Re: Uncommited insert blocks update?

Hi,

 I could imagine you are waiting on some unique index.  The insert could be locking index entries - if you are updating those  indexed fields, the update has to wait for the insert to finish,  otherwise it could'nt verify uniqueness.

 I am just guessing though....

HTH Karsten

In article <951165814.16990.0.nnrp-13.9e984b29_at_news.demon.co.uk>,   "Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote:
>
> Could you clarify this slightly -
>
> Are you saying that an update statement
> from one session attempting to update
> a row inserted but not yet committed
> by another session appeared to hang
> waiting for the commit/rollback ?
>
> --
>
> Jonathan Lewis
> Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk
>
> Lothar Armbüster wrote in message
> <924.86T521T12442960lothar.armbruester_at_rheingau.netsurf.de>...
> >Hello out there,
> >
> >today I observed some strange behaviour of our Oracle 8.1.5 database
on
> >NT.
> >We had a long running insert blocking updates on the same table.
> >That means the update seemed to wait for a lock but when i selected
> >the rows which should be updated there were nothing. Obviously the
update
> >statement waited for the insert to complete in order to change the
newly
> >inserted rows (or some of them). I expected that rows that cannot be
seen
> >by a select would not be affected in an update.
> >Is the observed behaviour a bug or a feature of Oracle?
> >
> >I could imagine some use of such a behaviour when trying to update
unique
> >columns which are inserted but not commited yes but I want to know
if this
> >is the desired behaviour or if it is just a bug.
> >
> >Any ideas?
> >
> >Lothar
> >
> >--
> >Lothar Armbrüster | lothar.armbruester_at_rheingau.netsurf.de
> >Hauptstr. 26 | lothar.armbruester_at_t-online.de
> >D-65346 Eltville |
> >
>
>

Sent via Deja.com http://www.deja.com/
Before you buy. Received on Tue Feb 22 2000 - 04:32:32 CST