Re: SPARQL performance for ORDER BY on large datasets

2009/8/27 Andreas Langegger <al@jku.at>:
> On Aug 27, 2009, at 12:11 AM, Seaborne, Andy wrote:
>>
>> it's a tradeoff of generality and data validity assumptions.
>
> and generality of RDF compared to multiple tables with their own indexable
> attributes in RDBMS.
>
> I'm wondering for some time already if there is any triple store that allows
> to define custom indexes on special predicates or subsets of the whole set
> of triples/quads? All the existing stores I know index over all triples in
> different combinations (spo, pso, ...). Is there any research going on
> towards partial indexes over user-defined subsets of triples? E.g. an index
> over all xsd:dateTime literals.

Mulgara indexes through all of the XSD datatypes, including
xsd:dateTime literals. It's mostly used for optimizing when finding
intervals.

Unfortunately, there is no optimization to take advantage of this when
doing an ORDER BY on a literal field (though now that it's been
mentioned, I'll have to get to it).

Regards,
Paul Gearon

Received on Friday, 28 August 2009 11:29:38 UTC