fillna mit einem namedtuple oder anderen Klasse pandas

Question

Gibt es eine Möglichkeit, in Python mit einem namedtuple füllen?

Ich erhalte diese TypeError:

from collections import namedtuple 
import pandas as pd 
import numpy as np 

df = pd.DataFrame([0, 0, 0, 0, np.nan, 0, 0, 0]) 

nametup = namedtuple('mynp', ['arg1', 'arg2']) 
q = nametup(None, None) 
df.fillna(q) 

Traceback (most recent call last): 
    File "C:\Anaconda2\lib\site-packages\IPython\core\interactiveshell.py", line 2881, in run_code 
    exec(code_obj, self.user_global_ns, self.user_ns) 
    File "<ipython-input-25-363ec560dd77>", line 9, in <module> 
    df.fillna(q) 
    File "C:\Anaconda2\lib\site-packages\pandas\core\frame.py", line 2762, in fillna 
    downcast=downcast, **kwargs) 
    File "C:\Anaconda2\lib\site-packages\pandas\core\generic.py", line 3101, in fillna 
    'you passed a "{0}"'.format(type(value).__name__)) 
TypeError: "value" parameter must be a scalar or dict, but you passed a "mynp" 

auch schon versucht, diese:

df.replace(np.nan, q) 
Traceback (most recent call last): 
    File "C:\Anaconda2\lib\site-packages\IPython\core\interactiveshell.py", line 2881, in run_code 
    exec(code_obj, self.user_global_ns, self.user_ns) 
    File "<ipython-input-31-6f8a86f11bbb>", line 1, in <module> 
    df.replace(np.nan, q) 
    File "C:\Anaconda2\lib\site-packages\pandas\core\generic.py", line 3440, in replace 
    raise TypeError(msg) # pragma: no cover 
TypeError: Invalid "to_replace" type: 'float' 

jede Abhilfe? Vielen Dank!

Answer 1

Es ist nicht einfach, aber braucht Series durch Objekte zu erstellen und dann NaN ersetzen:

nametup = namedtuple('mynp', ['arg1', 'arg2']) 
q = nametup(None, None) 

s = pd.Series([q]*len(df.index)) 
print (s) 
0 (None, None) 
1 (None, None) 
2 (None, None) 
3 (None, None) 
4 (None, None) 
5 (None, None) 
6 (None, None) 
7 (None, None) 
dtype: object 

Lösung mit mask:

df[0] = df[0].mask(df[0].isnull(), s) 
print (df) 
       0 
0    0 
1    0 
2    0 
3    0 
4 (None, None) 
5    0 
6    0 
7    0 

Eine andere Lösung mit combine_first oder fillna von Seriess:

df[0] = df[0].combine_first(s) 
#similar solution 
#df[0] = df[0].fillna(s) 
print (df) 
       0 
0    0 
1    0 
2    0 
3    0 
4 (None, None) 
5    0 
6    0 
7    0