[FIXED] Creating expanded 1D array from 1D array elements without for loop

Issue

I want to create an expanded 1D array from elements of 1D array using some numpy operation and without using a for loop. I would essentially read in the elements of 1D array and count up a few numbers(pre-determined value) from it and proceed to the next element to expand. See an example below.

1D array = [1 9 20 56 78 120]

expanded 1D array = [1 2 3 9 10 11 20 21 22 56 57 58 78 79 80 120 121 122]

is this possible? Thank you

I honestly don't have any idea how to do this without using a for loop

Solution

a = np.array([1, 9, 20, 56 ,78, 120])
n = 3
out = (np.arange(n) + a[:,None]).ravel()

Edit: explanation

a[:, None] is a shortcut for a.reshape(a.size, 1). It creates a 2 dimension array. When I want to sum with np.arange(n), this array is promoted to shape (1, n). Then numpy broadcasts all 1 to the matching dimension.

np.arange(n) + a[:,None] has shape (a.size, n). Then ravel is used to flatten it as a 1D array.

Answered By - Olivier Gauthé