Get An Array Back From An Itertools.chain Object

Suppose I have list_of_numbers = [[1, 2], [3], []] and I want the much simpler object list object x = [1, 2, 3]. Following the logic of this related solution, I do list_of_numbers

Solution 1:

list. If you do list(chain) it should work. But use this just for debugging purposes, it could be inefficient in general.

Solution 2:

If your ultimate goal is to get a Numpy array then you should use numpy.fromiter here:

>>> import numpy as np
>>> from itertools import chain
>>> list_of_numbers = [[1, 2], [3], []]
>>> np.fromiter(chain(*list_of_numbers), dtype=int)
array([1, 2, 3])
>>> list_of_numbers = [[1, 2]*1000, [3]*1000, []]*1000
>>> %timeit np.fromiter(chain(*list_of_numbers), dtype=int)
10 loops, best of 3: 103 ms per loop
>>> %timeit np.array(list(chain(*list_of_numbers)))
1 loops, best of 3: 199 ms per loop

Solution 3:

You can do it with list(chain).