Re: [Gnumed-devel] simplifying gmdrugs.sql

[Hilmar Berger]

Hi,

On Wed, 1 Jan 2003, Horst Herb wrote:

> I had a few complaints about the complexity of gmdrugs.sql, and I am starting
> having my own concerns regarding performance with so many subqueries
> neccessary.
I agree that gmdrugs is somewhat complex and yes, there may be room for
improvement. On the other hand I'm quite convinced that pharmaceutical
data are of a rather complex nature. If we really aim to have a database
that should hold drug data at a relatively high level of detail and from
different sources (that is, different national drug databases) I
almost tend to believe that we will need the level of detail/complexity
that is already implemented.
I don't see how we will get a very much simpler design without loosing a
lot of significant information. Plus it will be much simpler to discard
information / complexity from an existing database once we see performance
hits caused by to many queries than to go the other way around.
I would suggest staying with the core structure and test its functionality
and performance first. I admit that there might be parts of the schema
that could be done better in a different way. We should focuse on
improving those parts instead of doing a complete rewrite.
The simplified schema has some very useful suggestions, on the other hand
it is lacking tables on packages, amount of substance used in a
preparation etc.
Please give us more detailed information on your complaints about the
general structure of gmdrugs so that we can discuss those problems.

Hilmar