Re: [Gnumed-devel] Might ref.consumable substances need to allow an amount of zero?

[Karsten Hilbert]

On Mon, Sep 05, 2011 at 08:38:20AM -0700, Jim Busser wrote:

> > Unless, of course, one desires to use proper case in which
> > case the wrong casing *can* eventually be pruned.
> > 
> >> Can we accept to make the constraint
> >> 
> >>    unique(upper(description), amount, upper(unit))
> > 
> > Why not - at insertion of substances - make the existence
> > check case insensitive ?
> 
> I think I understand… I can apply the above constraint to the *data pack 
> installer* and in that way avoid the duplication (or at least one source of 
> it) to occur.
> 
> (A script can be more strict than the schema, it only cannot be more liberal.)

That's right.

> By having avoided the duplication, any inappropriate
> 
>       ramipril 10 MG
> 
> in the reference table will likely be the only reference
> row of ramipril 10 MG to be keyed to clinical data and at
> any point where a user (or admin) would make the correction
> in the reference row, it becomes instantly fixed for all
> clinical records which the change does not even (at least
> directly) touch. :-)

Correct. Even a script can do that (lowercase the MG) - say,
when the existence check is run.

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346