[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gnumed-devel] gmClinical.sql and general comments
From: |
Karsten Hilbert |
Subject: |
Re: [Gnumed-devel] gmClinical.sql and general comments |
Date: |
Sat, 14 Jan 2006 12:04:12 +0100 |
User-agent: |
Mutt/1.5.11 |
On Sat, Jan 14, 2006 at 12:44:50PM +1100, Richard wrote:
> > There is no such thing as "sit together" in relational
> > databases.
> You mis-understand my comment. Of course there is not.
OK, you understand this. I thought so.
> However if you have one
> hundred tables, and you are working on the vaccination section, and you have
> to keep scrolling up and down a large list were one table is called
> link_vacciantion.. bal bal, the next is called
> schedule_vaccines bla bla , the next is called
> vacc_dowhatver,, you have to keep moving up and down to find them to work on
> them.
Ah, I see, the tool you are using is not suited for the
task. It makes you work on names, on fingers pointing at the
moon, not on relationships - on the moon itself. A "proper"
tool would group tables relationally, or even let you define
content-groups you are interested in with depth-to-follow
(since relations can be cyclic graphs). There may not be
such a tool easily available, I agree.
> If you aggregate the names again, like I suggested with the file names as gm
> +
> schema + table + etc. they are easier to find.
Let's make this extra sure: are you talking about *table*
names or *file* names or both ?
> Though I've not put these in alphabetical order here, they sort so in the
> database tools when you are using them (Ah!!!, Light bulb goes on.... I see
> now, you don't use database tools so it is of no consequence to you!!!, but
> for those of us who do (like moi) it increases my productivity 100 fold.
I agree this can be the case. For me, I work with the source
SQL files in Midnight Commander. Such they are logically
grouped anyways. By relationship. Sort of. And Midnight
Commander doesn't try to second-guess my intentions
reordering the tables by some "rather random" criterion I
never intended.
> *taught myself elemental postgres
Good. It shouldn't be much different from any other SQL
database, just more standards compliant.
> *created a 40gig full product information database, populated it with data,
> done sqls to bring out the text for formatting into html
You might want to think about making views from the
formatting sql.
> *created 10gig prescribing database populated with upto date data
>
> *created contacts database and populated with dummy data
>
> *created coding database populated with complete data
In GNUmed that would map to the reference part. In the
clinical schema there's a "shadow" table - much like what
Hilmar suggested for the prescribing part - for the codes in
use.
> without such gui-tools I wouldn't get to first base, so humour me.
fine, whatever makes you productive
> Yes, you can't hold it all in your head.
I can but it's making things easier to not have to.
Karsten
--
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
Re: [Gnumed-devel] gmClinical.sql and general comments, Richard Terry, 2006/01/13