[Bioperl-l] Re: Bioperl-DB question

Hilmar Lapp hlapp at gmx.net
Mon Oct 25 12:28:35 EDT 2004


Michael,

the location of the error suggests that a statement handle was returned  
undef and hence failed to prepare. Is it the very first term that fails  
or do several terms succeed until one fails? If you supply --printerror  
as a command-line argument a preceding DBI error should become visible.  
If so, please copy&paste verbatim and post it.

When you say you instantiated the Oracle schema and view, does this  
mean you instantiated the view-based API for strict Biosql compliance,  
or did you install the alias-based API? I.e., did you run  
BS-create-Biosql-API.sql (alias-based) or BS-create-Biosql-API2.sql  
(view-based). Bioperl-db can't deal with the view-based yet due to  
column naming. If that is the API you need though (instead of the  
alias-based) then it is relatively easy to add; it will become another  
driver. Alternatively, as an immediate fix install the alias-based API  
under a different user, after prefixing the referenced tables with the  
schema owner (and of course then you need to grant access to the tables  
which is not pretty).

If you instantiated the alias-based API and --printerror does not cause  
a preceding DBI error to be printed, supply --debug, capture the output  
(extensive!) in a file, and send it to me. Also, please use the  
unaltered distribution for this.

	-hilmar

On Monday, October 25, 2004, at 07:55  AM, michael.g.moore at gsk.com  
wrote:

>
> Hi,
>
> We've instantiated the Oracle schema and view for the BioPerl-DB tree  
> (latest CVS checkout ~1 week old and using ) and I am having some  
> problems
>
> I've managed to get the NCBI taxonomy to load.
>
> When I try the Gene Ontology data I use the following program
>
> ./load_ontology --dbname USTST240 --driver=Oracle --dbuser=biosql  
> --dbpass=biosql --safe --updobsolete --namespace "Gene Ontology"  
> --format go --fmtargs "-def_file,GO.defs" function.ontology  
> component.ontology process.ontology
>
> I get a
>
> 'Cant call method 'bind_param" on an undefined value at  
> /blah-blah-blah/BasePersistenceAdaptor.pm line (in mine its 948 but I  
> added some debugging so I think its around 942 in the original code)
>
> One of the pieces of SQL that is being constructed from this part of  
> the module does not seem to make sense
>
> select  
> term.term_oid,term.indetifier,term.name,term,definition,term.is_obsolet 
> e,term.ont_oid
> from term
> where indentifier = ?
>
> The reason that this seems wrong is that the columns names are from  
> the SYMGENE schema and should not be available to the biosql view that  
> was constructed  on the SYMGENE schema.
>
> Any help would be appreciated.
>
> Mike
>
>
>
>
> Mike Moore
>
> BioInformatics Engineering and Integration
> GlaxoSmithKline
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------





More information about the Bioperl-l mailing list