nunique() for unique counts in Pandas

Introduction

We use nunique() to count how many different values are in a list or column. It helps us understand variety in data.

Counting how many different products customers bought.

Finding how many unique cities appear in an address list.

Checking how many unique dates are in a sales record.

Seeing how many different categories exist in a dataset.

Syntax

Pandas

DataFrame['column_name'].nunique(dropna=True)

# or for whole DataFrame
DataFrame.nunique(dropna=True)

dropna=True means it ignores empty or missing values when counting.

You can use nunique() on a single column or on the whole DataFrame.

Examples

Count unique cities in the 'City' column.

Pandas

df['City'].nunique()

Count unique values in each column of the DataFrame.

Pandas

df.nunique()

Count unique products including missing values as a unique category.

Pandas

df['Product'].nunique(dropna=False)

Sample Program

This code creates a small table with names, cities, and products. It then counts how many unique names, cities, and products are in the table. For products, it counts missing values as unique if any.

Pandas

import pandas as pd

data = {
    'Name': ['Alice', 'Bob', 'Alice', 'David', 'Bob', None],
    'City': ['New York', 'Los Angeles', 'New York', 'Chicago', 'Chicago', 'Chicago'],
    'Product': ['Book', 'Pen', 'Book', 'Pen', 'Notebook', 'Pen']
}

df = pd.DataFrame(data)

# Count unique names ignoring missing values
unique_names = df['Name'].nunique()

# Count unique cities
unique_cities = df['City'].nunique()

# Count unique products including missing values
unique_products_incl_na = df['Product'].nunique(dropna=False)

print(f"Unique Names: {unique_names}")
print(f"Unique Cities: {unique_cities}")
print(f"Unique Products (including NA): {unique_products_incl_na}")

Important Notes

If you want to count missing values as unique, set dropna=False.

By default, nunique() ignores missing values.

Use nunique() to quickly understand diversity in your data.

Summary

nunique() counts how many different values are in data.

It can be used on one column or the whole DataFrame.

Missing values are ignored by default but can be included.

Practice

1. What does the nunique() function do in pandas?

easy

A. Sorts the values in ascending order

B. Calculates the sum of all values in a DataFrame

C. Replaces missing values with zero

D. Counts the number of unique values in a Series or DataFrame column

Solution

1. Step 1: Understand the purpose of nunique()

   The nunique() function counts how many different (unique) values exist in a pandas Series or DataFrame column.

2. Step 2: Compare with other functions

   Other options describe sum, sorting, or filling missing values, which are not what nunique() does.

3. Final Answer:

   Counts the number of unique values in a Series or DataFrame column -> Option D

4. Quick Check:

   nunique() = unique count [OK]

Hint: Remember: 'nunique' means 'number of unique' [OK]

Common Mistakes:

• Confusing with sum() function
• Thinking it sorts data
• Assuming it fills missing values

2. Which of the following is the correct syntax to count unique values in the 'Age' column of a DataFrame named df?

easy

A. df['Age'].nunique()

B. df.nunique('Age')

C. df.nunique['Age']

D. df.nunique.Age()

Solution

1. Step 1: Identify correct method call on a column

   To count unique values in a specific column, select the column first with df['Age'], then call nunique().

2. Step 2: Check syntax of options

   df['Age'].nunique() uses correct syntax: df['Age'].nunique(). Other options misuse parentheses or indexing.

3. Final Answer:

   df['Age'].nunique() -> Option A

4. Quick Check:

   Use df['column'].nunique() for unique count [OK]

Hint: Select column first, then call nunique() [OK]

Common Mistakes:

• Passing column name as argument to nunique()
• Using dot notation with parentheses incorrectly
• Trying to index nunique as if a dictionary

3. Given the DataFrame:

import pandas as pd
data = {'City': ['NY', 'LA', 'NY', 'Chicago', 'LA', 'NY'], 'Score': [10, 20, 10, 30, 20, 40]}
df = pd.DataFrame(data)
print(df['City'].nunique())

What is the output?

medium

A. 4

B. 6

C. 3

D. 2

Solution

1. Step 1: Identify unique cities in the 'City' column

   The cities are ['NY', 'LA', 'NY', 'Chicago', 'LA', 'NY']. Unique values are 'NY', 'LA', and 'Chicago'.

2. Step 2: Count unique values

   There are 3 unique cities.

3. Final Answer:

   3 -> Option C

4. Quick Check:

   Unique cities count = 3 [OK]

Hint: Count distinct city names ignoring repeats [OK]

Common Mistakes:

• Counting total rows instead of unique values
• Including duplicates in count
• Confusing with number of columns

4. You run this code:

import pandas as pd
data = {'A': [1, 2, 2, 3], 'B': [4, 4, 5, 6]}
df = pd.DataFrame(data)
print(df['A'].nunique(axis=1))

But it gives an error. What is the problem?

medium

A. The axis parameter is invalid for nunique()

B. nunique() cannot be used on DataFrames

C. The DataFrame has duplicate values causing error

D. Missing parentheses after nunique

Solution

1. Step 1: Check nunique() parameters

   The nunique() function in pandas does not support the axis parameter to count unique values row-wise.

2. Step 2: Understand correct usage

   It works column-wise by default. To count unique values per row, a different approach is needed.

3. Final Answer:

   The axis parameter is invalid for nunique() -> Option A

4. Quick Check:

   nunique() has no axis argument [OK]

Hint: Remember: nunique() counts column-wise only [OK]

Common Mistakes:

• Trying to use axis=1 with nunique()
• Assuming nunique() works row-wise by default
• Confusing with other functions that accept axis

5. You have a DataFrame df with columns 'Name', 'City', and 'Age'. You want to find how many unique cities each person has appeared in, but the data has multiple rows per person. Which code correctly counts unique cities per 'Name'?

hard

A. df['City'].nunique(groupby='Name')

B. df.groupby('Name')['City'].nunique()

C. df.nunique('City', by='Name')

D. df.groupby('City')['Name'].nunique()

Solution

1. Step 1: Understand the goal

   We want to count unique cities for each person (grouped by 'Name').

2. Step 2: Use groupby and nunique correctly

   The correct syntax is to group by 'Name' and then apply nunique() on the 'City' column.

3. Step 3: Check options

   df.groupby('Name')['City'].nunique() matches this syntax. Others misuse parameters or group by wrong columns.

4. Final Answer:

   df.groupby('Name')['City'].nunique() -> Option B

5. Quick Check:

   Group by 'Name', then count unique 'City' [OK]

Hint: Group by person, then count unique cities with nunique() [OK]

Common Mistakes:

• Using nunique() without groupby
• Passing wrong parameters to nunique()
• Grouping by 'City' instead of 'Name'