Converting binary into categorical-CodePudding

I have a dataframe in the form of

     black  orange   yellow    green
  1    0      1         0        1
  2    0      0         0        1
  3    1      0         0        0

I would like to create another column that would tell which colours are present, so the final output SHOULD be:

     black  orange   yellow    green  colours
  1    0      1         0        1    orange and green
  2    0      0         0        1       green 
  3    1      0         0        0       black

Thanks in advance. P.S. Forgot to add there are a lot more columns than this - plenty of more colours.

CodePudding user response：

Use DataFrame.dot with the columns names and then remove last 5 values:

df['colours'] = df.dot(df.columns   ' and ').str[:-5]
print (df)
   black  orange  yellow  green           colours
1      0       1       0      1  orange and green
2      0       0       0      1             green
3      1       0       0      0             black

CodePudding user response：

Or use apply:

>>> df['colours'] = df.apply(lambda x: ' and '.join(x[x == 1].index), axis=1)
>>> df
   black  orange  yellow  green           colours
1      0       1       0      1  orange and green
2      0       0       0      1             green
3      1       0       0      0             black
>>>