Andy Seaborne created JENA-1069:
-----------------------------------
Summary: Simpify the internal naming of the default graph (Quad
constants)
Key: JENA-1069
URL: https://issues.apache.org/jira/browse/JENA-1069
Project: Apache Jena
Issue Type: Improvement
Components: ARQ
Affects Versions: Jena 3.0.0
Reporter: Andy Seaborne
Assignee: Andy Seaborne
There are two different constants for the default graph for what seemed at the
time, slightly different situations. The idea was that quads from the default
graph of a dataset would be {{Quad.defaultGraphIRI}} and quads created by
parsers, would be {{Quad.defaultGraphNodeGenerated}}. The idea was that the
output of the parser was "default graph" but it was being redirected into a
named graph. This does not now seem useful.
{{Quad.defaultGraphIRI}} = {{<urn:x-arq:DefaultGraph>}}
{{Quad.defaultGraphNodeGenerated}} = {{<urn:x-arq:DefaultGraphNode>}}
In SPARQL, the application is using {{GRAPH <urn:x-arq:DefaultGraph>}}. Using
{{<urn:x-arq:DefaultGraphNode>}} wil not work consistently.
The proposal is to combine these two as {{Quad.defaultGraphIRI}}, possible
migrate to {{Quad.defaultGraph}}, or even a class or enum for the graph naming
constants.
These constants should never have been written to disk in database, certainly
not {{<urn:x-arq:DefaultGraphNode>}}, but just to make sure, the operation
{{Quad.isDefaultGraph(Node)}} can take the old name into account for a release
cycle.
Nearly all the changes are in tests where the expected output has been
explicitly written into test results.
There is another, little used, constant -- {{Quad.tripleInQuad}} -- for when a
triple is in a quad just to be transported around, meaning "meaningless",
"unknown" or "doesn't matter". This would not change. It exists to be
different from {{Quad.defaultGraphIRI}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
