Vectorized way of adding to an array that is indexed by another array - Python/NumPy

Question

I have 3 Numpy arrays, a, b and c. b and c are very large arrays and have the same length. Each element of b is 0, 1 or 2 and also the length of a is 3. Now I wonder if there is a way to eliminate the following for loop:

for i in range(len(b)):
    a[b[i]] += c[i]

Any comment would be greatly appreciated.

You want to eliminate the for loop and still have a modified that way? — Seif
– Seif, Commented Jan 4, 2017 at 16:11
You have to iterate there is no other way, you can maybe merge 2 for loops so you only iterate once, but you cannot avoid it — Seif
– Seif, Commented Jan 4, 2017 at 16:17
Who talked about a ? I said if you have a for loop before that one you can try to put both treatment in the same loop. Show more code if you want help. — Seif
– Seif, Commented Jan 4, 2017 at 16:34
Give a small example so there's no ambiguity as what you expect. — hpaulj
– hpaulj, Commented Jan 4, 2017 at 16:40

user2357112 · Accepted Answer · 2017-01-04 18:13:35Z

3

NumPy ufuncs have an at method for cases like this:

numpy.add.at(a, b, c)

This does what everyone expects a[b] += c to do for an array b of indices before they see that it doesn't work.

answered Jan 4, 2017 at 18:13

user2357112

286k32 gold badges490 silver badges569 bronze badges

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

1 Comment

ali Over a year ago

Thank you very much. It helps me very much.

Divakar · Accepted Answer · 2017-01-04 18:29:38Z

2

You can use np.bincount for such ID based summing, like so -

a += np.bincount(b,c,minlength=a.size)

Runtime test -

In [136]: # Large arrays as inputs
     ...: a = np.random.rand(3)
     ...: c = np.random.rand(10000)
     ...: b = np.random.randint(0,3,10000)
     ...: 
     ...: # Make copies for testing
     ...: a1 = a.copy()
     ...: a2 = a.copy()
     ...: 

In [137]: def bincount_app(a, b, c): # bincount approach as func
     ...:     a += np.bincount(b,c,minlength=a.size)
     ...:     

In [138]: %timeit np.add.at(a1, b, c) # @user2357112's soln
1000 loops, best of 3: 1.29 ms per loop

In [139]: %timeit bincount_app(a2, b, c)
10000 loops, best of 3: 36.6 µs per loop

edited Jan 4, 2017 at 18:29

answered Jan 4, 2017 at 18:09

Divakar

222k19 gold badges273 silver badges374 bronze badges

1 Comment

ali Over a year ago

Thank you very much. I really appreciate your time and help.

Collectives™ on Stack Overflow

Vectorized way of adding to an array that is indexed by another array - Python/NumPy

2 Answers 2

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related