[ 
https://issues.apache.org/jira/browse/TINKERPOP-2006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16563574#comment-16563574
 ] 

Svante Schubert commented on TINKERPOP-2006:
--------------------------------------------

1) Yes, "Collection<String> vertexKeySet.." and same for the edge can still be 
made final, good catch.

2) No, I did not make any performance test, do you have one for Tinkerpop, what 
would you suggest? 
In any case, I leave it to your taste to adopt it to your desire 
\{#emotions_dlg.wink}

3) There is a key declared in the header of the for loop, what is your 
suggestion? I do not fully understand your suggestion. Again, I have no strong 
feelings, in case you like to adopt the source. I only want to remove the 
reported problem.

4) The scenario comes from creating a graph with Tinkerpop using same-named 
property on edge and vertex and serializing it to GraphML. A regression test 
might be the extending an existing Graph example with the color property for an 
edge and vertex, which would be the correct location/class to create in 
Tinkerpop such a regression test?

> GraphML serialization invalid if a vertex and edge have similar named property
> ------------------------------------------------------------------------------
>
>                 Key: TINKERPOP-2006
>                 URL: https://issues.apache.org/jira/browse/TINKERPOP-2006
>             Project: TinkerPop
>          Issue Type: Bug
>          Components: io
>    Affects Versions: 3.3.3
>            Reporter: Svante Schubert
>            Priority: Trivial
>
> I have created a graph using the Tinkerpop Graph and required for Gephi 
> visualization a property called color on edges and vertices to be coloured.
> The current gremlin-core serialization creates the following two key elements 
> in GraphML
> <key id="color" for="node" attr.name="color" attr.type="string"></key>
> <key id="color" for="edge" attr.name="color" attr.type="string"></key>
> the id attribute is an internal name, but have to be different.
> I would suggest a patch to check for an intersection of the edge & vertices 
> keys and add for those keys an additional differentiating letter like:
> <key id="colorV" for="node" attr.name="color" attr.type="string"></key>
> <key id="colorE" for="edge" attr.name="color" attr.type="string"></key>
>  
> Going to provide a pullrequest on GitHub.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to