Gerald Brandt wrote:
> Hi James,
>
> Actually, it all COULD be placed into the base tables, and in fact it was
> before I separated them. I just thought that I would get a speed increase
> by more cleanly separating the data. Then I sat down and thought about it
> some more, and now I'm not sure. Hence the question...
>
> Gerald
I think I understand now. As far as the original peformance question you
asked, I am not in a position to answer it, but as a software tester (one of
the many hats I wear) I would suggest creating some sort of script or program
that goes in loop and generates the SQL code to generate 500 sets of these
tables. Then I would write another program to fill the tables with pseudo
random data. Then I would write another program to do various inserts, and
selects againsts this data, only being the devious tester that I am, I would
create a shell script or perl script perhaps that would kick off about a
hundred of these, maybe a thousand if the system would handle (actually if it
doesn't that would be good, because you kind of know a limit to your system,
but if its a live system...) and during all this, I would be generating logs
to keep track of the time it takes to do all these transactions. I might at
that point generate another script to pull all this data together and see if I
could make something out of all the data...I might even put the data in a
postgreSQL database.
Anyway, that's a lot of work, but it is a sure way to find out just what will
happen before you spend a lot of time developing a polished user app that
bases its internals on this. It would probably be fun even, but...it
certainly would at least eat up an afternoon...james