np.unique() for unique values in NumPy

Introduction

We use np.unique() to find all the different values in a list or array. It helps us see what unique items are there without repeats.

When you want to know all the different colors used in a photo's pixel data.

When you have survey answers and want to see all the unique responses.

When cleaning data to find and remove duplicate entries.

When counting how many different categories appear in a dataset.

Syntax

NumPy

numpy.unique(ar, return_index=False, return_inverse=False, return_counts=False, axis=None)

ar is the input array or list.

You can ask for extra info like where the unique values came from or how many times they appear by setting return_index, return_inverse, or return_counts to True.

Examples

This finds unique values in a simple list and prints them sorted.

NumPy

import numpy as np
arr = [1, 2, 2, 3, 4, 4, 4]
unique_values = np.unique(arr)
print(unique_values)

This finds unique values and how many times each appears.

NumPy

arr = np.array([5, 3, 5, 2, 3])
unique_vals, counts = np.unique(arr, return_counts=True)
print(unique_vals)
print(counts)

This finds unique rows in a 2D array.

NumPy

arr = np.array([[1, 2], [2, 3]])
unique_axis0 = np.unique(arr, axis=0)
print(unique_axis0)

Sample Program

This program shows how to get unique values from an array and also count how many times each unique value appears.

NumPy

import numpy as np

# Sample data with repeated values
data = np.array([10, 20, 20, 30, 10, 40, 50, 50, 50])

# Find unique values
unique_vals = np.unique(data)
print("Unique values:", unique_vals)

# Find unique values and their counts
unique_vals, counts = np.unique(data, return_counts=True)
print("Counts of each unique value:", counts)

Important Notes

np.unique() always returns sorted unique values.

Use return_counts=True to see how often each unique value appears.

For multi-dimensional arrays, use the axis parameter to find unique rows or columns.

Summary

np.unique() helps find all different values in data.

You can get extra info like counts or original positions.

It works on lists and arrays, including multi-dimensional ones.