On 7/22/07, Vincenzo Romano <vincenzo.romano@gmail.com> wrote:
> On Sunday 22 July 2007 19:20:08 Tom Lane wrote:
> > Vincenzo Romano <vincenzo.romano@gmail.com> writes:
> > > In the original setup, the "UNIQUE" constraint had been dropped
> > > *before* doing the tests. So the "slow" case is without the
> > > UNIQUE constraint but with an index. The NOT NULL was instead
> > > there.
> >
> > With what index, pray tell?
> >
> > regards, tom lane
>
> Sorry for the incomplete sentence.
> Read it as:
>
> In the original setup, the "UNIQUE" constraint had been dropped
> *before* doing the tests. So the "slow" case is without the
> UNIQUE constraint but with an index on NOT NULL fields.
Control question: did you recreate non-unique index after dropping
the UNIQUE constraint?
Regards,
Dawid