hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Purtell <apurt...@apache.org>
Subject Re: Using SPARQL against HBase
Date Wed, 31 Mar 2010 16:27:10 GMT
HBase has nice properties for efficiently storing, for example, a sparse adjacency representation
of a graph, very large graphs. I'm sure it could be used to store an enormous number of RDF
triples. But this is a long long way from something that can respond to SPARQL queries. The
RDF store I've seen that most closely resembles HBase is "bigdata": http://www.bigdata.com/bigdata/blog/.
I've only heard of it, not looked at it.

> From: Stack
> The in-memory graph would hold all data or just metadata?
> You might look at IHBase in the 'indexed' contrib to see how
> it loads an index on region open (It subclasses hbase so it
> can catch key transistions)
> Why HBase and not a native graph database?


View raw message