How to randomly select elements from subset of dataframe?

Question

I have dataframe in the following form:

W1 W2 W3 W4 0 1 1 0 1 1 1 1 1 0 0 0 0 1 0 1

For every row, I want to randomly select single element that is 1 and make other ones zero. Initial zeros stay zeros Eg

W1 W2 W3 W4 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1

I have very convoluted solution that uses iterrows() , but I'm looking for a pandastic one.

Answer 1

Idea is extract positions, shuffling and then remove duplicates by first column 0 - by rows:

#get positions of 1
a = np.where(df == 1)

#create nd array
X = np.hstack((a[0][:, None], a[1][:, None]))
#shuffling
np.random.shuffle(X)

#remove duplicates
vals = pd.DataFrame(X).drop_duplicates(0).values

#set 1
arr = np.zeros(df.shape)
arr[vals[:,0],vals[:,1]] = 1

df = pd.DataFrame(arr.astype(int), columns=df.columns, index=df.index)
print (df)
   W1  W2  W3  W4
0   0   0   1   0
1   0   0   0   1
2   1   0   0   0
3   0   1   0   0

Answer 2

IIUC, you want to randomly select 1 from every row and make the rest 0. Here's one approach. Sample the indices and based on indices assign 1. ie

idx = pd.DataFrame(np.stack(np.where(df==1))).T.groupby(0).apply(lambda x: x.sample(1)).values
# array([[0, 2],
#        [1, 1],
#        [2, 0],
#        [3, 3]])

ndf = pd.DataFrame(np.zeros(df.shape),columns=df.columns)

ndf.values[idx[:,0],idx[:,1]] = 1

   W1  W2  W3  W4
0   0   0   1   0
1   1   0   0   0
2   1   0   0   0
3   0   1   0   0

Answer 3

Here is mixture of functional and pandastic approach:

df = pd.DataFrame({'w1': [0, 1,1,0],
                   'w2': [1, 1,0,1],
                   'w3': [1, 1,0,0],
                   'w4': [0, 1,0,1]})
df
   w1  w2  w3  w4
0   0   1   1   0
1   1   1   1   1
2   1   0   0   0
3   0   1   0   1


def choose_one(row):
    """
    returns array with randomly chosen positive value and 0 otherwise
    """
    one = np.random.choice([i for i, v in enumerate(row) if v])
    return [0 if i != one else 1 for i in range(len(row))]

apply for each row

df.apply(choose_one, 1)

   w1  w2  w3  w4
0   0   1   0   0
1   0   1   0   0
2   1   0   0   0
3   0   0   0   1