Summing through an array containing other arrays

Question

I would like to sum through an array made of other sub-arrays.

import numpy as np

a1 = np.array([1,2,3])
a2 = np.array([3,4,5])
a3 = np.array([6,7,8])

a = np.concatenate(([a1],[a2],[a3]))

print(len(a))

for i in range(1,len(a)):                                 (1)
    for j in i :                                          (2)
        for k in j :                                      (3)
            goal = np.sum(np.cos(k))
        print(goal)
        
   
        
goal  to achieve  : 

    goal = cos(1)+cos(2)+cos(3) 
    goal = cos(3)+cos(4)+cos(5) 
    goal = cos(6)+cos(7)+cos(8)

I thought of looping first on the array containing every other sub-array (1) , then for each sub-array (2) and finally inside each sub array (3) .

I'll get a traceback :

File "*****************", line 18, in <module>
    for j in i :

TypeError: 'int' object is not iterable

In my real problem, I might have more than 3 sub-arrays. I think it can be done but I'm still discovering Python.

Barmar · Accepted Answer · 2021-12-10 00:40:59Z

1

You have far too many levels of looping. You just need one loop to iterate over the rows of the array, then you can sum that row.

for row in a:
    goal = np.sum(np.cos(row))
    print(goal)

answered Dec 10, 2021 at 0:40

Barmar

789k57 gold badges555 silver badges669 bronze badges

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

3 Comments

pho Over a year ago

np.sum takes an axis argument that's much faster than looping over rows

Barmar Over a year ago

@PranavHosangadi I suspected there was a numpy one-liner, but I don't know numpy well. I suggest you post that answer.

Charles Over a year ago

Simple, straightforward. Thank you very much for your answer !

hpaulj · Accepted Answer · 2021-12-10 01:50:33Z

1

In [212]: a1 = np.array([1,2,3])
     ...: a2 = np.array([3,4,5])
     ...: a3 = np.array([6,7,8])
     ...: 
     ...: a = np.concatenate(([a1],[a2],[a3]))
     ...:

The resulting 2d array:

In [213]: a
Out[213]: 
array([[1, 2, 3],
       [3, 4, 5],
       [6, 7, 8]])

cos of each element:

In [214]: np.cos(a)
Out[214]: 
array([[ 0.54030231, -0.41614684, -0.9899925 ],
       [-0.9899925 , -0.65364362,  0.28366219],
       [ 0.96017029,  0.75390225, -0.14550003]])

sum rows:

In [215]: np.cos(a).sum(axis=1)
Out[215]: array([-0.86583703, -1.35997393,  1.56857251])

it's easier to verify that this is summing rows with:

In [216]: a.sum(axis=1)
Out[216]: array([ 6, 12, 21])

With numpy it's best to think in terms of the whole multidimensional array, and actions that work on the whole thing, or on specific dimensions. This is both faster and generally more useful.

answered Dec 10, 2021 at 1:50

hpaulj

233k14 gold badges260 silver badges392 bronze badges

1 Comment

Charles Over a year ago

Thank you very much for this explaination, I start to understand how numpy is working.

Sakhund · Accepted Answer · 2021-12-10 00:56:40Z

0

Instead of j in i write j in a[i]

edited Dec 10, 2021 at 0:56

answered Dec 10, 2021 at 0:47

Sakhund

2621 gold badge7 silver badges33 bronze badges

7 Comments

Barmar Over a year ago

Where do you call np.cos()?

Barmar Over a year ago

He doesn't want the sum of everything. He wants the sum of the cosines of each row.

Sakhund Over a year ago

Oh so the problem occurs when you try to cos and the list has another list?

Barmar Over a year ago

Read the part of the question after "goal to achieve"

Sakhund Over a year ago

Now I see whats wrong

|

Collectives™ on Stack Overflow

Summing through an array containing other arrays

3 Answers 3

3 Comments

1 Comment

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

1 Comment

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related