Python how to convert multiple columns into common OneHotEncoding style categorical columns without duplicating same column names

user_16903537 0
Viewed 30
cat_1 = ["A", "B", "C", "D"]
cat_2 = ["C", "A", "E", "A"]
cat_3 = ["NaN","F","NaN","NaN"]
  
# dictionary of lists 
dict = {'image_1': cat_1, 'image_2': cat_2, 'image_3': cat_3} 
    
df = pd.DataFrame(dict)
    
df

Above I have given the example data, when using "pd.get_dummies(df)" I get the same categorical data's name duplicated with various names. In other words when used pd.get_dummies(df) I get:

enter image description here

As seen value "A" is repeated for image_1 and image_2.

What I am aiming for commonize a single column for every unique value instead of extra columns for the same value

enter image description here

one-hot-encoding pandas python
2 Answers

You can stack first, then aggregate:

pd.get_dummies(df.stack()).groupby(level=0).sum()

NB. If you use real NaNs, they will be absent by default. Also, in case you have several values of a kind in a row and don't want to have numbers higher than 1, use max instead of sum as aggregation function.

Output:

   A  B  C  D  E  F
0  1  0  1  0  0  0
1  1  1  0  0  0  1
2  0  0  1  0  1  0
3  1  0  0  1  0  0
user_163434640

In your case do get_dummies

df = df.mask(df=='NaN')
out = pd.get_dummies(df,prefix_sep='', prefix='').groupby(level=0, axis=1).sum()
Out[25]: 
   A  B  C  D  E  F
0  1  0  1  0  0  0
1  1  1  0  0  0  1
2  0  0  1  0  1  0
3  1  0  0  1  0  0
user_79645270
Related