[
https://issues.apache.org/jira/browse/BEAM-91?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17134200#comment-17134200
]
Pedro Teixeira commented on BEAM-91:
------------------------------------
Hello, it's not entirely clear (to someone who is trying to catch up with Beam)
what's the status of this one here.
I am reading this document, so will cross link it here too, if helps someone
else:
[https://docs.google.com/document/d/14WRfxwk_iLUHGPty3C6ZenddPsp_d6jhmx0vuafXqmE/edit#]
Some questions came to mind:
1) Is this still on the roadmap? Are there people working on this? Is Tyler the
owner of this design decision?
2) Any related work in the context of other runners worth reading / researching
about?
3) Perhaps given more real life use cases (since 2016, year of this issue) is
it still worth the research + impl effort? Or workarounds (e.g upserts in
sinks) are sufficient for most? perhaps "common strategies" for the lack of
retractions could be added to the design document for reference to better
understand known trade-offs?
4) Are there any immediate tasks community contributors could help with on this
topic?
5) Going forward, is the current idea to break API / SDK for current pipelines?
or most preferable add new optional interfaces for metadata & handlers?
> Retractions
> -----------
>
> Key: BEAM-91
> URL: https://issues.apache.org/jira/browse/BEAM-91
> Project: Beam
> Issue Type: New Feature
> Components: beam-model
> Reporter: Tyler Akidau
> Priority: P2
> Labels: stale-P2
> Original Estimate: 672h
> Remaining Estimate: 672h
>
> We still haven't added retractions to Beam, even though they're a core part
> of the model. We should document all the necessary aspects (uncombine,
> reverting DoFn output with DoOvers, sink integration, source-level
> retractions, etc), and then implement them.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
