Aldo Bucchi wrote:
Kingsley,
On Fri, Dec 4, 2009 at 6:14 PM, Kingsley Idehen <[email protected]> wrote:
Aldo Bucchi wrote:
Hi,
I am operating over the RDF_QUAD table directly and I find myself in
the need to transform all values obtained from column O to the
explicit SQL type in a programmatic way.
Is there any stored procedure I can use for this matter?
A quick glance at "sparql.sql" did not help and (re)implementing the
heuristics to access the RDF Object storage table does not seem like a
good idea ;)
Thanks!
A
Aldo,
Please start with what you are trying to achieve first. Don't go to that
table directly.
What we are trying to achieve is to generate some relational tables by
"flattening" part of the graph. We need this to perform aggregate
analysis that does not run fast enough in SPARQL ( probably lots and
lots of self joins as relations are spread accross several statements
in a corpus of over 100 million triples ).
We did try other approaches but this one seems to work best.
Ah, why I end up with the specific need for this "procedure" is
because we are generating the "plain" aggregate by querying the quad
store ( again for efficiency ) and we end up with iri_id_8 and rdf box
values down the road.
I basically need the analog to id_to_iri() but for the quad Object column.
I understand breaking the abstraction layer is not adviced but this is
100% read-only and we won't mess up the caches. It's just cheaper for
us to work this way.
Regards,
A
I assume the last post re. link to use of SQL inside SPARQL expressions
shows the way?
Kingsley
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO OpenLink Software Web: http://www.openlinksw.com
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
