I have a long list of (i, j, value) pairs, with multiple values for each position, the number of values varies for each position.

At the moment I am initializing an empty “3d” list with

sorted = numpy.empty((m, n, 0)).tolist()

then I loop over each entry in the array and append the value to the list at i, j.

for pixel in value_list:
    sorted[pixel[0]][pixel[1]].append(pixel[2])

This is slow and hard to parallelize since I have to write to the same lists and don´t know how long they are going to be at the end.

Is there a faster solution to this? I am sure this is a common problem but maybe I just don´t now what I have to look for.

This is slow and hard to parallelize since I have to write to the same lists and don´t know how long they are going to be at the end.

Is there a faster solution to this? I am sure this is a common problem but maybe I just don´t now what I have to look for.