-
Notifications
You must be signed in to change notification settings - Fork 4.5k
[BEAM-191] Remove WindowedValue.valueInEmptyWindows #316
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
A value in empty windows expands to no values, so it can be dropped at any time, perhaps unintentionally. This has bitten runner authors, including Spark & Dataflow. While creating such a thing in memory is not automatically problematic, it is also not really useful. So this change removes it.
|
R: @bjchambers or @amitsela or really anyone. References are this dev thread which led to #179 fixing BEAM-189. It has also be re-filed as BEAM-230. |
|
Jenkins test is failing IT test, is it running with the DataflowPipelineRunner - meaning as a Dataflow service ? If so, I think you miss an update there.. (ed3676eda3f9031e): java.lang.NoSuchMethodError: org.apache.beam.sdk.util.WindowedValue.valueInEmptyWindows(Ljava/lang/Object;)Lorg/apache/beam/sdk/util/WindowedValue; ...... (9349dfed44f1940a): Workflow failed. Causes: (b529cf23ada96ef9): S04:WriteCounts/DataflowPipelineRunner.BatchTextIOWrite/Write/DataflowPipelineRunner.BatchWrite/View.AsSingleton/BatchViewAsSingleton/DataflowPipelineRunner.GroupByWindowHashAsKeyAndWindowAsSortKey/DataflowPipelineRunner.GroupByKeyAndSortValuesOnly/Read+WriteCounts/DataflowPipelineRunner.BatchTextIOWrite/Write/DataflowPipelineRunner.BatchWrite/View.AsSingleton/BatchViewAsSingleton/ParDo(IsmRecordForSingularValuePerWindow) failed. |
|
It seems that ValueInEmptyWindows is unused now, if we don't want to remove it yet, should we |
|
Ah, yes, good catch. I will wipe out |
|
Getting in just under 1-month-idle.. Any update here? |
|
Yes, there are some prerequisites that need to be addressed first. |
|
I am going to close this while I fix the bug in the runner. |
…uments (apache#316) * Began refactoring away from GetDocument proto which contains no `read_time` field and in general is a shortcut around `BatchGetDocuments` * Correctly instantiate snapshots for missing documents * Removed stale NotFound exception * Removed unnecessary empty list check * Linting fix * Expanded batch_get change to async classes * Updated variable name * Added get_batch to async classes * Improved consumption of async generators * Fixed test coverage * Fixed broken mock in test * Linting * Reverted the move of AsyncIter Co-authored-by: Craig Labenz <craiglabenz@google.com>
Be sure to do all of the following to help us incorporate your contribution
quickly and easily:
[BEAM-<Jira issue #>] Description of pull requestmvn clean verify. (Even better, enableTravis-CI on your fork and ensure the whole test matrix passes).
<Jira issue #>in the title with the actual Jira issuenumber, if there is one.
Individual Contributor License Agreement.
A value in empty windows expands to no values, so it can be dropped at
any time, perhaps unintentionally. This has bitten runner authors, including
Spark & Dataflow.
While creating such a thing in memory is not automatically problematic, it
is also not really useful. So this change removes it.