[BioSQL-l] Taxa

[Hilmar Lapp]

I propose to make the following changes to tables Taxa and 
Bioentry_Taxa, ranked by decreasing priority (still, I'd like to 
make all of these changes).

1) Introduce UK constraints. This makes it much safer (and 
potentially faster due to amenability to caching) to find records). 
My suggestions for UK candidates are common_name, full_binomial (see 
2.), and ncbi_taxa_id. (There are no UKs yet.)

2) Introduce an additional column full_binomial ('full' in the 
bioperl Bio::Species sense: with ssp if applicable). I think this is 
going to be searched against quite frequently, and I just don't 
think it's very practical to always extract this out of full_lineage.

3) Collapse Taxa and Bioentry_Taxa into one and add Taxa_Id as a 
nullable FK to Bioentry. Having this additional association table 
doesn't add any functionality we'd like to use, but instead can only 
decrease performance and enforceability. Also, it necessitates 
otherwise pointless code in the adaptors.

4) For consistency, rename Taxa and all Taxa related columns to 
Taxon (Taxa is plural, all other table names are singular).

What do people think? Any outcries? If I make these changes, 
adaptors in different projects will have to be changed (I'll change 
the ones involved in Bioperl-db).

	-hilmar
--
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------