cannot replace [''] with method pad on a DataFrame

Question

running python2.7

Question1

I want to replace empty string '' with None in my dataframe "test":

    from numpy.random import randn
    test = pd.DataFrame(randn(3,2))
    test.iloc[0,0]=''
    test.replace('', None)

Give me error TypeError:

cannot replace [''] with method pad on a DataFrame.

What went wrong?

Question 2:

from numpy.random import randn

test = pd.DataFrame(randn(3,2))
# this works
test.iloc[0,1]= 'A'
test[1] = test[1].replace('A','b')
# this does not
test.iloc[0,0]=''
test.loc[0] = test[0].replace('', None)


test 
          0         1
0                   b
1  0.042052  -1.44156
2  0.462131 -0.303288

I am expecting

test 
          0         1
0    None           b
1  0.042052  -1.44156
2  0.462131 -0.303288

Answer 1

None is being interpreted as a lack of an argument. Try:

test.replace({'':None})

Question 2:

test.where(test != '', None)

Answer 2

Just use not-a-number. Pandas would convert your None into not-a-number anyways*. Second, do not forget to assign your result to a new variable or set the inplace paramater to True instead. test.replace('', np.NaN) does nothing on its own.

import pandas as pd
import numpy as np
from numpy.random import randn
test = pd.DataFrame(randn(3,2))
test.iloc[0,0]=''
test = test.replace('', np.NaN)

*Pandas works best and is most efficient when all the data has the same datatype. That's numpy.float64 in your case. np.NaN is also a numpy float. If you wanted to have None in your dataframe, everything would need to be stored as object datatype which is less efficient. np.NaN is probably what you need.