[
https://issues.apache.org/jira/browse/JENA-1445?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16283543#comment-16283543
]
Andy Seaborne edited comment on JENA-1445 at 12/8/17 2:13 PM:
--------------------------------------------------------------
Could you explain the benefit and explain "quad loading"? {{ja:data}} seems
like a clear general predicate to load RDF data.
TriG and N-Quads have Turtle and N-Triples as subsets (this was intentional by
the RDF 1.1 working group).
It does need to work with conneg - given a URL, the code does not know the MIME
type until the connection is opened.
Elsewhere in Jena, loading N-Quads into a graph causes a warning if named graph
quads are encountered and triples go into the graph. Having a general predicate
and also a (sub-property?) "must be triples" for named and default graph use to
(impl: add an StreamRDF that throws an error on quads) would protect
(transactional load only) against an attempt to put quads into a graph.
was (Author: andy.seaborne):
Could you explain the benefit and explain "quad loading"? {{ja:data}} seems
like a clear general predicate to load RDF data.
TriG and N-Quads have Turtle and N-Triples as subsets (this was intentional by
the RDF 1.1 working group).
It does need to work with conneg - given a URL, the code does not know the MIME
type until the connection is opened.
Elsewhere in Jena, loading N-Quads into a graph causes a warning if named graph
quads are encountered and triples go into the graph. Having a general predicate
and also a (sub-property?) "must be triples" for named graph use to (impl: add
an StreamRDF that throws an error on quads) would protect (transactional load
only) against an attempt to put quads into a graph.
> Add more error checking for the memory/general assemblers.
> ----------------------------------------------------------
>
> Key: JENA-1445
> URL: https://issues.apache.org/jira/browse/JENA-1445
> Project: Apache Jena
> Issue Type: Improvement
> Affects Versions: Jena 3.6.0
> Reporter: Andy Seaborne
> Fix For: Jena 3.7.0
>
>
> We aim to have in-memory dataset assemblers make "the best choice" between
> general and TIM implementations.
> But a description with a mixture of descriptions for models can pass
> (unrecognized proeprties are ignored).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
