[
https://issues.apache.org/jira/browse/BEAM-14474?focusedWorklogId=771026&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-771026
]
ASF GitHub Bot logged work on BEAM-14474:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 16/May/22 20:31
Start Date: 16/May/22 20:31
Worklog Time Spent: 10m
Work Description: TheNeuralBit commented on code in PR #17682:
URL: https://github.com/apache/beam/pull/17682#discussion_r874129298
##########
sdks/python/apache_beam/dataframe/frames_test.py:
##########
@@ -1723,28 +1729,24 @@ def test_groupby_callable(self):
def test_groupby_apply(self):
df = GROUPBY_DF
-
- def median_sum_fn(x):
- return (x.foo + x.bar).median()
-
Review Comment:
nit: I would actually keep these where they are and just duplicate the
filterwarnings calls, so that the tests are more self-contained and easy to
inspect (but I'm also fine with the move if you prefer it that way)
##########
sdks/python/apache_beam/dataframe/frames_test.py:
##########
@@ -1601,6 +1602,11 @@ def test_pivot_no_index_provided_on_multiindex(self):
class GroupByTest(_AbstractFrameTest):
"""Tests for DataFrame/Series GroupBy operations."""
+ @staticmethod
+ def median_sum_fn(x):
+ warnings.filterwarnings("ignore", message="Mean of empty slice")
+ return (x.foo + x.bar).median()
Review Comment:
Could we use a
[context](https://docs.python.org/3/library/warnings.html#temporarily-suppressing-warnings)
here to make sure the filters get reset?
Issue Time Tracking
-------------------
Worklog Id: (was: 771026)
Time Spent: 1h (was: 50m)
> Clean up/Fix runtime warning messages in python unit test standard error
> ------------------------------------------------------------------------
>
> Key: BEAM-14474
> URL: https://issues.apache.org/jira/browse/BEAM-14474
> Project: Beam
> Issue Type: Test
> Components: sdk-py-core
> Reporter: Yi Hu
> Priority: P3
> Time Spent: 1h
> Remaining Estimate: 0h
>
> Currently the log of python unit test results contains thousands of lines of
> runtime warning / future warning / deprecation warnings, making trouble
> shooting during development inconvenient. Most of the warning messages are
> not necessary at all and could be suppressed or fixed if possible.
> For example, the most popped message is hundreds of lines of 'RuntimeWarning:
> Mean of empty slice
> return np.nanmean(a, axis, out=out, keepdims=keepdims)'
> raised in dataframe tests, which could be suppressed.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
