope of tuning the
> bulk update to execute in under this amount of time? (And what is
> a set-based update?)

A set-based update is one where the WHERE clause of the UPDATE statement returns more than one row (ie. a set of rows)

> Based on what criteria? Can the web-layer users change this criteria
> out from under you

No.

> , or is it only used by the batch processes?

It's only used by batch processes.

>
> Isn't it illegal to commit within a cursor (and then continue to use the
> cursor)?

Only illegal to commit within a cursor that uses the FOR UPDATE clause

> Probably not or definitely not. What if it dies during processing,
> would you have some way to know which ones were already updated and
> which weren't, just by inspecting the data?

The bulk job is entirely re-runnable so no problem there.

> What happens when the
> web-layer users see data from this 15_000 where some has been
> updated and some haven't?

The web users are only interested in one row at a time, and the timeliness of the bulk updates is not that crucial. So no problem there.

Thanks again for your help.

Paul. Received on Thu Feb 13 2003 - 10:29:14 CST