robertwb commented on a change in pull request #11835:
URL: https://github.com/apache/beam/pull/11835#discussion_r432093237
##########
File path: sdks/python/apache_beam/transforms/trigger_test.py
##########
@@ -518,6 +519,28 @@ def format_result(k_v):
'B-3': {10, 15, 16},
}.items())))
+ def test_never(self):
+ with TestPipeline() as p:
+
+ def construct_timestamped(k_t):
+ return TimestampedValue((k_t[0], k_t[1]), k_t[1])
+
+ def format_result(k_v):
+ return ('%s-%s' % (k_v[0], len(k_v[1])), set(k_v[1]))
+
+ result = (
+ p
+ | beam.Create([1, 1, 2, 3, 4, 5, 10, 11])
+ | beam.FlatMap(lambda t: [('A', t), ('B', t + 5)])
+ | beam.Map(construct_timestamped)
+ | beam.WindowInto(
+ FixedWindows(10),
+ trigger=Never(),
+ accumulation_mode=AccumulationMode.DISCARDING)
+ | beam.GroupByKey()
+ | beam.Map(format_result))
+ assert_that(result, equal_to([]))
Review comment:
Ack. (Interestingly, I originally was expecting it to fire at EoW.) The
ULR doesn't yet support gc timers. I can look into fixing this (or at least
making the "Never" trigger correct).
As an aside, the non-trivial triggering and windowing in PAssert makes it
less than ideal for validating new runners.
----------------------------------------------------------------
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]
