numpy.full_like() in Python

Last Updated : 28 Sep, 2026

numpy.full_like() function is used to create a new NumPy array filled with a specified value while using an existing array as a reference. By default, the new array has the same shape and data type as the reference array. It is useful when:

Python 
import numpy as np

arr = np.array([[1, 2, 3],
                [4, 5, 6]])

result = np.full_like(arr, 7)

print(result)

Output
[[7 7 7]
 [7 7 7]]

Explanation:

  • Input array has 2 rows and 3 columns.
  • New array has the same shape.
  • Output also uses an integer data type.
  • Every element is filled with 7.

Syntax

numpy.full_like( a, f ill_value, dtype=None, order='K', subok=True, shape=None )

Parameters:

  • a: The input array used as a reference for the shape and data type.
  • fill_value: The value used to fill every element of the new array.
  • dtype: Specifies the data type of the output array. By default, the data type of a is used.
  • order: Specifies the memory layout of the output array.
  • subok: Determines whether a subclass of the input array should be preserved.
  • shape: Specifies a different shape for the output array.

Using a Different Fill Value

This parameter determines the value stored in every element.

Python 
import numpy as np

arr = np.array([10, 20, 30, 40])

result = np.full_like(arr, 5)

print(result)

Output
[5 5 5 5]

Explanation:

  • Input array contains four elements.
  • Output array also contains four elements.
  • Every element is initialized with 5.

Changing Data Type

By default, numpy.full_like() uses the data type of the input array. The dtype parameter can be used to change it.

Python 
import numpy as np

arr = np.array([1, 2, 3, 4])

result = np.full_like(arr, 2.5, dtype=float)

print(result)

Output
[2.5 2.5 2.5 2.5]

Explanation:

  • Original array contains integers.
  • dtype=float changes the output data type to floating point.
  • Decimal value 2.5 is retained.

Specifying a Different Shape

Shape parameter can override the shape of the input array.

Python 
import numpy as np

arr = np.array([1, 2, 3, 4])

result = np.full_like(arr, 9, shape=(2, 2))

print(result)

Output
[[9 9]
 [9 9]]

Explanation:

  • Input array has shape (4,).
  • shape=(2, 2) changes the output dimensions.
  • Four elements are arranged into 2 rows and 2 columns.

Difference Between numpy.full() and numpy.full_like()

Main difference is how the shape of the output array is specified. numpy.full() takes the required shape directly.

Python 
import numpy as np

result = np.full((2, 3), 8)
print("Using np.full():", result)

arr = np.array([[1, 2, 3], [4, 5, 6]])
result_like = np.full_like(arr, 8)

print("Using np.full_like():", result_like)

Output
Using np.full(): [[8 8 8]
 [8 8 8]]
Using np.full_like(): [[8 8 8]
 [8 8 8]]

numpy.full_like() vs numpy.zeros_like()

Both functions create an array based on another array. The difference is the value used to initialize the elements.

  • numpy.zeros_like() fills the array with zeros.
  • numpy.full_like() fills the array with a specified value.
Python 
import numpy as np

arr = np.array([[1, 2],
                [3, 4]])

print(np.zeros_like(arr))
print(np.full_like(arr, 6))

Output
[[0 0]
 [0 0]]
[[6 6]
 [6 6]]
Comment

Explore