[
https://issues.apache.org/jira/browse/TINKERPOP3-699?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stephen mallette updated TINKERPOP3-699:
----------------------------------------
Issue Type: Improvement (was: Bug)
I guess the guiding principle for TP3 has been to provide solutions that will
scale on multi-billion+ edge graphs. We've tried to do that everywhere that it
makes sense to do so. The old GraphSON that produced "valid JSON" didn't work
so well in that regard and we determined a need for a JSON based format that
was better represented as an adjacency list like Gryo - thus GraphSON became
consistent with Gryo and would work on any size graph. I think the fact that
the GraphSON is not utterly valid is a reasonable consequence for this change
and not terribly harmful, unless you just have a very small graph you intend to
ship around.
That said, for that use case it might be useful to get some valid JSON from
{{writeGraph()}} so that its easily consumable by other libs An option on the
{{GraphSONWriterBuilder}} to {{wrapAdjacencyListInObject}} would probably be an
acceptable addition. That would basically take "what we have now" and do:
{code}
{ "graph": [
...what we have now...
]}
{code}
The default would be to remain unwrapped. An option on the
{{GraphSONReaderBuilder}} would need to be added to signal that the JSON was in
this format but perhaps it- could just be detected??
Anyway, if you'd like to submit a pull request around an idea like this, please
feel free to do so. Tests should be added to {{IoTest}} to validate the
operations of this features.
> GraphSON writeGraph not producing valid json object
> ---------------------------------------------------
>
> Key: TINKERPOP3-699
> URL: https://issues.apache.org/jira/browse/TINKERPOP3-699
> Project: TinkerPop 3
> Issue Type: Improvement
> Components: documentation
> Reporter: Jim Lloyd
> Assignee: stephen mallette
> Fix For: 3.0.0.GA
>
>
> The following code in the M9-rc3 console produces an incorrect t.json file:
> gremlin> graph = TinkerFactory.createModern()
> ==>tinkergraph[vertices:6 edges:6]
> gremlin> g = graph.traversal()
> ==>graphtraversalsource[tinkergraph[vertices:6 edges:6], standard]
> gremlin> f = new FileOutputStream("t.json")
> ==>java.io.FileOutputStream@7927bd9f
> gremlin> mapper =
> graph.io(IoCore.graphson()).mapper().embedTypes(true).create()
> ==>org.apache.tinkerpop.gremlin.structure.io.graphson.GraphSONMapper@63f34b70
> gremlin>
> graph.io(IoCore.graphson()).writer().mapper(mapper).create().writeGraph(f,
> graph)
> ==>null
> gremlin> f.close()
> ==>null
> The file is a stream of independent vertices rather than a single json object
> following the GraphSON schema. I.e. the output looks like the this:
> {"id":1,"label":"person", ... }
> {"id":2,"label":"person", ... }
> {"id":3,"label":"software", ... }
> Where ... represents omitted fields.
> A separate issue but related issue is that the documentation examples use
> "CoreIo" but the actual name of that class is "IoCore", as shown in the
> console session above.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
