如何从我的类别类型列中删除 NaN 值？ 我用过.dropna 但它不起作用

Question

I want to remove the NaN values from Dep Freq & Sex Orien. I used.dropna and get this result

How do I fix this? Thank you.

Answer 1

According to the pandas.DataFrame.dropna documentation, dropna method is used to completely drop the columns or rows. Are you going to drop completely columns with existing Nan values? Or what do you mean by “I want to remove Nan values”?

Answer 2

Try this to drop the Nan value columns

df.dropna(axis=1) 

Answer 3

you can find more info on how to call dropna() method at pandas.DataFrame.dropna and below listing example of usage with row if does have missing and columns drop if any contains NaN so resulting is empty dataframe since all cols does have missing:

In [4]: import pandas as pd

In [5]: import numpy as np

In [6]: df = pd.DataFrame(np.random.randn(5, 3), index=['a', 'c', 'e', 'f', 'h'],columns=['one', 'two', 'three'])

In [7]: df
Out[7]: 
        one       two     three
a -1.040103  1.964200  1.519638
c -0.796710  1.654887 -0.614065
e  1.899870  0.810478  1.294714
f -0.913869  1.052014 -0.114583
h  0.186190 -0.156173 -2.323759

In [8]: df2 = df.reindex(['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h'])

In [9]: df2
Out[9]: 
        one       two     three
a -1.040103  1.964200  1.519638
b       NaN       NaN       NaN
c -0.796710  1.654887 -0.614065
d       NaN       NaN       NaN
e  1.899870  0.810478  1.294714
f -0.913869  1.052014 -0.114583
g       NaN       NaN       NaN
h  0.186190 -0.156173 -2.323759

In [10]: df2.dropna()
Out[10]: 
        one       two     three
a -1.040103  1.964200  1.519638
c -0.796710  1.654887 -0.614065
e  1.899870  0.810478  1.294714
f -0.913869  1.052014 -0.114583
h  0.186190 -0.156173 -2.323759

In [11]: df2.dropna(axis="columns")
Out[11]: 
Empty DataFrame
Columns: []
Index: [a, b, c, d, e, f, g, h]

In [12]: df2
Out[12]: 
        one       two     three
a -1.040103  1.964200  1.519638
b       NaN       NaN       NaN
c -0.796710  1.654887 -0.614065
d       NaN       NaN       NaN
e  1.899870  0.810478  1.294714
f -0.913869  1.052014 -0.114583
g       NaN       NaN       NaN
h  0.186190 -0.156173 -2.323759

In [13]: df2.dropna(inplace=True)

In [14]: df2
Out[14]: 
        one       two     three
a -1.040103  1.964200  1.519638
c -0.796710  1.654887 -0.614065
e  1.899870  0.810478  1.294714
f -0.913869  1.052014 -0.114583
h  0.186190 -0.156173 -2.323759

if you look to drop observations or columns in specific columns:

In [26]: df2[['one','three']].dropna(axis='columns')
Out[26]: 
Empty DataFrame
Columns: []
Index: [a, b, c, d, e, f, g, h]

In [27]: df2[['one','three']].dropna(axis=0)
Out[27]: 
        one     three
a -1.040103  1.519638
c -0.796710 -0.614065
e  1.899870  1.294714
f -0.913869 -0.114583
h  0.186190 -2.323759

Answer 4

You can use the replace( ) function

df.replace('NaN', 'Word XY')

Example

Table with NaN

import pandas as pd

df = pd.DataFrame([
['1', 'Fares', 32, True],
['2', 'Elena', 23, 'NaN'],
['NaN', 'Steven', 40, True],
['4', 'Max', 24, 'NaN'],
['5', 'Mike', 20, False],
['NaN', 'John', 40, True]])

df.columns = ['id', 'name', 'age', 'decision']

df

Output:

Now use the replace function( )

df.replace('NaN', ' ')

Desired Output:

So what this does is simply replace the string 'NaN' with an empty whitespace. You can add whatever word you want for the replacement.

df.replace('NaN', '######') would replace every NaN with number signs.

I hope I could help you a bit.