> Retested and made some configuraion changes to Firebird (FB was
> using more memory than it should).

what were the config changes? how much memory is too much from your
judgement? or do you mean more than PG?

>
> I have now ran test against PostgresSQL (since most of you felt that
> a wasn't comparing apples to apples)
>
> Same Table, Data and Indices are in both Databases.
> Both Databases were ran on the same machine.
> I didn't make ANY changes to the configuration file of PostgresSQL
> (didn't know what most of them were for.)
>
> More Info on the table layout can be found here:
> http://groups.yahoo.com/group/firebird-support/message/61124?
> threaded=1
>
>
> NOTE: The computer was rebooted after each run.

this will certainly defeat FB's caching which I think is quite different to
PGs. So this may not be an appropriate comparison. FB always runs better on
the second go whereas PG's second query will not be aa much of an
improvement.

>
> SELECT COUNT(*) FROM myTable;
> PG = 0.52 secs FB = 1.38 min
>
> SELECT DISTINCT store_no FROM myTable;
> PG = 0.51 secs FB = 1.31 min
>
>
> SELECT store_no FROM myTable GROUP BY 1;
> PG = 0.27 secs FB = 2.12 min
>
>
> As you can see, FB is running better than before, but it still runs
> much worst than PostgresSQL (without ANY configuration changes)
>
> GROUP BY are terrible in FB !!
>

then you need to conclude what you will from this. But I would still say
that even with superior performance on these queries from PG, I would not
design my system to use these queries. I would skin the cat a completely
different way and I would get the same result in less than a second.
Evenqueries which run for 30 seconds is totally unacceptable to me. I prefer
subsecond results but you cannot get this with a demand to scan 2 million
records.
Alan