Open In App

numpy.pad() function in Python

Improve
Improve
Like Article
Like
Save
Share
Report

numpy.pad() function is used to pad the Numpy arrays. Sometimes there is a need to perform padding in Numpy arrays, then numPy.pad() function is used. The function returns the padded array of rank equal to the given array and the shape will increase according to pad_width.

Syntax: numpy.pad(array, pad_width, mode=’constant’, **kwargs) 

Parameters :

  • array: the array to pad
  • pad_width: This parameter defines the number of values that are padded to the edges of each axis.
    mode : str or function(optional)
  • **kwargs: allows you to pass keyword variable length of argument to a function. It is used when we want to handle the named argument in a function.

Return:
A padded array of rank equal to an array with shape increased according to pad_width.

Example 1:

Python3




# Python program to explain
# working of numpy.pad() function
import numpy as np
  
  
arr = [1, 3, 2, 5, 4]
  
# padding array using CONSTANT mode
pad_arr = np.pad(arr, (3, 2), 'constant'
                 constant_values=(6, 4))
  
print(pad_arr)


Output:

[6 6 6 1 3 2 5 4 4 4]

Example 2:

Python3




# Python program to explain
# working of numpy.pad() function
import numpy as np
  
  
arr = [1, 3, 2, 5, 4
  
# padding array using 'linear_ramp' mode
pad_arr = np.pad(arr, (3, 2), 'linear_ramp',
                 end_values=(-4, 5))   
  
print(pad_arr)


Output:

[-4 -2 -1  1  3  2  5  4  4  5]

Example 3:

Python3




# Python program to explain
# working of numpy.pad() function
import numpy as np
  
  
arr = [1, 3, 9, 5, 4]
  
# padding array using 'maximum' mode
pad_arr = np.pad(arr, (3,), 'maximum')
  
print(pad_arr)


Output:

[9 9 9 1 3 9 5 4 9 9 9]

Example 4:

Python3




# Python program to explain
# working of numpy.pad() function
import numpy as np
  
  
arr = [[1, 3],[5, 8]] 
  
# padding array using 'minimum' mode
pad_arr = np.pad(arr, (3,), 'minimum')       
  
print(pad_arr)


Output:

[[1 1 1 1 3 1 1 1]
[1 1 1 1 3 1 1 1]
[1 1 1 1 3 1 1 1]
[1 1 1 1 3 1 1 1]
[5 5 5 5 8 5 5 5]
[1 1 1 1 3 1 1 1]
[1 1 1 1 3 1 1 1]
[1 1 1 1 3 1 1 1]]


Last Updated : 01 Oct, 2020
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads