ENH: assert_frame_equal to show column name for columns of dtype datetime64[ns] #43457
Description
Is your feature request related to a problem?
I wish I could use pandas.testing.assert_frame_equal to determine which of my datetime64[ns] columns are not equal when comparing two dataframes. Currently the output is AssertionError: numpy array are different without indicating which columns are different.
Describe the solution you'd like
When two dataframes with datetime64[ns] columns are compared using pandas.testing.assert_frame_equal and there are different values, the output should also have the column name that has this difference, e.g. AssertionError: DataFrame.iloc[:, 1] (column name="b") are different.
Additional context
Minimal reproducible code below. If you change one of the datetime values to an integer, the dtype of the column will be object, in which case, the desired behaviour is shown.
import pandas as pd
from pandas.testing import assert_frame_equal
df_left = pd.DataFrame({'a': [1, 2], 'b': [
datetime(2021, 1, 1),
# 1,
datetime(2021, 1, 2)
]})
df_right = pd.DataFrame({'a': [1, 2], 'b': [
datetime(2021, 1, 1),
# 1,
datetime(2021, 1, 3)
]})
assert_frame_equal(left=df_left, right=df_right)