rohdesamuel commented on a change in pull request #14327:
URL: https://github.com/apache/beam/pull/14327#discussion_r610024537
##########
File path: sdks/python/apache_beam/dataframe/frames_test.py
##########
@@ -526,13 +526,12 @@ def test_dataframe_cov_corr(self):
df = pd.DataFrame(np.random.randn(20, 3), columns=['a', 'b', 'c'])
df.loc[df.index[:5], 'a'] = np.nan
df.loc[df.index[5:10], 'b'] = np.nan
- self._run_test(lambda df: df.corr().round(8), df)
- self._run_test(lambda df: df.cov().round(8), df)
- self._run_test(lambda df: df.corr(min_periods=12).round(8), df)
- self._run_test(lambda df: df.cov(min_periods=12).round(8), df)
- self._run_test(lambda df: df.corrwith(df.a).round(8), df)
- self._run_test(
- lambda df: df[['a', 'b']].corrwith(df[['b', 'c']]).round(8), df)
+ self._run_test(lambda df: df.corr(), df)
+ self._run_test(lambda df: df.cov(), df)
+ self._run_test(lambda df: df.corr(min_periods=12), df)
+ self._run_test(lambda df: df.cov(min_periods=12), df)
+ self._run_test(lambda df: df.corrwith(df.a), df)
+ self._run_test(lambda df: df[['a', 'b']].corrwith(df[['b', 'c']]), df)
Review comment:
That would be true if there was only logic that wrapped the
DataFrame.corr method, but there is manual logic here that unwraps the columns
into every 2-combination and correlates those. It's a non-blocking issue so up
to you.
--
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]
