pandas2.0后mean()更新

df = pd.DataFrame(
    {'key1': ['A', 'A', 'B', 'B', 'A'], 'key2': ['one', 'two', 'one', 'two', 'one'], 'data1': [2, 3, 4, 6, 8],
     'data2': [3, 5, np.nan, 3, 7]})

print(df.groupby(by='key1').mean())

报错TypeError

在demo中mean()会将key2也传入方法内执行

解决方式：

This behavior was changed in pandas 2.0: https://pandas.pydata.org/docs/whatsnew/v1.5.0.html#numeric-only-default-value

Specify .mean(numeric_only=True) if you'd like to keep the old behavior.