Re: [Gnumed-devel] Bootstrapping data - permissions, latin vs unicode, other

[Karsten Hilbert]

On Tue, Jan 04, 2011 at 10:27:07AM -0800, Jim Busser wrote:

> My objective is to get canadian drug data in, so I am wondering whether this 
> would involve one or more shell scripts paired with one or more sql files.
> 
> Also, whether the permission needed at runtime would be either the
> 
>       system postgres account password or
>       password for postgres user gm-dbo

The latter, sometimes the former.

> Also, where to look for examples to learn to understand.

It needs more thought than the FreeDiams stuff because one
cannot just delete the data an re-insert. One needs to
update existing data and insert missing data.

> BTW I looked at
> 
>       .../gnumed/gnumed/server/sql/gmCountryZones.sql
> 
> and wondered
> 
> 1) why is encoding "Latin 1"

That's what the text data in this file is encoded as.

> and does this make any problems for downstream unicode needs?

No.

> 2) why does it contain the states of *some* countries, but others (like 
> Canada's) are in
> 
>       gnumed/gnumed/server/sql/country.specific

No particular reason, just history.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346