EA: fillna should accept same type

[jbrockmendel]

cat = pd.Categorical(["A", "B", None, "A"])
ser = pd.Series(cat).fillna("B")

>>> filled = cat.fillna(ser)  # <-- works

>>> cat.fillna(filled)
TypeError: 'value' parameter must be a scalar, dict or Series, but you passed a Categorical

@jbrockmendel Is this issue appropriate for first-time contributors?

[jbrockmendel]

Is this issue appropriate for first-time contributors?

Probably not; it gets pretty deep into the weeds. #32402 might be a good fit if you're comfortable with cython, otherwise i suggest searching for the Good First Issue label

[mroeschke]

I think this work on master. Could use a test

In [22]: cat = pd.Categorical(["A", "B", None, "A"])
    ...: ser = pd.Series(cat).fillna("B")

In [23]: >>> filled = cat.fillna(ser)

In [24]: >>> cat.fillna(filled)
Out[24]:
['A', 'B', 'B', 'A']
Categories (2, object): ['A', 'B']

[Bhavay-2001]

hey @mroeschke , Could you please tell on which cases should we really test it on?

[mroeschke]

@Bhavay192 This entire code snippet should be tested

[Bhavay-2001]

Hey @mroeschke , thanks for replying. Soo according to you I should take different values of numpy arrays and test it on that??

[mroeschke]

The exact code snippet above should be written as a unit test and added to the testing suite: https://pandas.pydata.org/pandas-docs/stable/development/contributing_codebase.html#writing-tests

[Bhavay-2001]

Hey @mroeschke, Since I'm quite new to open source and unit testing. Could you please help with in finding out what all test should I Perform on it??
I can take test of input values and what should be the expected output. Any other test that I can perform??