[Python] Fastest way of creating a pair-wise matching matrix from a single numpy array

I want to do quick pair-wise matchings of all elements in a numpy array, ignoring self-matching.

For example given a numpy array like

a = np.array(range(4))


I want to get a matrix like

array([[0, 1],
[0, 2],
[0, 3],
[1, 0],
[1, 2],
[1, 3],
[2, 0],
[2, 1],
[2, 3],
[3, 0],
[3, 1],
[3, 2]])


In my real case, I have different sizes of arrays varying from 1 to 1000. I need to constantly call a function that does this pair-wise matching. I wrote the code below

import numpy as np
from itertools import permutations

a = np.array(range(1000))
xs, ys = zip(*[(x, y) for x, y in permutations(a, 2)])
res = np.vstack([np.array(xs), np.array(ys)]).T
print(res)


I want to know if there are faster ways (maybe more numpyic ways) to do this?

I would like to see some benchmarkings of runtime vs. array size.

Continue reading...