Numpy iterative array operation

Question

I have started to remove more and more of my for loops by using numpy array operations instead. I am however stuck on the following case. Any help welcome.

I start with a single known value (0.55) in an array A of length L and another array B of length L as well. I want to populate the remaining values in the first array using cross multiplication.

This gives me the desired output:

def cross_mult(B, starting_A = 0.55):
    A= np.zeros(B.shape)
    A[0] = starting_A   
    for i in range(B.shape[0])[1:]:
        A[i] = A[i-1] * B[i] / B[i-1]
    return A

This attempt without the for loop fails:

def cross_mult(B, starting_A = 0.55):
    A= np.zeros(B.shape)
    A[0] = starting_A 
    A[1:] = A[:-1] * B[1:] / B[:-1]
    return A

I get:

array([0.55      , 0.60401715, 0.       ])

Instead of a fully populated array with the three values in it.

Patrick · Accepted Answer · 2019-05-02 14:16:13Z

1

The issue with vectorizing this function is that the calculation at each index actually does depend on the index before it, so it's necessary for the calculations to happen sequentially.

To my knowledge, you're stuck using a loop in this situation :)

If you really want to avoid using a loop explicitly though, you can use accumulate:

def cross_mult(B, starting_A = 0.55):
    A= np.zeros(B.shape)
    A[0] = starting_A 
    A[1:] = B[1:] / B[:-1]
    return np.multiply.accumulate(A)

edited May 2, 2019 at 14:16

answered May 2, 2019 at 13:45

Patrick

6108 silver badges11 bronze badges

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

Collectives™ on Stack Overflow

Numpy iterative array operation

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related