[MOBY-l] MOBY Central database schema

[Mark Wilkinson]

Hi Jason,

We seem to be close to consensus :-)

Below I have trimmed out a piece of the GO database schema to show you what I
mean by the term2term table.  It allows us to build the hierarchy as a DAG or
a simple tree, whichever we prefer.  We can go either route at the beginning,
but it is good to have the flexibility to change later on without having to
re-write the schema.  It also allows each "edge" to have a particular
relationship type.   I think these three tables, together with the ones we
have already discussed, will be sufficient to keep us in business for quite
some time!

             Table: term
                     acc
                     id
                     is_obsolete
                     is_root
                     name
                     term_type


             Table: term2term

                        id
                        is_inheritance
                        is_obsolete
                        modification_id
                        relationship_type     # I have no idea what is stored
in this slot...??
                        relationship_type_id
                        term1_id
                        term2_id



             Table: relationship_type
                     id
                     type_desc
                     type_name



Ewan tells me that he has now given both Jason and Midori CVS access now.
Please check that.  Lucas - I just checked and you are NOT yet a developer.
I'm sending another request to Ewan right now - he'll write to you directly.
Hopefully he's still at work...

Gotta run!

M





"Jason E. Stewart" wrote:

> Hey Mark,
>
> Just so we don't tramp on each other's feet to much, here's the schema
> I've built for Postgres, let's standardize it, check it into CVS and
> work on it together.
>
> I prefer verbose column names and not mixing case - because different
> DBMS handle case differently.
>
> Also I had made a dedicated InputParameter table and a dedicated
> OutputParameter table, but there isn't any point really, the type
> should do that for us.
>
> Instead of a Service table and an Object table I used a single
> Ontology table.
>
> Also, I'm including the skeleton of a Perl Module that I wrote to test
> it all out.
>
> Cheers,
> jas.
>
>

--
--------------------------------
"Speed is subsittute fo accurancy."
________________________________

Dr. Mark Wilkinson
Bioinformatics Group
National Research Council of Canada
Plant Biotechnology Institute
110 Gymnasium Place
Saskatoon, SK
Canada