Summary  
Demonstrates how to extract all distinct entries from a data column and count them using unique() and nunique() methods.

General domain of usage  
Exploratory data analysis

Data often gets **duplicated** in DataFrames. For instance, in the `countries` DataFrame, the `'continent'` column has repeated entries. There's a method that retrieves an **array of distinct values** from a specific DataFrame column.

import pandas as pd 

country_data = {'country' : ['Thailand', 'Philippines', 'Monaco', 'Malta', 'Sweden', 'Paraguay', 'Latvia'],
          'continent' : ['Asia', 'Asia', 'Europe', 'Europe', 'Europe', 'South America', 'Europe'],
          'capital':['Bangkok', 'Manila', 'Monaco', 'Valletta', 'Stockholm', 'Asuncion', 'Riga']}
countries = pd.DataFrame(country_data)
print(countries)

Next, apply the `unique()` method to the `'continent'` and `'country'` columns:


import pandas as pd

country_data = {'country' : ['Thailand', 'Philippines', 'Monaco', 'Malta', 'Sweden', 'Paraguay', 'Latvia'],
          'continent' : ['Asia', 'Asia', 'Europe', 'Europe', 'Europe', 'South America', 'Europe'],
          'capital':['Bangkok', 'Manila', 'Monaco', 'Valletta', 'Stockholm', 'Asuncion', 'Riga']}
countries = pd.DataFrame(country_data)
unique_countries = countries['country'].unique()
unique_continents = countries['continent'].unique()
print(unique_countries)
print(unique_continents)

Sometimes, instead of retrieving all unique values, you might just want to know **how many distinct values** exist in a column. In such cases, you can use the **`nunique()`** method. It returns the **count** of unique entries in a column — not the values themselves.

import pandas as pd

country_data = {'country' : ['Thailand', 'Philippines', 'Monaco', 'Malta', 'Sweden', 'Paraguay', 'Latvia'],
          'continent' : ['Asia', 'Asia', 'Europe', 'Europe', 'Europe', 'South America', 'Europe'],
          'capital':['Bangkok', 'Manila', 'Monaco', 'Valletta', 'Stockholm', 'Asuncion', 'Riga']}
countries = pd.DataFrame(country_data)
print(countries['continent'].nunique())

import unittest
import importlib
import numpy as np


# Helper function for dynamic test names and assertions
def _dynamic_test(test_case, condition, success_message, failure_message):
    if condition:
        test_case._testMethodName = success_message
        test_case.assertTrue(True, success_message)
    else:
        test_case._testMethodName = failure_message
        test_case.fail(failure_message)


class TestUserCode(unittest.TestCase):

    def test_unique_years_is_declared(self):
        import user_code
        _dynamic_test(
            self,
            hasattr(user_code, 'unique_years'),
            "The `unique_years` variable is declared.",
            "Expected `unique_years` to be declared."
        )

    def test_unique_years_contains_correct_values(self):
        import user_code
        expected_value = sorted([2017, 2016, 2019])
        try:
            if isinstance(user_code.unique_years, np.ndarray):
                actual_value = sorted(user_code.unique_years)
                condition = actual_value == expected_value
                failure_message = f"Expected `unique_years` to contain `{expected_value}`, but got `{actual_value}`."
            else:
                condition = False
                failure_message = f"`unique_years` is not a `numpy.ndarray`."

        except AttributeError:
            condition = False
            failure_message = "The `unique_years` variable is not declared."

        _dynamic_test(
            self,
            condition,
            f"`unique_years` contains correct values.",
            failure_message
        )

    def test_unique_fueltype_is_declared(self):
        import user_code
        _dynamic_test(
            self,
            hasattr(user_code, 'unique_fueltype'),
            "The `unique_fueltype` variable is declared.",
            "Expected `unique_fueltype` to be declared."
        )

    def test_unique_fueltype_contains_correct_values(self):
        import user_code
        expected_value = sorted(["petrol", "diesel"])
        try:
            if isinstance(user_code.unique_fueltype, np.ndarray):
                actual_value = sorted(user_code.unique_fueltype)
                condition = actual_value == expected_value
                failure_message = f"Expected `unique_fueltype` to contain `{expected_value}`, but got `{actual_value}`."
            else:
                condition = False
                failure_message = f"`unique_fueltype` is not a `numpy.ndarray`."
        except AttributeError:
            condition = False
            failure_message = "The `unique_fueltype` variable is not declared."

        _dynamic_test(
            self,
            condition,
            f"`unique_fueltype` contains correct values.",
            failure_message
        )

    def test_count_unique_fueltypes_is_declared(self):
        import user_code
        _dynamic_test(
            self,
            hasattr(user_code, 'count_unique_fueltypes'),
            "The `count_unique_fueltypes` variable is declared.",
            "Expected `count_unique_fueltypes` to be declared."
        )

    def test_count_unique_fueltypes_is_correct(self):
        import user_code
        expected_value = 2
        try:
            actual_value = user_code.count_unique_fueltypes
            condition = actual_value == expected_value
            failure_message = f"Expected `count_unique_fueltypes` to be `{expected_value}`, but got `{actual_value}`."
        except AttributeError:
            condition = False
            failure_message = "The `count_unique_fueltypes` variable is not declared."

        _dynamic_test(
            self,
            condition,
            f"`count_unique_fueltypes` is correct.",
            failure_message
        )


if __name__ == '__main__':
    unittest.main()

test_main.py

Pandas is an extremely user-friendly library for data analysis. It's also designed to handle large datasets, using data structures like DataFrame and Series. This makes it an invaluable tool for Data Science. In this guide, you'll get acquainted with a range of statistical functions, including how to find correlations, modes, medians, and maximum and minimum values within a dataset. You'll also learn how to handle missing values and manipulate specific values, as well as how to remove them.

Dive into the fundamentals of the pandas library. Learn to create and manipulate Series and DataFrames, explore their structure, and build confidence in performing essential data operations.

Master how to load and manage data from CSV and TXT files. Gain practical experience importing datasets and preparing them for further analysis using pandas tools.

Learn to explore, clean, and summarize data effectively. Understand how to handle missing values, extract insights, and perform basic statistical operations with pandas.

Unique Values

解答