Numpy Array Slicing

Summary: in this tutorial, you’ll learn about the numpy array slicing that extracts one or more elements from a numpy array.

Numpy array slicing on on-dimensional arrays

NumPy arrays use brackets [] and : notations for slicing like lists. By using slices, you can select a range of elements in an array with the following syntax:

[m:n]Code language: Python (python)

This slice selects elements starting with m and ending with n-1. Note that the nth element is not included. In fact, the slice m:n can be explicitly defined as:

[m:n:1]Code language: Python (python)

The number 1 specifies that the slice selects every element between m and n.

To select every two elements, you can use the following slice:

[m:n:2]Code language: Python (python)

In general, the following expression selects every k element between m and n:

[m:n:k]Code language: Python (python)

If k is negative, the slice returns elements in reversed order starting from m to n+1. The following table illustrates the slicing expressions:

See the following example:

import numpy as np

a = np.arange(0, 10)

print('a=', a)
print('a[2:5]=', a[2:5])
print('a[:]=', a[:])
print('a[0:-1]=', a[0:-1])
print('a[0:6]=', a[0:6])
print('a[7:]=', a[7:])
print('a[5:-1]=', a[5:-1])
print('a[0:5:2]=', a[0:5:2])
print('a[::-1]=', a[::-1])Code language: Python (python)

Output:

a= [0 1 2 3 4 5 6 7 8 9]
a[2:5]= [2 3 4]
a[:]= [0 1 2 3 4 5 6 7 8 9] 
a[0:-1]= [0 1 2 3 4 5 6 7 8]
a[0:6]= [0 1 2 3 4 5]
a[7:]= [7 8 9]
a[5:-1]= [5 6 7 8]
a[0:5:2]= [0 2 4]
a[::-1]= [9 8 7 6 5 4 3 2 1 0]Code language: Python (python)

Numpy array slicing on multidimensional arrays

To slice a multidimensional array, you apply the square brackets [] and the : notation to each dimension (or axis). The slice returns a reduced array where each element matches the selection rules. For example:

import numpy as np

a = np.array([
    [1, 2, 3],
    [4, 5, 6],
    [7, 8, 9]
])


print(a[0:2, :])Code language: Python (python)

Output:

[[1 2 3]
 [4 5 6]]Code language: Python (python)

In this example, array a is a 2-D array. In the expression a[0:2, :]:

First, the 0:2 selects the element at index 0 and 1, not 2 that returns:

[[1 2 3]
[4 5 6]]Code language: Python (python)

Then, the : select all elements. Therefore the whole expression returns:

[[1 2 3]
 [4 5 6]]Code language: Python (python)

Consider another example:

import numpy as np

a = np.array([
    [1, 2, 3],
    [4, 5, 6],
    [7, 8, 9]
])

print(a[1:, 1:])Code language: Python (python)

Output:

[[5 6]
 [8 9]]Code language: Python (python)

In the expression a[1:, 1:]:

First, 1: selects the elements starting at index 1 to the last element of the first axis (or row), which returns:

[[4 5 6]
[7 8 9]]Code language: Python (python)

Second, 1: selects the elements starting at index 1 to the last elements of the second axis (or column), which returns:

[[5 6]
 [8 9]]Code language: Python (python)

Summary

• Use slicing to extract elements from a numpy array
• Use a[m:n:p] to slice one-dimensional arrays.
• Use a[m:n:p, i:j:k, ...] to slice multidimensional arrays