[ https://issues.apache.org/jira/browse/JENA-1391?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16187481#comment-16187481 ]
Andy Seaborne commented on JENA-1391: ------------------------------------- The {{Context}} isn't going to be dynamic? I can imagine wanting different settings on the union (e.g. query timeouts) to the base graphs so it is the way it is formed. That fact it diverges as the underlying ones if they change can be seen as a feature as well as a bug. I agree that a multimap is not feasible - other code will assume the type of the symbol value setting. So that leaves any defined combination - left-to-right is what I'd be least surprised at. Callbacks - I do not see a call to {{addCallback}} so we can probably solve any problems by removing the code as being an unnecessary complexity. (And the add/remove don't need synchronized.) Nowdays, {{Context}}s are static and copied to freeze them. > Add Convenience Methods to Dataset > ---------------------------------- > > Key: JENA-1391 > URL: https://issues.apache.org/jira/browse/JENA-1391 > Project: Apache Jena > Issue Type: Improvement > Components: ARQ > Affects Versions: Jena 3.4.0 > Reporter: Adam Jacobs > Assignee: A. Soroka > Priority: Trivial > > The Dataset interface could provide several convenience methods similar to > the Model interface, allowing usability of RDF quads on par with RDF triples. > Specific examples include, > # add(Dataset) > # remove(Dataset) > # union(Dataset) > # intersection(Dataset) > # difference(Dataset) > # isEmpty() > Following is a possible implementation of these methods. > {code:java} > default Dataset add(Dataset d) { > this.getDefaultModel().add(d.getDefaultModel()); > d.listNames().forEachRemaining(name -> > this.getNamedModel(name).add(d.getNamedModel(name))); > return this; > } > default Dataset remove(Dataset d) { > this.getDefaultModel().remove(d.getDefaultModel()); > d.listNames().forEachRemaining(name -> > this.getNamedModel(name).remove(d.getNamedModel(name))); > return this; > } > default Dataset union(Dataset d) { > return DatasetFactory.create().add(this).add(d); > } > default Dataset difference(Dataset d) { > Dataset output = DatasetFactory.create(); > > output.setDefaultModel(this.getDefaultModel().difference(d.getDefaultModel())); > this.listNames().forEachRemaining(name -> { > Model difference = > this.getNamedModel(name).difference(d.getNamedModel(name)); > if (!difference.isEmpty()) output.addNamedModel(name, difference); > }); > return output; > } > default Dataset intersection(Dataset d) { > Dataset output = DatasetFactory.create(); > > output.setDefaultModel(this.getDefaultModel().intersection(d.getDefaultModel())); > Set<String> names = this.names(); > names.retainAll(d.names()); > names.forEach(name -> { > Model intersection = > this.getNamedModel(name).intersection(d.getNamedModel(name)); > if (!intersection.isEmpty()) output.addNamedModel(name, > intersection); > }); > return output; > } > default Set<String> names() { > Set<String> names = new HashSet<>(); > this.listNames().forEachRemaining(names::add); > return names; > } > default boolean isEmpty() { > return this.asDatasetGraph().isEmpty(); > } > {code} -- This message was sent by Atlassian JIRA (v6.4.14#64029)