Re: [Health] High Level Database architecture question

[Marc Murray (MOH)]

On Wed, 2016-10-05 at 12:20 -0400, Vernon Oberholzer wrote:
>     Hi,
>     
>     Traditionally, many healthcare systems have used some form of the
>     Mumps(M) database. https://en.wikipedia.org/wiki/MUMPS
>     This database is not a traditional relational database, and is
>     claimed to be particularly well suited to healthcare data because
>     of it's flexibility and speed.
>     
>     With GNU Health using the Postgresql relational database as it's
>     foundation, have there been any issues relating to flexibility or
>     performance, or have these factors been satisfactory thus far ?
>     
>     Thank-you,
>     -Vernon
>     

Dear Vernon, 

I have been using GNU Health (you know, Tryton, PostgreSQL and the lot) in production for nearly 2 years. Currently, I have about 80 users across 5 installations and those numbers are increasing weekly. Flexibility with PostgreSQL or the Tryton ecosystem has never been an problem at all. One of the things that makes Postgres awesome is that one can do schema modifications within a transaction. And, if you really want to store unstructured data (e.g. key:value pairs as offered by MUMPS), you can use the hstore or jsonb formats, native to Postgres.

Oh, and it's also ACID compliant.

If you're new Tryton and new to Python, the going may get tough when you want to customise the application. To help with that, I think one good source of documentation is the source-code of other modules. 

Good luck. 

---
Marc