numpy.stack() in Python

Last Updated : 08 Aug, 2022

NumPy is a famous Python library used for working with arrays. One of the important functions of this library is stack().

Important points:

stack() is used for joining multiple NumPy arrays. Unlike, concatenate(), it joins arrays along a new axis. It returns a NumPy array.
to join 2 arrays, they must have the same shape and dimensions. (e.g. both (2,3)--> 2 rows,3 columns)
stack() creates a new array which has 1 more dimension than the input arrays. If we stack 2 1-D arrays, the resultant array will have 2 dimensions.

Syntax: numpy.stack(arrays, axis=0, out=None)

Parameters:

arrays: Sequence of input arrays (required)
axis: Along this axis, in the new array, input arrays are stacked. Possible values are 0 to (n-1) positive integer for n-dimensional output array. For example, in the case of a resultant 2-D array, there are 2 possible axis options :0 and 1. axis=0 means 1D input arrays will be stacked row-wise. axis=1 means 1D input arrays will be stacked column-wise. We shall see the example later in detail. -1 means last dimension. e.g. for 2D arrays axis 1 and -1 are same. (optional)
out: The destination to place the resultant array.

Example #1 : stacking two 1d arrays

Python

import numpy as np  # input array a = np.array([1, 2, 3]) b = np.array([4, 5, 6])  # Stacking 2 1-d arrays c = np.stack((a, b),axis=0) print(c)

output -

array([[1, 2, 3],
[4, 5, 6]])

Notice, output is a 2-D array. They are stacked row-wise. Now, let's change the axis to 1.

Python

# stack 2 1-d arrays column-wise np.stack((a,b),axis=1)

output -

array([[1, 4],
[2, 5],
[3, 6]])

Here, stack() takes 2 1-D arrays and stacks them one after another as if it fills elements in new array column-wise.

Python

#stacking 2 arrays along -1 axis np.stack((a,b),axis=-1)

output -

array([[1, 4],
[2, 5],
[3, 6]])

-1 represents 'last dimension-wise'. Here 2 axis are possible. 0 and 1. So, -1 is same as 1.

Example #2 : stacking two 2d arrays

Python3

# input arrays x=np.array([[1,2,3], [4,5,6]]) y=np.array([[7,8,9], [10,11,12]])

1. stacking with axis=0

Python3

np.stack((x,y),axis=0)

output -

array([[[ 1, 2, 3],
[ 4, 5, 6]],

[[ 7, 8, 9],
[10, 11, 12]]])

Imagine as if they are stacked one after another and made a 3-D array.

2. stacking with axis=1

Python3

np.stack((x,y),axis=1)

Output - 3D array. 1st dimension has 1st rows. 2nd dimension has 2nd rows. [Row-wise stacking]

array([[[ 1, 2, 3],
[ 7, 8, 9]],

[[ 4, 5, 6],
[10, 11, 12]]])

3. stacking with axis =2

Python3

np.stack((x,y),axis=2)

Output - 3D array. 1st dimension has 1st rows. 2nd dimension has 2nd rows. [Column-wise stacking]

array([[[ 1, 7],
[ 2, 8],
[ 3, 9]],

[[ 4, 10],
[ 5, 11],
[ 6, 12]]])

Example #2 : stacking more than two 2d arrays

1. with axis=0 : Just stacking.

Python3

x=np.array([[1,2,3],             [4,5,6]]) y=np.array([[7,8,9],             [10,11,12]]) z=np.array([[13,14,15],             [16,17,18]])  np.stack((x,y,z),axis=0)

output -

array([[[ 1, 2, 3],
[ 4, 5, 6]],

[[ 7, 8, 9],
[10, 11, 12]],

[[13, 14, 15],
[16, 17, 18]]])

2. with axis =1 (row-wise stacking)

Python3

np.stack((x,y,z),axis=1)

output -

array([[[ 1, 2, 3],
[ 7, 8, 9],
[13, 14, 15]],

[[ 4, 5, 6],
[10, 11, 12],
[16, 17, 18]]])

3. with axis =2 (column-wise stacking)

Python

np.stack((x,y,z),axis=2)

output-

array([[[ 1, 7, 13],
[ 2, 8, 14],
[ 3, 9, 15]],

[[ 4, 10, 16],
[ 5, 11, 17],
[ 6, 12, 18]]])

Example #3 : stacking two 3d arrays

1. axis=0. Just stacking

Python3

#2 input 3d arrays  m=np.array([[[1,2,3],             [4,5,6],             [7,8,9]],              [[10,11,12],             [13,14,15],             [16,17,18]]])  n=np.array([[[51,52,53],             [54,55,56],             [57,58,59]],              [[110,111,112],             [113,114,115],             [116,117,118]]])  # stacking np.stack((m,n),axis=0)

output -

array([[[[ 1, 2, 3],
[ 4, 5, 6],
[ 7, 8, 9]],

[[ 10, 11, 12],
[ 13, 14, 15],
[ 16, 17, 18]]],

[[[ 51, 52, 53],
[ 54, 55, 56],
[ 57, 58, 59]],

[[110, 111, 112],
[113, 114, 115],
[116, 117, 118]]]])

2. with axis=1

Python3

np.stack((m,n),axis=1)

output - Imagine as if the resultant array takes 1st plane of each array for 1st dimension and so on.

array([[[[ 1, 2, 3],
[ 4, 5, 6],
[ 7, 8, 9]],

[[ 51, 52, 53],
[ 54, 55, 56],
[ 57, 58, 59]]],

[[[ 10, 11, 12],
[ 13, 14, 15],
[ 16, 17, 18]],

[[110, 111, 112],
[113, 114, 115],
[116, 117, 118]]]])

3. with axis = 2

Python3

np.stack((m,n),axis=2)

output -

array([[[[ 1, 2, 3],
[ 51, 52, 53]],

[[ 4, 5, 6],
[ 54, 55, 56]],

[[ 7, 8, 9],
[ 57, 58, 59]]],

[[[ 10, 11, 12],
[110, 111, 112]],

[[ 13, 14, 15],
[113, 114, 115]],

[[ 16, 17, 18],
[116, 117, 118]]]])

4. with axis = 3

Python3

np.stack((m,n),axis=3)

output -

array([[[[ 1, 51],
[ 2, 52],
[ 3, 53]],

[[ 4, 54],
[ 5, 55],
[ 6, 56]],

[[ 7, 57],
[ 8, 58],
[ 9, 59]]],

[[[ 10, 110],
[ 11, 111],
[ 12, 112]],

[[ 13, 113],
[ 14, 114],
[ 15, 115]],

[[ 16, 116],
[ 17, 117],
[ 18, 118]]]])

numpy.stack() in Python

jana_sayantan

Improve

Article Tags :

Practice Tags :

python

numpy.stack() in Python

Important points:

Parameters:

Example #1 : stacking two 1d arrays

Example #2 : stacking two 2d arrays

Example #2 : stacking more than two 2d arrays

Example #3 : stacking two 3d arrays

Similar Reads