[Biopython-dev] [Bug 2833] Features insertion on previous bioentry_id

[Cymon Cox]

2009/6/3 Peter <biopython at maubp.freeserve.co.uk>

> On Tue, Jun 2, 2009 at 9:29 PM, Cymon Cox <cy at cymon.org> wrote:
> >
> > Whoa, I see now that in Loader._load_bioentry_table that if the
> > rec.annotations["gi"] is missing, it gets filled with the
> accession.version:
> >
> >        if "gi" in record.annotations :
> >            identifier = record.annotations["gi"]
> >        else :
> >            identifier = record.id
> >
> > So biopythons BioSQL identifiers are not equivalent to GenBank
> identifiers.
> > I wonder why this is done and identifier is not just left NULL, and the
> > unique constraint maintained by accession/version...
> >
>
> Remember, it isn't just GenBank files that get imported into BioSQL.
> While the record.id is the accession.version when loading a GenBank
> file, this is not the case in general.
>
> Consulting the CVS log, this was changed BioSQL/Loader/py revision
> 1.33 to cope with loading a FASTA file into a BioSQL database (Bug
> 2425). Presumably I was trying to mimic the BioPerl loading of FASTA
> files. Before this change, the bioentry.identifier was taken as the GI
> number if available.
>
> i.e. This change wasn't anything directly to do with the uniqueness rules.


Thanks Peter.

Yes, it seems to have been done to mimic BioPerl - but I'm still curious as
to why it is done at all...

Anyway, I seem to be chasing my tail here:
http://bugzilla.open-bio.org/show_bug.cgi?id=2681#c5

Cheers, C.
--