Re: [Dbpedia-discussion] SPARQL access to DB appears to be corrupted

[Peter A. <ans...@gm...>]

2009/11/5 Kingsley Idehen <ki...@op...>:
> Ted Thibodeau Jr wrote:
>> Hello, all --
>>
>> I've dug a little further...
>>
>> The issue is that the English data sets are pretty clean, but the
>> non-English data sets aren't ... and we loaded all of them into
>> a single graph, so now the discrepancies are *very* visible.
>>
>> We're going to drop the whole kit and reload, partitioning by lang
>> this time.
>>
>> There will be dirty output for a bit, hopefully not more than
>> a few hours.
>>
>> Apologies for the inconvenience.
>>
>> Be seeing you,
>>
>> Ted
>>
>>
>>
> Just to add to the above.
>
> We are going to put each of the language specific data sets in their own
> Named Graphs. The main DBpedia Graph IRI: <http://dbpedia.org> will be
> the host of the "en" data set (how its been to date). This also means
> the default HTML+RDFa based exploration/browser pages will be scoped to
> the same Graph (as it was pre 3.4).

What will the GRAPH URI's look like? I tried a query on the
dbpedia.org endpoint to see if it could find out what graphs were
present but it refuses to perform the query because it is estimated to
take too long. [1]

The query was : select distinct ?graph where { graph ?graph { [] ?p ?o
} } LIMIT 100

> The SPARQL endpoint will support access to all the data across Named
> Graphs, if your SPARQL patterns aren't scoped to any specific Graph IRI(s).

Would this involve the pagelinks set as well? How do I set this
functionality up on my own virtuoso installation by the way?

Cheers,

Peter

[1] http://preview.tinyurl.com/ygctx6k