[
https://issues.apache.org/jira/browse/BEAM-6120?focusedWorklogId=170160&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-170160
]
ASF GitHub Bot logged work on BEAM-6120:
----------------------------------------
Author: ASF GitHub Bot
Created on: 28/Nov/18 09:18
Start Date: 28/Nov/18 09:18
Worklog Time Spent: 10m
Work Description: robertwb commented on a change in pull request #7127:
[BEAM-6120] Support retrieval of large gbk iterables over the state API.
URL: https://github.com/apache/beam/pull/7127#discussion_r236997715
##########
File path: model/pipeline/src/main/proto/beam_runner_api.proto
##########
@@ -588,6 +588,10 @@ message StandardCoders {
// of the element
// Components: The element coder and the window coder, in that order
WINDOWED_VALUE = 8 [(beam_urn) = "beam:coder:windowed_value:v1"];
+
+ // Encodes an iterable of elements, some of which may be stored in state.
+ // Components: Coder for a single element.
+ STATE_BACKED_ITERABLE = 9 [(beam_urn) =
"beam:coder:state_backed_iterable:v1"];
Review comment:
Yes, such composition is possible if one extends coder protocol to have the
notion of lazy decoding and encoding for component coders, with the all
supporting infrastructure, in every SDK. However, I don't think we have the
bandwidth to tackle that complexity now, and 95% of the time this is useful is
the case of consuming a values iterable from a GBK from the runner. This will
allow us to at least have parity with the non-portability framework, and we can
look at going beyond that (if the underlying runners support such large records
once we pass them across the FnAPI boundary) later.
I'm happy to annotate this as experimental if that would help.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 170160)
Time Spent: 1h 50m (was: 1h 40m)
> Support retrieval of large gbk iterables over the state API.
> ------------------------------------------------------------
>
> Key: BEAM-6120
> URL: https://issues.apache.org/jira/browse/BEAM-6120
> Project: Beam
> Issue Type: Improvement
> Components: sdk-py-harness
> Reporter: Robert Bradshaw
> Assignee: Robert Bradshaw
> Priority: Major
> Time Spent: 1h 50m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
