How can you fill a NumPy array with another array (possibly with different shape)

Question

Not sure how to ask my question, maybe the title doesn't reflect my question correctly, and if that is the case feel free to correct me.

I have two arrays

a = np.zeros((2, 2))
b = np.ones((3, 3))

I want a func that would yield the following result:

c = func(a, b)
print(c)

# array([[1., 1.],
#       [1., 1.]])

But if we switch position of a and b when we pass them to the function, we will get the following result:

c = func(b, a) # Note the change
print(c)

# array([[0., 0., 1.],
#       [0., 0., 1.],
#       [1., 1., 1.]])

Edit: for clarification, here is another example to answer @mercury's question.

a = np.zeros((2, 3))
b = np.ones((3, 2))

c = func(a, b)
print(c)

# array([[1., 1., 0.],
#       [1., 1., 0.]])

Some clarifications: what if a has shape (2,3) and b has shape (3,2) (Each array is bigger in a different dimension). Or are the arrays always nxn square arrays? — Mercury
– Mercury, Commented Aug 16, 2020 at 10:28
@Mercury I added another example to the post that should answer your question — Tomergt45
– Tomergt45, Commented Aug 16, 2020 at 10:31
So you want to keep the shape of the first array and replace all elements of it with the elements of the second array, if the second array has these elements? — Niklas Mertsch
– Niklas Mertsch, Commented Aug 16, 2020 at 11:02

Tomergt45 · Accepted Answer · 2020-08-16 10:41:32Z

1

I think I found an answer, not sure about it tho.

def func(a, b):
    x1, x2 = np.min((a.shape, b.shape), 0)
    c = a.copy()
    c[:x1, :x2] = b[:x1, :x2]
    return c

I'll be happy if someone who understands the question could reply if this actually solves it, and if so, how should this operation be called?

answered Aug 16, 2020 at 10:41

Tomergt45

6791 gold badge9 silver badges25 bronze badges

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

4 Comments

Niklas Mertsch Over a year ago

Looks good to me. If I understood your question correctly, something like this would be my answer.

Tomergt45 Over a year ago

@NiklasMertsch thank you for your answer, how would you call this operation btw? I thought something along the lines of fill_array but I'm not sure if it fully reflects the idea of the function.

Niklas Mertsch Over a year ago

This is your brainchild, name it however you want. If you solve some specific problem with it, maybe that problems gives you some inspiration. fill_existing(base, fill)?

Tomergt45 Over a year ago

@NiklasMertsch Thank you for your answer. If you are familiar with neural networks, I am using this function to connect the weights of two layers in a neural network while compensating for missing connections (for genetic a genetic algorithm).

Collectives™ on Stack Overflow

How can you fill a NumPy array with another array (possibly with different shape)

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related