Is there a better way to use converting the float values in fillna() into an object dtype?

Question

Is there a better way to do this conversion of the fillna values with object datatype :

df =df.fillna({"first_col":0.0, "second_col":0.0})

Answer 1

I think you need check downcast parameter, Dataframe.fillna

Answer 2

I have first converted it to the int type and then the object type.

# float (-> fillna) -> int -> object

# converting to int type
df["first_col"] = df["first_col"].astype("int")
df["second_col"] = df["second_col"].astype("int")

# converting to obj. type
df["first_col"] = df["first_col"].astype("object")
df["second_col"] = df["second_col"].astype("object")