Trying to minimize code to find sum of sub_arrays of a numpy array

Question

I have a Numpy array of the following form:

a = [[1,2,3],[4,5,6],[3,4,2]]

I want to sum all the sublists in the following form:

b = [6,15,9]

I have the following code that does it:

ss = numpy.zeros(shape = [])
for item in a:
    print item
    s = item.sum()
    print s
    b = np.append(ss,s)

Here is the result: b = [6,15,9]

Can I do it without this explicit for loop? As in is/are there a numpy function(s) that can make my life easy? This is because the array a is pretty big ~ 10^6 entries.

Are the sub_arrays of the same length? If so then np.array(a) is a 2d array, and you can sum over axis=1. Otherwise some sort of loop is required, with a list comprehension being the most compact: [sum(x) for x in [[1,2,3],[4,5,6],[3,4,2]]]. — hpaulj
– hpaulj, Commented Jan 4, 2014 at 6:08

score 3 · Accepted Answer · 2014-01-04 05:07:11Z

3

ss = a.sum(1)

This will do if a is a numpy array. If it is a list of numpy arrays or a list-of-lists you want to use:

ss = numpy.sum(a, 1)

Thanks @DSM for improvement.

edited Jan 4, 2014 at 5:07

answered Jan 4, 2014 at 4:59

user3072164

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

3 Comments

DSM Over a year ago

If a is an ndarray, you could also use the method rather than the function, i.e. a.sum(1). numpy.sum(a, 1) will actually work even if a is a list-of-lists.

Abhinav Kumar Over a year ago

@Nabla thanks for your answer. Btw, why use 1 inside the paranthesis in ss = a.sum(1) ?

user3072164 Over a year ago

@AbhinavKumar The number defines the axis of the two-dimensional array along which is to be summed, see the documentation for details.

Collectives™ on Stack Overflow

Trying to minimize code to find sum of sub_arrays of a numpy array

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related