Slicing numpy array taking every nth element

Question

I have an numpy array of shape 24576x25 and i want to extract 3 array out of it. Where the first array contains the every 1st,4th,7th,10th,... element while second array contains 2nd,5,8,11th,... element and third array with 3rd,6,9,12th,... The output array sizes would be 8192x25. I was doing the following in MATLAB

c = reshape(a,1,[]);
x = c(:,1:3:end);
y = c(:,2:3:end);
z = c(:,3:3:end);

I have tried a[:,0::3] in python but this works only if i have array of shape divisible by 3. What can i do?

X,Y = np.mgrid[0:24576:1, 0:25:1]
a = X[:,::,3]
b = X[:,1::3]
c = X[:,2::3]

does not work either. I need a,b,c.shape = 8192x25

If you need 8192x25, then it is a[0::3]. @bigbounty answer definitely works. — paime
– paime, Commented Jul 7, 2020 at 11:38
it gives me the same size but the elements picking is wrong.. as it should return 0,3,6,9,,... but its giving me 0 1 2 3,... — A R.
– A R., Commented Jul 7, 2020 at 12:35

JaHoff · Accepted Answer · 2020-07-07 13:11:01Z

2

A simple tweak to your original attempt should yield the results you want:

X,Y = np.mgrid[0:24576:1, 0:25:1]
a = X[0::3,:]
b = X[1::3,:]
c = X[2::3,:]

answered Jul 7, 2020 at 13:11

JaHoff

2061 silver badge5 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

bigbounty · Accepted Answer · 2020-07-07 12:37:07Z

0

import numpy as np
a = np.arange(24576*25).reshape((24576,25))
a[::3]

a[::3].shape gives you (8192, 25)

edited Jul 7, 2020 at 12:37

answered Jul 7, 2020 at 11:13

bigbounty

17.5k7 gold badges45 silver badges76 bronze badges

3 Comments

A R. Over a year ago

i did the same, doesn't return 8192x25 x3

A R. Over a year ago

this gives me (8192,25) but not the correct elements as it should give 0,3,6,.. but I am not getting it

bigbounty Over a year ago

@AR. When you go through the elements of a[::3] - it will give you 0,3,6

Collectives™ on Stack Overflow

Slicing numpy array taking every nth element

2 Answers 2

Comments

3 Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

3 Comments

Related