How to Find Unique Values in pandas DataFrame or Series

To find unique values in pandas, use the unique() method on a Series or DataFrame column. This returns an array of distinct values without duplicates.

Syntax

The unique() method is called on a pandas Series or DataFrame column to get unique values.

• Series.unique(): Returns unique values in the Series as a NumPy array.
• DataFrame['column'].unique(): Returns unique values from a specific column.

python

unique_values = df['column_name'].unique()

Example

This example shows how to find unique values in a DataFrame column using unique(). It prints the unique values found.

python

import pandas as pd

data = {'fruits': ['apple', 'banana', 'apple', 'orange', 'banana', 'kiwi']}
df = pd.DataFrame(data)

unique_fruits = df['fruits'].unique()
print(unique_fruits)

Output

['apple' 'banana' 'orange' 'kiwi']

Common Pitfalls

Common mistakes when finding unique values include:

• Calling unique() on a DataFrame instead of a Series or column, which causes an error.
• Expecting unique() to return a list instead of a NumPy array.
• Not handling missing values which appear as nan in the output.

python

import pandas as pd

data = {'fruits': ['apple', 'banana', 'apple', None, 'banana', 'kiwi']}
df = pd.DataFrame(data)

# Wrong: calling unique on DataFrame
# unique_values = df.unique()  # This will raise an AttributeError

# Correct: call unique on a column
unique_values = df['fruits'].unique()
print(unique_values)

Output

['apple' 'banana' None 'kiwi']

Quick Reference

Summary tips for finding unique values in pandas:

• Use Series.unique() to get unique values as a NumPy array.
• Use DataFrame['column'].unique() to find unique values in a specific column.
• To count unique values, use nunique().
• Missing values appear as nan or None in the output.

Key Takeaways

Use the unique() method on a pandas Series or DataFrame column to get unique values.

unique() returns a NumPy array of distinct values including None or nan if present.

Do not call unique() directly on a DataFrame; select a column first.

Use nunique() to count the number of unique values quickly.

Handle missing values carefully as they appear in the unique values output.