What is NumPy Array Splitting?

Splitting is a reverse operation of joining. Joining merges sequence of arrays where splitting splits or breaks one array into multiple arrays.NumPy supports two methods termed as split() method and array_split().

method to break a single array into multiple arrays. We will discuss both in this article. 

  a. split(): 

  syntax:      

split(ary , indices_or_sections , axis)

 array: ndarray (array we want to split into sub-arrays)

 indices_or_sections : integer or 1-D array.

 if is an integer N, the array will be divided into N equal arrays along the axis. If such a split is not possible, an error will be raised.

  If the 1-D array of sorted integers, the entries indicate where along the axis the array is split. For example: [1,2]would for axis=0, result in

• arry[:1]
• arry[1,2]
• arry[2:]         

If the array index exceeds the dimension along the axis, an empty sub-array is returned.

 Axis: int or optional (the axis along which splitting will perform default 0)

 It returns sub-arrays and if the indices_or_section is an integer value but the split is not possible then ValueError is raised.

Example Code: 

   #python code to split an array when indices_or_sections is an integer.

 import numpy as np

 x =np.arange(10)

 print(np.split(x,5))

 #output:

 #[array([0, 1]), array([2, 3]), array([4, 5]), array([6, 7]), array([8, 9])]

#python code to split an array when indices_or_sections is a 1-D array.

import numpy as np

x =np.arange(10)

print(np.split(x,[4,6]))

#output: [array([0, 1, 2, 3]), array([4, 5]), array([6, 7, 8, 9])]

b). array_split(): 

 This function divides a single array into multiple sub-arrays. It accepts the array we want to split, indices_or_sections, and the axis along which splitting will be performed. 

  Syntax:

 array_split(arry , indices_or_sections , axis)

 array: ndarray (array we want to split into sub-arrays)

           indices_or_sections: int or 1-D

If is an integer N, the array will be divided into N equal arrays along the axis. If such a split is not possible, an error will be raised.If a 1-D array of sorted integers, the entries indicate where along the axis the array is split. For example: [1,2]would for axis=0, result in

• arry[:1]
• arry[1,2]
• arry[2:]

   If the array index exceeds the dimension along the axis, an empty sub-array is returned.

Axis: int or optional (the axis along which splitting will be performed default 0)

The difference between the split() and array_split() method is that when the elements are less in the source array for splitting, array_split() worked properly but the split() method would fail...

For Example:

  # This code demonstrate array_split() method working

import numpy as np

x =np.arange(11)

print(np.array_split(x,6))

#output: [array([0, 1]), array([2, 3]), array([4, 5]), array([6, 7]), 

array([8, 9]), array([10])]

# This code demonstrates the split() method works.

import numpy as np

x =np.arange(11)

print(np.split(x,6))

#output: 
raise ValueError(

ValueError: array split does not result in an equal division)

       

 

  # python code to split a 2-D array into two 2-D sub-arrays along the default axis.   

import numpy as np

arry =np.array([[3,7,89,12], [56,80,81,13],[1,2,6,4],[12,34,56,78]])

print(np.array_split(arry,2))

#output:[array([[ 3, 7, 89, 12],

       [56, 80, 81, 13]]), array([[ 1, 2, 6, 4],

       [12, 34, 56, 78]])]

# python code to split a 2-D array into two 2-D sub-arrays along 

axis = 1.   

import numpy as np

arry =np.array([[3,7,89,12], [56,80,81,13],[1,2,6,4],[12,34,56,78]])

print(np.array_split(arry,2, axis=1))

#output:[array([[ 3, 7],

       [56, 80],

       [ 1, 2],

       [12, 34]]), array([[89, 12],

       [81, 13],

       [ 6, 4],

       [56, 78]])]

Note: similar alternates to vstack() , dstack() are available as 

vsplit() , dsplit().