多列上的 Pandas get_dummies

Question

I have a dataset with multiple columns that I wish to one hot encode. However, I don't want to have the encoding for each one of them since said columns are related to the said items. What I want is one "set" of dummies variables that uses all the columns. See my code for a better explanation.

Suppose my dataframe looks like this:

In [103]: dum = pd.DataFrame({'ch1': ['A', 'C', 'A'], 'ch2': ['B', 'G', 'F'], 'ch3': ['C', 'D', 'E']})

In [104]: dum
Out[104]:
 ch1 ch2 ch3
0   A   B   C
1   C   G   D
2   A   F   E

If I execute

pd.get_dummies(dum)

The output will be

   ch1_A  ch1_C  ch2_B  ch2_F  ch2_G  ch3_C  ch3_D  ch3_E
 0      1      0      1      0      0      1      0      0
 1      0      1      0      0      1      0      1      0
 2      1      0      0      1      0      0      0      1

However, what I would like to obtain is something like this:

 A B C D E F G
 1 1 1 0 0 0 0
 0 0 1 1 0 0 1
 1 0 0 0 1 1 0

Instead of having multiple columns representing the encoding, eg ch1_A and ch1_C , I only wish to have one group ( A , B , and so on) with value 1 when any of the values in the columns ch1 , ch2 , ch3 show up.

To clarify, in my original dataset, a single row won't contain the same value (A,B,C...) more than once; it will just appear on one of the columns.

Answer 1

Using stack and str.get_dummies

dum.stack().str.get_dummies().sum(level=0)
Out[938]: 
   A  B  C  D  E  F  G
0  1  1  1  0  0  0  0
1  0  0  1  1  0  0  1
2  1  0  0  0  1  1  0

Answer 2

You could use pd.crosstab to create a frequency table:

import pandas as pd

dum = pd.DataFrame({'ch1': ['A', 'C', 'A'], 'ch2': ['B', 'G', 'F'], 'ch3': ['C', 'D', 'E']})

stacked = dum.stack()
index = stacked.index.get_level_values(0)
result = pd.crosstab(index=index, columns=stacked)
result.index.name = None
result.columns.name = None

print(result)

yields

   A  B  C  D  E  F  G
0  1  1  1  0  0  0  0
1  0  0  1  1  0  0  1
2  1  0  0  0  1  1  0

Answer 3

Call it this way

x = pd.get_dummies(dum, prefix="", prefix_sep="")

And then print using

print(x.to_string(index=False))

Answer 4

You can create dummies for separate columns and concat the results:

temp = pd.concat([pd.get_dummies(dum[col]) for col in dum], axis=1)

    A   C   B   F   G   C   D   E
0   1   0   1   0   0   1   0   0
1   0   1   0   0   1   0   1   0
2   1   0   0   1   0   0   0   1

temp.groupby(level=0, axis=1).sum()

    A   B   C   D   E   F   G
0   1   1   1   0   0   0   0
1   0   0   1   1   0   0   1
2   1   0   0   0   1   1   0