[
https://issues.apache.org/jira/browse/BEAM-6120?focusedWorklogId=170396&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-170396
]
ASF GitHub Bot logged work on BEAM-6120:
----------------------------------------
Author: ASF GitHub Bot
Created on: 28/Nov/18 19:58
Start Date: 28/Nov/18 19:58
Worklog Time Spent: 10m
Work Description: lukecwik 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_r237239144
##########
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, I thought about the Data API case to not require chunked values and
found out a bunch of subtle complexity there related to making messages still
fit within 2gbs since we may make a bad coder choice such as
iterable<length_prefix<blob>> instead of length_prefix<iterable<blob>>.
Is there a significant difference in the implementation complexity of having
the support for reading state within the length prefix coder while having
support for lazy decoding within the iterable portion? (haven't been able to
look through the code to understand the nuance)
Yes, marking this experimental to move the implementation forward will make
sense but will defer to your judgement and hopefully the ML thread is fruitful.
----------------------------------------------------------------
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: 170396)
Time Spent: 2h 10m (was: 2h)
> 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: 2h 10m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
