Array modified via two others in python [duplicate]

Question

Suppose we have three 1-D arrays:

A (say of length 5),
B (the same length, 5 in the example)
C (much longer, say of length 100)

C is initially filled with zeros. A gives indices of C elements which should be changed (they may repeat), and B gives values which should be added to initial zeros of C. For example, if A = [1, 3, 3, 3, 29] and B = [2, 3, 4, 2, 3], C[1] should become 2, C[3] - 9, C[29] - 3; all other C elements should remain 0. I've written it as a for-loop:

for i in range(len(A) - 1):
    C[A[i]] = C[A[i]] + B[i]

But is there a more efficient way to do the same in numpy in the vector form?

I don't think you mean len(A)-1; that will skip the last element (29 & 3), as range doesn't include the upper bound. range(3) == [0,1,2]. — DSM
– DSM, Commented Aug 29, 2013 at 15:36

DSM · Accepted Answer · 2013-08-29 15:44:29Z

3

I think you might be able to use bincount, at least for the 1-D case:

>>> A = np.array([1,3,3,3,29])
>>> B = np.array([2,3,4,2,3])
>>> np.bincount(A, B)
array([ 0.,  2.,  0.,  9.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
        0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,
        0.,  0.,  0.,  3.])

(Aside: duplicated values when using numpy indexing can behave very strangely, and it's very easy to be led astray by the behaviour you see in simple cases. I avoid them entirely, as the behaviour is almost never what I want.)

answered Aug 29, 2013 at 15:44

DSM

355k67 gold badges606 silver badges504 bronze badges

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

Collectives™ on Stack Overflow

Array modified via two others in python [duplicate]

1 Answer 1

Comments

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Linked

Related