sum parts of numpy.array

Question

Let's say I have the following array:

a = np.array([[1,2,3,4,5,6], 
              [7,8,9,10,11,12],
              [3,5,6,7,8,9]])

I want to sum the first two values of the first row: 1+2 = 3, then next two values: 3+4 = 7, and then 5+6 = 11, and so on for every row. My desired output is this:

array([[ 3,  7, 11],
       [15, 19, 23],
       [ 8, 13, 17]])

I have the following solution:

def sum_chunks(x, chunk_size):
    rows, cols = x.shape
    x = x.reshape(x.size / chunk_size, chunk_size)
    return x.sum(axis=1).reshape(rows, cols/chunk_size)

But it feels unnecessarily complicated, is there a better way? Perhaps a built-in?

Answer 1

Just use slicing:

a[:,::2] + a[:,1::2]

This takes the array formed by every even-indexed column (::2), and adds it to the array formed by every odd-indexed column (1::2).