What is the fastest way to map group names of numpy array to indices?

Question

I\'m working with 3D pointcloud of Lidar. The points are given by numpy array that looks like this:

points = np.array([[61651921, 416326074, 39805], [6160525         


        

Answer 1

You could use Cython:

%%cython -c-O3 -c-march=native -a
#cython: language_level=3, boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True, infer_types=True

import math
import cython as cy

cimport numpy as cnp


cpdef groupby_index_dict_cy(cnp.int32_t[:, :] arr):
    cdef cy.size_t size = len(arr)
    result = {}
    for i in range(size):
        key = arr[i, 0], arr[i, 1], arr[i, 2]
        if key in result:
            result[key].append(i)
        else:
            result[key] = [i]
    return result

but it will not make you faster than what Pandas does, although it is the fastest after that (and perhaps the numpy_index based solution), and does not come with the memory penalty of it. A collection of what has been proposed so far is here.

In OP's machine that should get close to ~12 sec execution time.