Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] SPARQL - getting all predicates for an ontology

manuelso manuelso at stanford.edu
Fri Jun 28 10:29:22 PDT 2013


Simon,

I just realized that this went unanswered. Sorry.

So in the RDF Download Service you are downloading the RDF serialization of the OWL ontology as the ontology admin uploaded it to BioPortal. In the SPARQL endpoint we add few extra statements. For instance we generate prefLabel statements for classes that are missing them. We either generate the label with last fragment of the class URI or we take a rdfs:label literal to generate. All the prefLabels in the bioportal namespace are labels that we asserted in the ontology graph. 

Manuel

On May 21, 2013, at 1:37 AM, Simon Kocbek <simon at dbcls.rois.ac.jp> wrote:

> Hi Manuel,
>  
> thank you for your answer.
>  
> Since I still wish to count numbers of specific/all predicates in a single ontology, I’ve decided to download ontologies as RDF/XML and parse them myself. I use the download service as described here:
> http://www.bioontology.org/wiki/index.php/BioPortal_REST_services#RDF_Download_Service
>  
> However, I have another question. For example, according to SPARQL query, FBdv contains 132 prefLabel relations/predicates (http://bioportal.bioontology.org/metadata/def/prefLabel). When I search for this relation in the downloaded ontology (vid: 1016), I cannot find any. Some other relations that I am interested in (e.g., subclassOf, label) agree with the SPARQL query.
>  
> Could you please explain this behaviour?
>  
> Thanks,
> Simon
>  
> From: manuelso [mailto:manuelso at stanford.edu] 
> Sent: Monday, May 20, 2013 12:11 PM
> To: Simon Kocbek
> Cc: support at bioontology.org
> Subject: Re: [bioontology-support] SPARQL - getting all predicates for an ontology
>  
> Hi Simon,
>  
> You are right we should notify the users about incomplete answers. We have an action item for this that should be addressed soon. However we can only do this on the UI front-end and not on programatic queries because HTTP SPARQL program does not have a neat way to handle this. This is a known issue in or community see related discussion here [1].
>  
> Re: Documentation on incomplete results see "Partial or Incomplete results" here [2]
>  
> Best,
>  
> Manuel
>  
> [1] http://lists.w3.org/Archives/Public/public-lod/2013Apr/0198.html
> [2] http://www.bioontology.org/wiki/index.php/SPARQL_BioPortal
>  
>  
>  
> On May 14, 2013, at 8:39 PM, Simon Kocbek <simon at dbcls.rois.ac.jp> wrote:
> 
> 
> Hi Manuel,
>  
> thank you for your answer.
>  
> Do you recommend any workaround for this limitation? There are also other queries that won’t work correctly as a result. For example:
>  
> “SELECT ?p, (COUNT(?p) as ?pCount)
> WHERE {
>   GRAPH <http://bioportal.bioontology.org/ontologies/RCD> {
>     ?s ?p ?o .
>   }
> }
> GROUP BY ?p”
>  
> also returns only 1200000 predicates, however there is no indication that the query had been interrupted. I believe users should be informed about this somehow.
>  
> Are these kind of limitations documented somewhere?
>  
> Thanks,
> Simon
>  
> From: manuelso [mailto:manuelso at stanford.edu] 
> Sent: Wednesday, May 15, 2013 12:59 AM
> To: Simon Kocbek
> Cc: support at bioontology.org
> Subject: Re: [bioontology-support] SPARQL - getting all predicates for an ontology
>  
> Hi Simon,
>  
> Unbound predicate queries are very inefficient and hit query timeouts. They tend to not return all the answers. Probably the query is being stopped by server always at the same point returning the same number.
>  
> It is not a bug it is just a configuration we have to assure that no query uses too many resources.
>  
> Manuel
>  
>  
> On May 14, 2013, at 12:38 AM, "Simon Kocbek" <simon at dbcls.rois.ac.jp> wrote:
> 
> 
> 
> Hi BioPortal,
>  
> I have a problem with getting the number of predicates for larger ontologies through the SPARQL interface.
>  
> The following query:
>  
> SELECT (COUNT(?p) as ?pCount)
> WHERE {
>   GRAPH <http://bioportal.bioontology.org/ontologies/[ontology]> {
>     ?s ?p ?o .
>   }
> }
>  
> will always return the same number (i.e., 1200000) for large [ontology] values (e.g, SNOMEDCT, RCD, NCBITAXON).
>  
> Is this a bug?
>  
> Thanks,
> Simon
>  
> _______________________________________________
> bioontology-support mailing list
> bioontology-support at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/bioontology-support
>  
>  

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20130628/23332f0f/attachment-0001.html>


More information about the bioontology-support mailing list