How can I write code that adds up all the numbers in between the indices? The arrays numbers and indices are correlated. The first two index values are 0-3 so the numbers between indices 0-3 are being added up 1 + 5 + 6 = 12. The expected value is what I am trying to find. I am trying to get the results without using a for loop.
numbers = np.array([1, 5, 6, 7, 4, 3, 6, 7, 11, 3, 4, 6, 2, 20]
indices = np.array([0, 3 , 7, 11])
Expected output:
[12, 41, 22]
I'm not sure how you got the expected output - from my calculation, the sum between the indices should be [12, 20, 25]. The following code calculates this:
numbers = np.array([1, 5, 6, 7, 4, 3, 6, 7, 11, 3, 4, 6, 2, 20])
indexes = np.array([0, 3, 7, 11])
tmp = np.zeros(len(numbers) + 1)
np.cumsum(numbers, out=tmp[1:])
result = np.diff(tmp[indexes])
The output of this is [12, 20, 25]
How does this work? It creates an array that is just one size larger than the numbers array (in order to have the first element be zero). Then it calculates the cumulative sum of the elements, starting at index 1 of the tmp array. Then, it takes the diff of the tmp array at the indices provided. As an example, it takes the different of the cumulative sum of the array from index 3 (value = 12) to index 7 (value = 32), 32-12 = 20.
You are likely looking for np.add.reduceat:
>>> np.add.reduceat(numbers, indices)
array([12, 20, 25, 28], dtype=int32)