NumPy split()

import numpy as np

# create a 1-D array
array1 = np.array([0, 1, 2, 3])

# split into two equal length sub-arrays 
subArrays= np.split(array1, 2)

print(subArrays)

'''
Output: [array([0, 1]), array([2, 3])]
'''

split() Syntax

The syntax of split() is:

numpy.split(array, indices_or_sections, axis = 0)

split() Arguments

The split() method takes the following arguments:

• array - the array to split
• indices - defines indices (as int or 1-D array) in which the array is split
• axis(optional) - axis along which the array is to be split

split() Return Value

The split() method returns the list of sub-arrays.

Example 1: Split an Array into Three Arrays

import numpy as np

array1 = np.array( [[1, 2], [3, 4], [5, 6]] )

# split into 3 arrays
splitArrays = np.split(array1, 3)

print(splitArrays)

Output

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

Example 2: Split an Array by Index

import numpy as np

array1 = np.array( [1, 2, 3, 4, 5, 6] )

# indices at which array is split
splitIndices = [2, 5, 8]

# split into subarrays 
splitArrays = np.split(array1, splitIndices)

print(splitArrays)

Output

[array([1, 2]), array([3, 4, 5]), array([6]), array([], dtype=int64)]

When an index exceeds the dimension of the array, the function returns an empty subarray.

Example 3: Split an Array Across Different Axes

The third parameter is used to split NumPy arrays across different axes. By default, axis is set to 0 (column-wise).

import numpy as np

# create a 3D array
array1 = np.array([[[1, 2], [3, 4]], 
                                [[5, 6], [7, 8]]])

# split the array column-wise (axis = 0 by default)
splitArrays = np.split(array1, 2)

print("\nSplit array column-wise:")
print(splitArrays)

# split the array row-wise
splitArrays = np.split(array1, 2, axis = 1)

print("Split array row-wise:")
print(splitArrays)

# split the array depth-wise
splitArrays = np.split(array1, 2, axis = 2)

print("\nSplit array depth-wise:")
print(splitArrays)

Output

Split array column-wise:
[array([[[1, 2],
        [3, 4]]]), array([[[5, 6],
        [7, 8]]])]
Split array row-wise:
[array([[[1, 2]],

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

       [[7, 8]]])]

Split array depth-wise:
[array([[[1],
        [3]],

       [[5],
        [7]]]), array([[[2],
        [4]],

       [[6],
        [8]]])]
> 

Example 4: Split an Array into Uneven Sub-arrays

The split() function cannot split an array into uneven arrays.

import numpy as np

array1 = np.array( [[1, 2], [3, 4], [5, 6]] )

# split into 2 arrays
splitArrays = np.split(array1, 2)

print(splitArrays)

Output

ERROR!
raise ValueError(
ValueError: array split does not result in an equal division

To split an array into unequal subarrays, you can use array_split() instead. Let's see an example.

import numpy as np

array1 = np.array( [[1, 2], [3, 4], [5, 6]] )

# split into 2 arrays
splitArrays = np.array_split(array1, 2)

for array in splitArrays:
    print(array,'\n')

Output

[[1 2]
 [3 4]] 

[[5 6]] 

Related Methods

We can also split an array along different axes using the following methods:

• NumPy hsplit() - split an array into multiple sub-arrays horizontally
• NumPy vsplit() - split an array into multiple sub-arrays vertically
• NumPy dsplit() - split an array into multiple sub-arrays along the 3rd axis (depth)

Let's see an example.

import numpy as np

# create a 3D array
array1 = np.array([[[1, 2], [3, 4]], 
                                [[5, 6], [7, 8]]])

# split the array column-wise
splitArrays = np.vsplit(array1, 2)

print("\nSplit array column-wise:")
for arr in splitArrays:
    print(arr)
    
# split the array row-wise
splitArrays = np.hsplit(array1, 2)

print("Split array row-wise:")
for arr in splitArrays:
    print(arr)

# split the array depth-wise
splitArrays = np.dsplit(array1, 2)

print("\nSplit array depth-wise:")
for arr in splitArrays:
    print(arr)

Output

Split array column-wise:
[[[1 2]
  [3 4]]]
[[[5 6]
  [7 8]]]
Split array row-wise:
[[[1 2]]

 [[5 6]]]
[[[3 4]]

 [[7 8]]]

Split array depth-wise:
[[[1]
  [3]]

 [[5]
  [7]]]
[[[2]
  [4]]

 [[6]
  [8]]]