[BioSQL-l] problem loading NCBI_taxonomy database into BioSQL bioseqdb

[Hilmar Lapp]

On Sep 3, 2008, at 6:44 PM, Nick Matzke wrote:

> Well, I'm not sure what I did, but some combination of these things  
> seems to have worked.

Great we got you going!

> [...]
> 3. (Make sure you have an empty version of the db, at least for me I  
> got errors if I had already loaded sequences etc. into it...I got  
> errors like this:
>
> ==========================================
> note: node (28;331111;27;species;;) is retired; failed to delete:  
> Cannot delete or update a parent row: a foreign key constraint fails  
> (`bioseqdb/bioentry`, CONSTRAINT `FKtaxon_bioentry` FOREIGN KEY  
> (`taxon_id`) REFERENCES `taxon` (`taxon_id`))
> note: node (70;300268;69;species;;) is retired; failed to delete:  
> Cannot delete or update a parent row: a foreign key constraint fails  
> (`bioseqdb/bioentry`, CONSTRAINT `FKtaxon_bioentry` FOREIGN KEY  
> (`taxon_id`) REFERENCES `taxon` (`taxon_id`))
> note: node (77;3002
> ==========================================
>

These aren't fatal, right? What is basically means is that your  
sequences referenced taxa that are not yet or not anymore in the NCBI  
taxonomy download.

The script doesn't yet process the change log that NCBI also produces.  
So if two nodes get merged into one, or one gets split into two new,  
the script can't migrate the data that you already have. Nodes that  
are in the database but not in the taxonomy dump from NCBI will be  
considered retired, and the script tries to delete them if there  
aren't any sequences yet pointing to them.

	-hilmar
-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================