Subject RE: [Firebird-Architect] Re: [IB-Architect] Rebuilding foreign keys system indexes
Author Ann W. Harrison
At 02:44 AM 4/19/2003 -0400, Claudio Valderrama C. wrote:
>Ann Harrison wrote:
> >
> > In fact, that is the way constraints are supposed to work. The
> > SQL standard considers UPDATE T1 SET A = 5 to be a single statement,
> > not on statement for each instance of T1.
>But I'm not sure the engine agrees. :-)

I'm quite sure it doesn't.

>... However, the so-called CHECK-constraints are
>triggers, triggers written behind scenes by the DSQL layer, triggers
>affected by the isolation level of the current transaction, triggers that
>will fire for each changed record

Actually, I think that the check triggers are run in system context,
though that doesn't affect your example.


We have answers.