[
https://issues.apache.org/jira/browse/TINKERPOP-3214?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18066117#comment-18066117
]
ASF GitHub Bot commented on TINKERPOP-3214:
-------------------------------------------
spmallette opened a new pull request, #3327:
URL: https://github.com/apache/tinkerpop/pull/3327
https://issues.apache.org/jira/browse/TINKERPOP-3214
Prevented duplicate identifiers by using the same pattern for tracking
vertex properties as is used for vertices and edges.
Not sure how controversial this change is for 3.7.x but basically this
change just tracks vertex properties just as we track vertices and edges. The
references to these are already in memory so it seems like the cost would be
negligible to do it this way. Can't think of cases where this would creates any
sort of breaking change to any behavior as I don't see how you'd rely on
duplicates for anything reasonable you could be doing. Seems best to just fix
it, but open to moving to another branch if I'm missing something here or if
there's a performance angle I've not considered.
VOTE +1
> TinkerGraph VertexProperty Id duplication
> -----------------------------------------
>
> Key: TINKERPOP-3214
> URL: https://issues.apache.org/jira/browse/TINKERPOP-3214
> Project: TinkerPop
> Issue Type: Bug
> Components: tinkergraph
> Affects Versions: 3.7.4
> Reporter: Stephen Mallette
> Priority: Blocker
>
> I assume this i just TinkerGraph. Tested in 3.8.0, but maybe this has been
> around longer?
> {code}
> gremlin> g = TinkerFactory.createAirRoutes().traversal()
> ==>graphtraversalsource[tinkergraph[vertices:3749 edges:57645], standard]
> gremlin> g.V().properties().hasId(0)
> ==>vp[date->2025-10-22 13:56:...]
> gremlin> g.V(4).property('city','newname')
> ==>v[4]
> gremlin> g.V(4).properties('city').id()
> ==>0
> gremlin> g.V().properties().hasId(0)
> ==>vp[date->2025-10-22 13:56:...]
> ==>vp[date->2025-10-22 13:56:...]
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)