Summing parts of an array in python(A rolling window to sum spicific parts of an array)

Question

Deal all,

I have data in the form of a numpy array and want to perform summation on parts of the array:

Example:

X=[[1, 2, 3,4],
  [5, 6, 7,8]]

I want the output to be the summation of elements that within 2x2 window.

Output Needed:

[[14],[22]]

This is similar to cross-correlation with ones(2,2) but i want the shift to be more than ones element.

This is just an example and the actual data is very large so i cannot manually calculate them. I need a for loop to do that but struggling to do it.

Your Help is appreciated.

Answer 1

here is one way:

window_size = 2 # 2x2
res = np.array(np.array_split(X.sum(axis=0),window_size)).sum(axis=1)

output:

>> [14 22]

Answer 2

You can reshape to 2 columns in "Fortran" order and sum:

X = np.array([[1, 2, 3, 4],
              [5, 6, 7, 8]])
X.reshape(-1, 2, order='F').sum(0)

output: array([14, 22])

With more columns you should probably generalize with:

X.reshape(-1, X.shape[1]//2, order='F').sum(0)