[
https://issues.apache.org/jira/browse/BEAM-8550?focusedWorklogId=356146&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-356146
]
ASF GitHub Bot logged work on BEAM-8550:
----------------------------------------
Author: ASF GitHub Bot
Created on: 09/Dec/19 12:51
Start Date: 09/Dec/19 12:51
Worklog Time Spent: 10m
Work Description: je-ik commented on pull request #8774: [BEAM-8550]
Requires time sorted input
URL: https://github.com/apache/beam/pull/8774#discussion_r355431292
##########
File path:
sdks/java/core/src/test/java/org/apache/beam/sdk/transforms/ParDoTest.java
##########
@@ -2363,6 +2367,136 @@ public int hashCode() {
return getClass().hashCode();
}
}
+
+ @Test
+ @Category({
+ ValidatesRunner.class,
+ UsesStatefulParDo.class,
+ UsesRequiresTimeSortedInput.class,
+ UsesStrictTimerOrdering.class
+ })
+ public void testRequiresTimeSortedInput() {
+ // generate list long enough to rule out random shuffle in sorted order
+ int numElements = 1000;
+ List<Long> eventStamps =
+ LongStream.range(0, numElements)
+ .mapToObj(i -> numElements - i)
+ .collect(Collectors.toList());
+ testTimeSortedInput(numElements, pipeline.apply(Create.of(eventStamps)));
+ }
+
+ @Test
+ @Category({
+ ValidatesRunner.class,
+ UsesStatefulParDo.class,
+ UsesRequiresTimeSortedInput.class,
+ UsesStrictTimerOrdering.class,
+ UsesTestStream.class
+ })
+ public void testRequiresTimeSortedInputWithTestStream() {
+ // generate list long enough to rule out random shuffle in sorted order
+ int numElements = 1000;
+ List<Long> eventStamps =
+ LongStream.range(0, numElements)
+ .mapToObj(i -> numElements - i)
+ .collect(Collectors.toList());
+ TestStream.Builder<Long> stream = TestStream.create(VarLongCoder.of());
+ for (Long stamp : eventStamps) {
+ stream = stream.addElements(stamp);
+ }
+ testTimeSortedInput(numElements,
pipeline.apply(stream.advanceWatermarkToInfinity()));
+ }
+
+ @Test
+ @Category({
+ ValidatesRunner.class,
+ UsesStatefulParDo.class,
+ UsesRequiresTimeSortedInput.class,
+ UsesStrictTimerOrdering.class,
+ UsesTestStream.class
+ })
+ public void testRequiresTimeSortedInputWithLateData() {
+ // generate list long enough to rule out random shuffle in sorted order
+ int numElements = 1000;
+ List<Long> eventStamps =
+ LongStream.range(0, numElements)
+ .mapToObj(i -> numElements - i)
+ .collect(Collectors.toList());
+ TestStream.Builder<Long> input = TestStream.create(VarLongCoder.of());
+ for (Long stamp : eventStamps) {
+ input = input.addElements(TimestampedValue.of(stamp,
Instant.ofEpochMilli(stamp)));
+ if (stamp == 100) {
+ // advance watermark when we have 100 remaining elements
+ // all the rest are going to be late elements
+ input = input.advanceWatermarkTo(Instant.ofEpochMilli(stamp));
+ }
+ }
+ testTimeSortedInput(
+ numElements - 100,
+ numElements - 1,
+ pipeline.apply(input.advanceWatermarkToInfinity()),
+ // Only direct runner is guaranteed to correctly drop elements after
watermark coming
+ // from different (sub-)partitions of TestStream. Cannot reference
DirectRunner by
+ // class to avoid cyclic dependencies.
+
pipeline.getOptions().getRunner().getSimpleName().equals("DirectRunner"));
Review comment:
Other runners (flink to be specific) might handle watermark differently.
Only direct runner synchronously advances watermark, so that what you mark as
"late" in TestStream will really end up late. When running the same TestStream
with flink, the watermark might get updated at different times and thus the
data might not be late and the output is indeed different.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to 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: 356146)
Time Spent: 4.5h (was: 4h 20m)
> @RequiresTimeSortedInput DoFn annotation
> ----------------------------------------
>
> Key: BEAM-8550
> URL: https://issues.apache.org/jira/browse/BEAM-8550
> Project: Beam
> Issue Type: New Feature
> Components: beam-model, sdk-java-core
> Reporter: Jan Lukavský
> Assignee: Jan Lukavský
> Priority: Major
> Time Spent: 4.5h
> Remaining Estimate: 0h
>
> Implement new annotation {{@RequiresTimeSortedInput}} for stateful DoFn as
> described in [design
> document|https://docs.google.com/document/d/1ObLVUFsf1NcG8ZuIZE4aVy2RYKx2FfyMhkZYWPnI9-c/edit?usp=sharing].
> First implementation will assume that:
> - time is defined by timestamp in associated WindowedValue
> - allowed lateness is explicitly zero and all late elements are dropped
> (due to being out of order)
> The above properties are considered temporary and will be resolved by
> subsequent extensions (backwards compatible).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
