Concatenate/stack unpacked numpy array in certain axis to other arrays

Question

I want to create a numpy array variables with a certain structure, as in the example below, consisting of the variables X and Y for a meshgrid and additional parameters params. I need this in the context of plotting a 3D surface plot for a function f(variables).

This is how the variables array should look like with some arbitrary params 5, 5, 5.

import numpy as np

X = np.linspace(0, 2, num=3)
Y = np.linspace(0, 2, num=3)
X, Y = np.meshgrid(X, Y)
variables = np.array([X, Y, 5, 5, 5])
print(variables)

With the desired output:

array([[[ 0.,  1.,  2.],
        [ 0.,  1.,  2.],
        [ 0.,  1.,  2.]],
       [[ 0.,  0.,  0.],
        [ 1.,  1.,  1.],
        [ 2.,  2.,  2.]],
       5, 5, 5])

And my questions is how can I construct this when having the parameters 5, 5, 5 in a numpy array also. (And explicitly not having something like [X, Y, [5, 5, 5]].)

params = np.array([5, 5, 5])

I already looked at different numpy functions like stack or concatenate, but could not figure it out. I'm rather new to numpy, so thanks for any help.

If np.array([X, Y, 5, 5, 5]) isn't your expected o/p, can you list out the same for that sample? — Divakar
– Divakar, Commented Sep 15, 2016 at 11:38
Yeah it would be something like variables = np.array([X, Y, params]) which would yield the structure stated above variables = np.array([X, Y, 5, 5, 5]). Did I explain myself? — mo_blu
– mo_blu, Commented Sep 15, 2016 at 11:42
I think it would be better if you could write down the expected o/p for that sample. — Divakar
– Divakar, Commented Sep 15, 2016 at 11:43
In other words, show what print(result) ought to look like for a small example. — senderle
– senderle, Commented Sep 15, 2016 at 11:43

Eric · Accepted Answer · 2016-09-15 13:30:53Z

2

You can't get there from here.

Numpy arrays are for storing cuboidal grids of data, where each entry is the same shape as every other entry.

The best you can do here is have a (5,)-grid, containing objects, where each of the first two objects is itself a numpy array:

params = np.array([5, 5, 5])
XY = np.array([X, Y, None])[:-1]  # trick numpy into using a 1d object array
                                  # there might be a better way to do this
variables = np.concatenate((XY, params))
variables

array([array([[ 0.,  1.,  2.],
       [ 0.,  1.,  2.],
       [ 0.,  1.,  2.]]),
       array([[ 0.,  0.,  0.],
       [ 1.,  1.,  1.],
       [ 2.,  2.,  2.]]),
       5, 5, 5], dtype=object)

Are you sure you need a numpy array at all here? Is this enough:

variables = (X, Y) + tuple(params)

answered Sep 15, 2016 at 13:30

Eric

98.1k54 gold badges257 silver badges389 bronze badges

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

1 Comment

mo_blu Over a year ago

Wow, both parts of your answer are possible solutions for me and also very helpful for understanding. So, thank you very much!

Collectives™ on Stack Overflow

Concatenate/stack unpacked numpy array in certain axis to other arrays

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related