Expand 2-D Numpy integer Array as binary but in parallel

Question

I want to convert arrays of integers to 0 or 1s, padding with 0s if the other array possesses the larger value.

Examples:

ex1 = np.array([[0],[3]])
=> array([[0,0,0],[1,1,1]])


ex2 = np.array([[2,1],[0,0]])
=> array([[1,1,1],[0,0,0]])


ex3 = np.array([ [2,1,2],[3,1,1] ])
=> array([[1,1,0,1,1,1]
          [1,1,1,1,1,0]])

How shall I achieve this? Can it also expand the N-dimension array?

Answer 1

Came up with this approach:

def expand_multi_bin(a):
    # Create result array
    n = np.max(a, axis=0).sum()
    d = a.shape[0]
    newa = np.zeros(d*n).reshape(d,n)

    row=0
    for x in np.nditer(a, flags=['external_loop'], order='F'):
        # Iterate each column
        for idx,c in enumerate(np.nditer(x)):
            # Store it to the result array
            newa[idx,row:row+c] = np.ones(c)
        row += np.max(x) 

    return newa

Though, given the multiple loops, highly skeptical that this is the best approach.