Fill in missing values in Pandas Dataframe wrong

Question

Suppose 'df' is dataframe object, 'ca' is one of the variables.

>>> df.ca.value_counts()
0.0    176
1.0     65
2.0     38
3.0     20
?        4
Name: ca, dtype: int64

As you can see, I have four missing values. I want to fill in them. Using below code:

>>> df.loc[df.ca == '?', 'ca'] = 0.0
0.0    176
1.0     65
2.0     38
3.0     20
0.0      4
Name: ca, dtype: int64

Why I got 5 unique values? I want to merge fifth row into first row, i.e.

0.0   176 + 4 = 180
1.0     65
2.0     38
3.0     20

How can I fix it?

Answer 1

Because '?' was one of your values, I know that df.ca is either dtype object or string. When you replace('?', 0.) you now have both string '0.0' and float 0.0. After you convert all to float, you shouldn't have a problem.

df.ca.replace('?', 0.).astype(float).value_counts()

0.0    180
1.0     65
2.0     38
3.0     20
dtype: int64