Adding Numpy Arrays across dictionaries

Question

I have a list of dictionaries with identical keys, where the values of each dictionary are NumPy arrays of the same size.

Here's the example

a = {"a": np.array([1,2]),"b": np.array([2,4]),}
b = {"a": np.array([5,7]),"b": np.array([12,34]),}
c = [a,b]

I would like to sum the values(Numpy arrays) for the same keys across all of the dictionaries.

I have tried Counter form collections but it only seems to work for scalar values. I tried to do it with two loops but got confused. I probably can do it with three loops but is there another more elegant way?

Thanks

Quang Hoang · Accepted Answer · 2021-03-23 19:08:03Z

2

Try with:

{k:a[k]+b[k] for k in a.keys()}

Or in general with reduce:

from functools import reduce
reduce(lambda x,y: {k:x[k]+y[k] for k in x.keys()}, c)

Output:

{'a': array([6, 9]), 'b': array([14, 38])}

answered Mar 23, 2021 at 19:08

Quang Hoang

151k11 gold badges64 silver badges86 bronze badges

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

1 Comment

Yury Moskaltsov Over a year ago

Yes, the reduce formula worked. It seems slightly unclear to me how it works, but I guess I will check it out. Thanks.

norok2 · Accepted Answer · 2021-03-23 19:28:07Z

1

You can just use sum() and dict-comprehension:

keys = a.keys()
{k: sum(x[k] for x in c) for k in keys}

This assumes that the keys in a are found in all dicts, otherwise you would need to check for key existence (e.g. sum(x[k] for x in c if k in x)).

The full example reads:

import numpy as np


a = {"a": np.array([1,2]),"b": np.array([2,4]),}
b = {"a": np.array([5,7]),"b": np.array([12,34]),}
c = [a,b]


keys = a.keys()
{k: sum(x[k] for x in c) for k in keys}
# {'a': array([6, 9]), 'b': array([14, 38])}

answered Mar 23, 2021 at 19:28

norok2

27.1k6 gold badges83 silver badges110 bronze badges

2 Comments

Yury Moskaltsov Over a year ago

nice and elegant

norok2 Over a year ago

keep in mind that reduce() may be faster as one of the loops is in C.

Collectives™ on Stack Overflow

Adding Numpy Arrays across dictionaries

2 Answers 2

1 Comment

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related