Negative indexing in NumPy

Introduction

Negative indexing helps you access elements from the end of a list or array easily. It saves time when you want to get last items without counting the total length.

You want the last element of a data array without knowing its length.

You need to slice the last few rows of a dataset quickly.

You want to access elements from the end in a loop or function.

You are working with time series data and want recent values.

You want to reverse access without creating a reversed copy.

Syntax

NumPy

array[-index]

Index -1 means the last element, -2 means second last, and so on.

This works the same way for 1D and multi-dimensional arrays.

Examples

Prints the last element of the array, which is 50.

NumPy

import numpy as np
arr = np.array([10, 20, 30, 40, 50])
print(arr[-1])

Prints the third last element, which is 30.

NumPy

print(arr[-3])

Accesses the last row and second last column, printing 8.

NumPy

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

Sample Program

This program shows how to use negative indexing to get elements from the end of 1D and 2D numpy arrays.

NumPy

import numpy as np

# Create a 1D array
numbers = np.array([5, 10, 15, 20, 25])

# Access last element using negative index
last_element = numbers[-1]

# Access second last element
second_last = numbers[-2]

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

# Access element from last row, last column
element = matrix[-1, -1]

print(f"Last element in numbers: {last_element}")
print(f"Second last element in numbers: {second_last}")
print(f"Element from last row and last column in matrix: {element}")

Important Notes

Negative indexing is very handy for quick access but be careful not to use an index smaller than -length, or it will cause an error.

It works well with slicing too, like arr[-3:] to get last three elements.

Summary

Negative indexing lets you count from the end of arrays easily.

-1 is the last element, -2 is second last, and so on.

It works for both 1D and multi-dimensional numpy arrays.