[redland-dev] rasqal - graph queries

Dave Beckett dave at dajobe.org
Tue Dec 15 03:51:17 CET 2009


A little update to this; Steve Harris is looking in to this to fix it
against the new rasqal (pre)release.  The query structure change and 4store
needs to get the graph information a slightly different way.  The very
latest rasqal SVN also has a different ABI/API as I said in other emails to
redland-dev, and that's also being fixed.

Dave

jakobitsch juergen wrote:
> hi, i was talking about the 4store tests that fail,
> 
> 4store however can be installed.
> 
> now if run the below query against the 4store sparql endpoint
> i get about the following result :
> 
> <sparql>
>  <head>
>   <variable name="g"/>
>  </head>
>  <!--
>  parser warning : Variable g was selected but is unused in the query. at line 1
>  -->
> </sparql>
> 
> 
> now if i remove the rasqal and raptor libs from /usr/local/lib and reinstall them from 
> opensuse 11.2 repository and then rebuild 4store i get my graph-uris as expected.
> 
> the datasource : i'm just putting some triples into different graphs - nothing fency 
> and surely without error.
> 
> wkr www.turnguard.com
> 
> 
> ----- Original Message ----
> From: Dave Beckett <dave at dajobe.org>
> To: jakobitsch juergen <tschyrgie at yahoo.com>
> Cc: redland-dev at lists.librdf.org
> Sent: Mon, December 7, 2009 7:23:31 PM
> Subject: Re: [redland-dev] rasqal - graph queries
> 
> On Sun, 6 Dec 2009, jakobitsch juergen wrote:
>> hi,
>>
>> i just installed 4store which runs a couple of test.
>> using the latest rasqal build from svn graph queries
>> like
>>
>> SELECT DISTINCT ?g WHERE { GRAPH ?g { ?s ?p ?o } }
>>
>> fail.
>>
>> they work howevery when i install rasqal from my opensuse 11.2 repository
>> and rebuild 4store.
>>
>> is this a known bug?
> 
> I don't think you've given enough information here - what do you mean by 
> 'fail'?  Do they fail to parse - syntax error - or fail to run with soem 
> runtime error, or return no results?  What datasets are you using?
> 
> Dave
> 
> 
>       
> 



More information about the redland-dev mailing list