Closed Jayson729 closed 1 week ago
Apparently this fails this test https://github.com/modin-project/modin/blob/main/modin/tests/pandas/test_series.py#L3430-L3439 because there is no longer a UserWarning being given in this case. I think changing this test to the following would solve this and still test the same thing, although we will have to import warnings
since that is not included in this file. Does this seem right?
def test_6782():
datetime_scalar = datetime.datetime(1970, 1, 1, 0, 0)
match = "Adding/subtracting object-dtype array to DatetimeArray not vectorized"
with warnings.catch_warnings():
warnings.filterwarnings("error", match, UserWarning)
pd.Series([datetime.datetime(2000, 1, 1)]) - datetime_scalar
Does this seem right?
@Jayson729 yes, it looks great! Only the category should be different (PerformanceWarning
):
https://github.com/pandas-dev/pandas/blob/c46fb76afaf98153b9eef97fc9bbe9077229e7cd/pandas/core/arrays/datetimelike.py#L1316
Note: It looks like 6782
test no longer worked as originally planned since pandas changed the warning category.
What do these changes do?
When creating a DataFrame or Series, no longer display
UserWarning: Distributing <class 'NoneType'> object. This may take some time.
when the size of the DataFrame or Series is small.flake8 modin/ asv_bench/benchmarks scripts/doc_checker.py
black --check modin/ asv_bench/benchmarks scripts/doc_checker.py
git commit -s
docs/development/architecture.rst
is up-to-date