Sum elements in window by n step in Numpy

Question

Is a Numpy way to make a sum each three elements in the interval with a step? For example:

import numpy as np
mydata = np.array([4, 2, 3, 8, -6, 10])

I would like to get this result:

np.array([9, 12])

I suppose that np.convole can do this, according to Summing elements in a sliding window - NumPy, but can I change the step from n=1 to n=3, in this case?

Try something like `mydata.reshape(-1, 3).sum(axis=1)`. This assume `mydata.size` is divisible by 3. Otherwise you need to pad the array. — Jérôme Richard, Aug 11 '22 at 10:43

score 0 · Accepted Answer · answered Aug 11 '22 at 10:52

0

Using reshape (requires the array length to be a multiple of n):

n = 3

mydata.reshape(-1, n).sum(1)

If you don't have a multiple, you can trim:

n = 3

mydata[:len(mydata)//n*n].reshape(-1, n).sum(1)

Using convolve, which should be much less efficient for large n as many values (n-1 out of n) are computed for nothing:

np.convolve(np.ones(n), mydata)[n-1::n]

Output: array([ 9, 12])

answered Aug 11 '22 at 10:52

mozway

All solutions are great! I like the third one, but trimming should not be done. It should be pad with 0 and sum the rest. – Krzyztopher Aug 11 '22 at 10:57
In this case you can use `numpy.pad` – mozway Aug 11 '22 at 11:05

1 Answers1