[ https://issues.apache.org/jira/browse/BEAM-3377?focusedWorklogId=100837&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-100837 ]
ASF GitHub Bot logged work on BEAM-3377: ---------------------------------------- Author: ASF GitHub Bot Created on: 10/May/18 21:07 Start Date: 10/May/18 21:07 Worklog Time Spent: 10m Work Description: mariapython commented on a change in pull request #5279: [BEAM-3377] Call assert_that per window via windowed (flag) URL: https://github.com/apache/beam/pull/5279#discussion_r187460644 ########## File path: sdks/python/apache_beam/testing/util.py ########## @@ -75,6 +75,25 @@ def __repr__(self): return InAnyOrder(iterable) +def equal_to_per_window(expected_dict_window_to_its_elements): + def matcher(elements): + actual_elements_in_window, window = elements + if window in expected_dict_window_to_its_elements: + expected_elements_in_window = list( + expected_dict_window_to_its_elements[window]) + sorted_expected = sorted(expected_elements_in_window) + sorted_actual = sorted(actual_elements_in_window) + if sorted_expected != sorted_actual: + # Results for the same window don't necessarily come all + # at once. Hence the same actual window may contain only + # subsets of the expected elements for the window. + # For example, in the presence of early triggers. + if all(elem in sorted_expected for elem in sorted_actual) is False: Review comment: Done. ---------------------------------------------------------------- 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: us...@infra.apache.org Issue Time Tracking ------------------- Worklog Id: (was: 100837) Time Spent: 3h 10m (was: 3h) > assert_that not working for streaming > ------------------------------------- > > Key: BEAM-3377 > URL: https://issues.apache.org/jira/browse/BEAM-3377 > Project: Beam > Issue Type: Bug > Components: sdk-py-core > Affects Versions: 2.2.0 > Reporter: MarĂa GH > Priority: Major > Labels: starter > Time Spent: 3h 10m > Remaining Estimate: 0h > > assert_that does not work for AfterWatermark timers. > Easy way to reproduce: modify test_gbk_execution [1] in this form: > > {code:java} > def test_this(self): > test_stream = (TestStream() > .add_elements(['a', 'b', 'c']) > .advance_watermark_to(20)) > def fnc(x): > print 'fired_elem:', x > return x > options = PipelineOptions() > options.view_as(StandardOptions).streaming = True > p = TestPipeline(options=options) > records = (p > | test_stream > | beam.WindowInto( > FixedWindows(15), > > trigger=trigger.AfterWatermark(early=trigger.AfterCount(2)), > accumulation_mode=trigger.AccumulationMode.ACCUMULATING) > | beam.Map(lambda x: ('k', x)) > | beam.GroupByKey()) > assert_that(records, equal_to([ > ('k', ['a', 'b', 'c'])])) > p.run() > {code} > This test will pass, but if the .advance_watermark_to(20) is removed, the > test will fail. However, both cases fire the same elements: > fired_elem: ('k', ['a', 'b', 'c']) > fired_elem: ('k', ['a', 'b', 'c']) > In the passing case, they correspond to the sorted_actual inside the > assert_that. In the failing case: > sorted_actual: [('k', ['a', 'b', 'c']), ('k', ['a', 'b', 'c'])] > sorted_actual: [] > [1] > https://github.com/mariapython/incubator-beam/blob/direct-timers-show/sdks/python/apache_beam/testing/test_stream_test.py#L120 -- This message was sent by Atlassian JIRA (v7.6.3#76005)