> I believe 6.5 is operating as it should --- 6.4 was producing inferior
> plans for small tables.
>
Yes, absolutely.
> But it is clearly a Bad Thing to allow the 6.5
> optimizer to believe that a relation is empty when it isn't. I concur
> with your suggestion to hack up CREATE INDEX so that creating an index
> before you load the table isn't quite such a losing proposition.
>
> > Please apply the patch I previously sent.
>
> Will do.
>
I think this will save us a lot of complaints. Thanx
Andreas