Combining two numpy arrays to form an array with the largest value from each array

Question

I want to combine two numpy arrays to produce an array with the largest values from each array.

import numpy as np

a  = np.array([[ 0.,  0.,  0.5],
               [ 0.1,  0.5,  0.5],
               [ 0.1,  0.,  0.]])

b  = np.array([[ 0.,  0.,  0.0],
               [ 0.5,  0.1,  0.5],
               [ 0.5,  0.1,  0.]])

I would like to produce

array([[ 0.,  0.,  0.5],
       [ 0.5,  0.5,  0.5],
       [ 0.5,  0.1,  0.]])

I know you can do

a += b

which results in

array([[ 0. ,  0. ,  0.5],
       [ 0.6,  0.6,  1. ],
       [ 0.6,  0.1,  0. ]])

This is clearly not what I'm after. It seems like such an easy problem and I assume it most probably is.

Answer 1

You can use np.maximum to compute the element-wise maximum of the two arrays:

>>> np.maximum(a, b)
array([[ 0. ,  0. ,  0.5],
       [ 0.5,  0.5,  0.5],
       [ 0.5,  0.1,  0. ]])

This works with any two arrays, as long as they're the same shape or one can be broadcast to the shape of the other.

To modify the array a in-place, you can redirect the output of np.maximum back to a:

np.maximum(a, b, out=a)

There is also np.minimum for calculating the element-wise minimum of two arrays.