pandas.Series.fillna change type of the column

Question

After changing type of columns in pandas.DataFrame from int64 to object, fillna, applied to this column, returns columns with int64 type again. For example:

import pandas as pd

data = pd.DataFrame({"a" : [2, 3]})

# changing type to 'object'
data['a'] = data['a'].astype('object')
print("type after astype -", data['a'].dtype)

# applying fillna
data["a"] = data["a"].fillna("no data")
print("type after fillna -", data['a'].dtype)

Will return:

type after astype - object
type after fillna - int64

How fix it, without using astype again.

Answer 1

You can use downcast=False to prevent this:

data['a'].fillna('no data', downcast=False)