numpy.stack() in Python (original) (raw)
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
a = np.array([ 1 , 2 , 3 ])
b = np.array([ 4 , 5 , 6 ])
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
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
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
x = np.array([[ 1 , 2 , 3 ],
`` [ 4 , 5 , 6 ]])
y = np.array([[ 7 , 8 , 9 ],
`` [ 10 , 11 , 12 ]])
1. stacking with axis=0
Python3
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
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
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
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
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
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 ]]])
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
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
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
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]]]])