[Gnumed-devel] Maintaining patient demographics

[Jim Busser]

It is possible in GNUmed to have a family as patients (sometimes an extended 
family) where three generations totalling several people all share the same 
telephone.

Does GNUmed normalize replicates of a communication channel, as a consequence 
of which any change to the value will be applied to all others keyed to that 
same value (which could be good or bad) or are such channels non-normalized?

In case one would wish to query in GNUmed, the expression in the report

        find patients by phone/fax/...

which has a line

        where dem.v_person_comms.url = '1234567890'

will find

        1234567890

but won't find

        (123) 456-7890

One option would be

        where dem.v_person_comms.url LIKE '%1%2%3%4%5%6%7%8%9%0'

however the above is very difficult to see clearly. Is there any 
postgres-supported expression which could evaluate a transform of (strip 
non-numerics from) a postgres column's values and allow them to be compared to

        '1234567890'

??

I was unable to get past the following which only supplies numeric strings 
*preserving*

        (
        )
        -

        select SUBSTRING(dem.v_person_comms.url, '[-0-9 \(\)]*')

-- Jim