I think you got me there. I have to agree with both points.
(Besides, you are the one coding this thing and I think you understand it
better than I do.)
Let me know if I can help.
Jeff
----- Original Message -----
From: "Josh Berkus" <josh@agliodbs.com>
To: "Jeffrey D. Brower" <jeff@Green-Visor.US>; <pgsql@mohawksoft.com>
Cc: <pgsql-performance@postgresql.org>; <pgsql-hackers@postgresql.org>
Sent: Friday, April 04, 2003 11:09 AM
Subject: Re: [PERFORM] [HACKERS] OSS database needed for testing
> Jeff, Mlw,
>
> > Absolutely. We could just use one large state or several small ones and
> > let folks download the whole thing if they wanted. Using that technique
> > you could control the size of the test quite closely and still make
> > something potentially quite valuable as a contribution beyond the bench.
>
> Hold on a second. The FCC database is still a better choice because it is
> more complex with a carefully defined schema. The Tiger database would
be
> good for doing tests of type 1 and 3, but not for tests of types 2 and 4.
>
> It would certainly be interesting to use the Tiger database as the basis
for
> an additional type of test:
>
> 6) Very Large Data Set: querying, then updating, 300+ selected rows from a
> 2,000,000 + row table.
>
> ... but I still see the FCC database as our best candidate for the battery
of
> tests 1-5.
>
> --
> Josh Berkus
> Aglio Database Solutions
> San Francisco
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster