"Sybrand Bakker" <gooiditweg_at_sybrandb.demon.nl> wrote in message
news:vo98avkolsvs8isfaepspgbvlkvru5rcnu_at_4ax.com...
> On Mon, 21 Apr 2003 14:44:43 GMT, "Eric Fortin" <emfortin_at_attbi.com>
> wrote:
...
>
> First of all: Oracle has multi-versioning. This means the end-user
> will always get to see the correct version of the record. In Oracle
> readers don't block writers, in sqlserver readers do block writers.
> You seem to have been exposed to sqlserver.
> Get the Expert-on-one book by Thomas Kyte and read the first three
> chapters. It's all there!!!
> You should lock a row only when you really are going to change
> anything. NOT when you SELECT a row. We use a generic application
> which can run against any database implemented this way, and their
> 'locking' (by means of the flag mechanism you describe) is a REAL pain
> in the ass. It means no 2 readers can ever read the same record, even
> if they don't plan to save anything.
>
> Oracle has mechanisms to abort a session which is inactive for a
> prolonged period of time. Allowing a lock to stay for 10 hours is
> just plain ridiculous.
>
>
> Sybrand Bakker, Senior Oracle DBA
>
> To reply remove -verwijderdit from my e-mail address
Thanks, I will buy the book and read it.
Received on Mon Apr 21 2003 - 13:00:10 CDT
