Hi Ken,
That link between div_allele_assay and div_stock is
a "reference" only. The path that you suggest is the
correct path. Look at my graphics picture of the
GDPDM schema. I didn't even draw that link between
div_allele_assay and div_stock. With your dynamic
path algorithm, leave out any link that I did not draw
on the schema diagram. That will solve any
ambiguities that you have.
Cheers,
Terry
On 1/10/06, Ken Youens-Clark <kc...@cs...> wrote:
> Terry || Ed || Isaak,
>
> Per a request from Isaak, I'm trying to create a display of a record
> from the "div_experiment" table that includes its related
> "div_scoring_tech_type." My code is using some graph theory to
> automatically navigate the schema to get between these tables. Using
> a shortest path-finding algorithm developed by Dijkstra, it has
> determined the path to be this:
>
> div_experiment
> div_obs_unit
> div_stock
> div_allele_assay
> div_scoring_tech_type
>
> However, looking at the schema, I think it should probably go like this:
>
> div_experiment
> div_obs_unit
> div_obs_unit_sample
> div_allele
> div_allele_assay
> div_scoring_tech_type
>
> Is there a preferred path, or are both valid? How do you feel about
> there being multiple possible paths through the schema? Is this a
> sign of weakness or robustness? I don't know the biology nearly
> enough to comment on how well the schema mirrors the data, but I'm
> worried by the fact that the schema isn't explicit in its definition
> as to how to get from one table to the other.
>
> ky
>
>
|
