Python: convert 1D list to 3D numpy array

Question

I have a list a and need to convert it to a numpy array b with shape (2, 3, 4) and elements in the following order.

a = [0, 12, 1, 13, 2, 14, 3, 15, 4, 16, 5, 17, 6, 18, 7, 19, 8, 20, 9, 21, 10, 22, 11, 23]

b = array([[[ 0,  1,  2,  3],
    [ 4,  5,  6,  7],
    [ 8,  9, 10, 11]],

   [[12, 13, 14, 15],
    [16, 17, 18, 19],
    [20, 21, 22, 23]]])

I tried a bit and got these two ways:

b = np.rollaxis(np.asarray(a).reshape(3, 4, 2), 2)
b = np.asarray(a).reshape(2,4,3, order="F").swapaxes(1, 2)

Is there any shorter way to do it?

You could also stack up a[::2] and a[1::2] and then reshape that. I don’t know if it’ll be much shorter. Ultimately, you’re doing two things: un-interleaving and array, and reshaping an array—and any readable solution is going to involve two steps like your existing ones. — abarnert
– abarnert, Commented Sep 7, 2018 at 20:00

user3483203 · Accepted Answer · 2018-09-07 20:45:02Z

Using reshape and transpose:

a.reshape(-1, 2).T.reshape(-1, 3, 4)

array([[[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],

       [[12, 13, 14, 15],
        [16, 17, 18, 19],
        [20, 21, 22, 23]]])

Timings on your sample array:

%timeit np.rollaxis(a.reshape(3, 4, 2), 2)
2.92 µs ± 10.9 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

%timeit a.reshape(2,4,3, order="F").swapaxes(1, 2)
1.1 µs ± 11.9 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)

%timeit a.reshape(-1, 2).T.reshape(-1, 3, 4)
1.08 µs ± 7.36 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)

I haven't timed this answer on massive arrays because I haven't figured out a way to generalize either of your solutions. One benefit of my solution is that it scales without any altering of the code:

a = np.zeros(48)
a[::2] = np.arange(24)
a[1::2] = np.arange(24, 48) 
a.reshape(-1, 2).T.reshape(-1, 3, 4)

array([[[ 0.,  1.,  2.,  3.],
        [ 4.,  5.,  6.,  7.],
        [ 8.,  9., 10., 11.]],

       [[12., 13., 14., 15.],
        [16., 17., 18., 19.],
        [20., 21., 22., 23.]],

       [[24., 25., 26., 27.],
        [28., 29., 30., 31.],
        [32., 33., 34., 35.]],

       [[36., 37., 38., 39.],
        [40., 41., 42., 43.],
        [44., 45., 46., 47.]]])

apitsch · Accepted Answer · 2018-09-07 20:24:32Z

0

Another way would be:

import numpy as np
np.reshape(sorted(a), (2, 3, 4))

In case you have converted a to an array already, instead do:

np.reshape(np.sort(a), (2, 3, 4))

edited Sep 7, 2018 at 20:24

answered Sep 7, 2018 at 20:01

apitsch

1,73218 silver badges32 bronze badges

1 Comment

Kasravnd Over a year ago

Inefficient for long arrays.

Collectives™ on Stack Overflow

Python: convert 1D list to 3D numpy array

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related