> > -> Index Scan using
> teststscell13_pkey on teststscell13 data1 (cost=0.0..3.9
> rows=1 width=16) (actual time=0.006..0.006 rows=0
> loops=285)
> >
> > doesn't make any sense: that table will never have any
> data.
> > I'd like to have a way to tell that to Postgresql...
>
> It's one index probe and takes virtually no time at all.
> That's not your problem.
>
Put that in a 60000 nested loop and it won't be "virtually no time at all" I'm afraid... to the planner that "3.9 cost"
almostthe same as an index scan on a populated table...
Hence the planner uses a different plan.
Otherwise I don't see why the 2 plans should be different...