Helen Borrie wrote:

> Not that I know of; although it would be a great idea to have a paper
> available. Have you tried googling the subject?

Yep.

> I'll be presenting topics on database design at the forthcoming Firebird
> conference in May. Indexing issues will be in there for certain. I'm sure
> all of the proceedings will come out as white papers afterwards....but it
> doesn't help you right now.

I'll still be interested then I expect.

> Do as Svein suggests: pick out the slow queries and raise a question
> here. If you don't have one, get hold of a client tool which shows you the
> plans for your queries. You can actually do this with command-line isql,
> using the -planonly switch.

Thanks for that info. I'm not the one actually working on it, so the facts from
our end may be slow in coming, but I'll tell you what I can.

These queries are being done to initialize a screen representing how work is
allocated on a factory floor. It reads 2 or 3 queries for each factory
workstation. Individually they only take a fraction of a second, but when they
are run in quick succession the roughly 30 queries take over 20s on our 200 MB
database.

The technique that sped up a smaller one didn't seem to work on the bigger one -
datapumping.

Regards

Mark