[
https://issues.apache.org/jira/browse/SPARK-3789?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14989855#comment-14989855
]
Sean Owen commented on SPARK-3789:
----------------------------------
[~sesshomurai] I'm more commenting on the lack of anyone stepping forward to
_maintain_ this functionality _in Spark_. For better or worse, I don't see
people who can shepherd and own this code in Spark anymore, and they're the
people you're addressing. I'd be most happy to be proved wrong, and
Ankur/Joseph pop up to merge this. I'm replying to tell you what I see rather
than leave nothing at all.
Back on the interesting separate question of the utility of graph analytics --
you're right that the lack of activity doesn't mean it's not good or useful.
FWIW virtually none of our customers use GraphX, and we interact with a pretty
good cross section of Big Companies. Many of the useful functions you identify
are not solved as graph problems in my experience, even if they could be (e.g.
recommenders, also viewed). Still, that's just a different POV from BigCo big
data use cases, filtered through my biases. I personally have a hard time
reconciling this with claims that it's very important relative to other
analytics or ML. It's still no reason to not have PySpark GraphX, but is an
explanation for why we somehow don't?
> [GRAPHX] Python bindings for GraphX
> -----------------------------------
>
> Key: SPARK-3789
> URL: https://issues.apache.org/jira/browse/SPARK-3789
> Project: Spark
> Issue Type: New Feature
> Components: GraphX, PySpark
> Reporter: Ameet Talwalkar
> Assignee: Kushal Datta
> Attachments: PyGraphX_design_doc.pdf
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
