Re: [HACKERS] regression bigtest needs very long time

wieck(at)debis(dot)com (Jan Wieck) writes:
> What I'm actually wondering about is why the hell using
> NUMERIC data type for fields where the database shouldn't
> calculate on. Why not using TEXT in that case?

He didn't say his application would be *all* I/O; he was just concerned
about whether the change would be a net loss if he did more I/O than
calculation. Seems like a reasonable concern to me.

> OTOH, I don't think that the format conversion base 10000->10
> overhead will be that significant compared against what in
> summary must happen until one tuple is ready to get sent to
> the frontend.

I agree, but it's still good if you can avoid slowing it down.

Meanwhile, I'd still like to see the runtime of the 'numeric'
regression test brought down to something comparable to one
of the other regression tests. How about cutting the precision
it uses from (300,100) down to something sane, like say (30,10)?
I do not believe for a moment that there are any portability bugs
that will be uncovered by the 300-digit case but not by a 30-digit
case.

regards, tom lane