[
https://issues.apache.org/jira/browse/BEAM-13078?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17549686#comment-17549686
]
Danny McCormick commented on BEAM-13078:
----------------------------------------
This issue has been migrated to https://github.com/apache/beam/issues/21260
> Python DirectRunner does not emit data at GC time
> -------------------------------------------------
>
> Key: BEAM-13078
> URL: https://issues.apache.org/jira/browse/BEAM-13078
> Project: Beam
> Issue Type: Bug
> Components: sdk-py-core
> Reporter: Zachary Houfek
> Priority: P1
>
> The following should succeed but does not:
> {code:java}
> test_options = PipelineOptions(flags=['--allow_unsafe_triggers'])
> with TestPipeline(options=test_options) as pipeline:
> pcoll = (
> pipeline
> | beam.Create([(1, 1), (1, 2), (1, 3), (1, 4)])
> | WindowInto(
> window.GlobalWindows(),
> trigger=trigger.AfterCount(5),
> accumulation_mode=trigger.AccumulationMode.ACCUMULATING)
> | beam.GroupByKey())
> assert_that(pcoll, equal_to([(1, [1, 2, 3, 4])])){code}
> However, it currently fails, because pcoll will be empty. It appears that
> the Direct Runner drops data if the trigger never fired.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
