Re: multitable inserts and sequences

Larry Elkins, 28.06.2013 03:58:

> In short, the order in which the tables are inserted is not determinant. Take a look at Note: ID 265826.1, referencing bug 2891576.
> So, if someone is considering this approach, the workaround is deferred constraints (our choice when facing this), or, temporarily
> disabling the constraint. The issue *wasn't* the sequence, it was the order in which the inserts were processed. So, just a head's
> up to those considering it.

Thanks for the answer. We will then keep on using PL/SQL and a loop that guarantees the sync between the sequence usage (even though there is no FK we do require the values to be "in sync".

Jonathan Lewis, 27.06.2013 21:04:
> I think the problem with this is that until it's documented to be working as expected you can't guarantee
> that it will work in the future however many times you test it now.

Good point ;)

Regards
Thomas

--
http://www.freelists.org/webpage/oracle-l