Re: isolation level serializable

Chuck <skilover_nospam_at_bluebottle.com> wrote:
> xhoster_at_gmail.com wrote:
>
> >>> Then I do another commit at the end of the last insert. There are no
> >>> commits, rollbacks, DDL, or any other SQL between the inserts,
> >>> however I still get FK violations at the end when I try enable the
> >>> constraints. Why?
> >
> > To know why, it would help to know what. What is the problem that is
> > causing the violation?
>
> As stated in the OP, it is a foreign key constraint violation on one the
> destination tables. A child row has no parent.

Right, so clearly either the child shouldn't be there or the parent should be there, but can you figure out which one? If the source table has time stamps or doesn't change extremely quickly, it should be possible to infer which is the case. Once you do, I'm still not sure what the next step would be, but it can't hurt.

>
> > Are you absolutely sure that the corresponding
> > constraint is in place and enabled and set to validate on the source
> > table?
>
> Yes. Absolutely.
>
> > Is your process the only one that writes into the s1 schema or can
> > other people be messing around with it at the same time as you are?
>
> My session is the only one that has access to the s1 schema.

I guess I'm out of ideas, then. When you figure it out, please let us know what was going on--it seems like quite a mystery.

Xho

-- 
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service                        $9.95/Month 30GB