Replace all but the first 1 in an array with 0

February 6, 2023

I am trying to find a way to replace all of the duplicate 1 with 0. As an example:

[[0,1,0,1,0],
[1,0,0,1,0],
[1,1,1,0,1]]

Should become:

[[0,1,0,0,0],
[1,0,0,0,0],
[1,0,0,0,0]]

I found a similar problem, however the solution does not seem to work numpy: setting duplicate values in a row to 0

>Solution :

Assume array contains only zeros and ones, you can find the max value per row using numpy.argmax and then use advanced indexing to reassign the values on the index to a zeros array.

arr = np.array([[0,1,0,1,0],
[1,0,0,1,0],
[1,1,1,0,1]])

res = np.zeros_like(arr)
idx = (np.arange(len(res)), np.argmax(arr, axis=1))
res[idx] = arr[idx]

res
array([[0, 1, 0, 0, 0],
       [1, 0, 0, 0, 0],
       [1, 0, 0, 0, 0]])