[FIXED] Invert binary NumPy array using a mask

Issue

I have a binary N-by-N (N~=200) NumPy array populated with zeroes and ones. I would like to apply a Boolean mask and 'swap' the values that correspond to True in the mask, so for example if I had:

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

I would like the resulting array to be:

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

I started this by initially creating a function swap_cell that swaps between the values, and then followed the approaches in this answer.

def swap_cell(x):
    if x == 1.0:
        return 0.0
    elif x == 0.0:
        return 1.0

arr_new = np.where(mask,swap_cell(arr),arr)

This code returns ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all(), which I understand is because of the 'if' statement in my swap_cell() function. I know there must be a much more pythonic way to accomplish this, but I thought this might work.

Solution

If you interpret the first (0,1) array as a boolean array, then you are doing the exclusive or operation of each entry in the first array with each entry of the second array.

Therefore, you can do this with the numpy function np.logical_xor() as follows:

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

np.array(np.logical_xor(arr, mask), dtype=int)

which returns

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

Answered By - Nick Alger