gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnumed-devel] Medication errors related to drug names, and the appr


From: Jim Busser
Subject: Re: [Gnumed-devel] Medication errors related to drug names, and the approach of TALLman lettering
Date: Mon, 05 Sep 2011 22:50:21 -0700

On 2011-09-05, at 4:22 AM, Karsten Hilbert wrote:

> Across-loanguages matching can be improved by ignoring
> trailing e's if that can be effected and still true to the
> purpose (reformatted for clarity):

Both forms should be applied (the e-containing form first) so that

        acetazolamide

becomes

        acetaZOLAMIDE (not acetaZOLAMIDe)

after which a second pass, substituting

        acetaZOLAMID

achieves the desired intention, no matter whether the original had been

        acetazolamide
        acetazolamid

so that, after two passes, they will each end up

        acetaZOLAMIDE (where originally ending in 'e')
        acetaZOLAMID (where originally not ending in 'e')

My approach had been to create the string of an SQL query (command)

        UPDATE …

one per row (i.e. one per drug of interest). In order to achieve both e and 
non-e forms, I could either create two files and combine them, or … (would it 
work to) set my special-purpose column 'overlay_string'  thus manufacturing the 
string

        UPDATE … desired drug case into existing drug case … ;

and afterward doing a second update, concatenating into my 'overlay_string' 
column:

        escaped return
        UPDATE … desired drug case (non-e form) into existing drug case 
(non-eform) … ;

achieving a single string that would contain both statements per drug (an 
e-form statement and a non-eform statement), in the aim that – when I would 
copy out the cvs file (in fact achieving an SQL file) – I would achieve two SQL 
statements per row?

-- Jim
        


reply via email to

[Prev in Thread] Current Thread [Next in Thread]