On 14/12/12 12:59, markh wrote:
Hello
I have a case I cannot see how to support, any advice gratefully
received.
I have a graph which has a collection of properties which enable me to
logically subset the graph to give me a graph of currently valid
statements. This concept of validity is an emergent property,
constantly changing as the graph evolves.
There are a number of queries I would like to run against the 'valid'
graph, for example examining complex property paths.
So, I would like to create a 'valid graph' on the fly, run queries
against it, then dump this graph. All other work happen with respect
to the full graph.
I thought I could use SPARQL subquery functionality, but it seems to
me that a subquery returns results, not a graph.
I hoped to have a query a bit like this:
SELECT ?someinfo
WHERE{
GRAPH <transient valid graph> {
?someinfo has conditions
}
{construct{
foo}
where
{validity constraints}
as <transient valid graph>}
}
Is there a part of the Jena/Fuseki implementation which might support
this?
SPARQL Query has sub-SELECTs, not sub-CONSTRUCTs. It does not need it
because the information used to make the CONSTRUCT graph comes from
variables of a WHERE pattern. The table produced by {SELECT} should be
enough.
What you might do is use a combination of SPARQL Update and SPARQL
Query. Do an update to create a graph in the dataset, query it, and
then drop the graph.
Andy
many thanks
mark
