potiuk commented on a change in pull request #11004:
URL: https://github.com/apache/airflow/pull/11004#discussion_r492175173
##########
File path: setup.py
##########
@@ -720,7 +720,7 @@ def is_package_excluded(package: str, exclusion_list:
List[str]):
'markdown>=2.5.2, <3.0',
'markupsafe>=1.1.1, <2.0',
'marshmallow-oneofschema>=2.0.1',
- 'pandas>=0.17.1, <2.0',
+ 'pandas>=1.1.2, <2.0',
Review comment:
This is a very good question whether it is experimental or not :). It
actually changed in 1.1.1 . (1.1.0 worked fine, 1.1.1 and 1.1.2 failed). This
is likely not a bug in Panda's but in the implementation of
"assert_frame_equal()".
https://pandas.pydata.org/pandas-docs/stable/whatsnew/v1.1.1.html
> Bug in pandas.testing.assert_series_equal() and
pandas.testing.assert_frame_equal() where extension dtypes were not ignored
when check_dtypes was set to False (GH35715)
I figured that the easiest way to fix it will be to change the test to pas
nan and expect nan rather than None. At the end this is entirely Panda's
problem not Airflow's what is the behaviour when you pass None to DataFrame and
want to export it to CSV.
I tested that nan <> 'nan' works fine for 1.1.0, 1.1.1 and 1.1.2 :). So all
shoudl be good.
----------------------------------------------------------------
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:
us...@infra.apache.org
