How to Use nunique in pandas to Count Unique Values

Use nunique() in pandas to count the number of unique values in a Series or DataFrame column. It returns an integer representing how many distinct values exist, optionally ignoring missing values with dropna.

Syntax

The basic syntax of nunique() is:

• Series.nunique(dropna=True): Counts unique values in a Series, ignoring NaN by default.
• DataFrame.nunique(axis=0, dropna=True): Counts unique values per column (axis=0) or per row (axis=1), ignoring NaN by default.

Parameters:

• dropna: Boolean, default True. If True, NaN values are not counted as unique.
• axis: For DataFrame, 0 counts unique values per column, 1 per row.

python

Series.nunique(dropna=True)
DataFrame.nunique(axis=0, dropna=True)

Example

This example shows how to use nunique() on a pandas DataFrame to count unique values in each column and in a Series.

python

import pandas as pd

data = {'A': [1, 2, 2, 3, 4, 4, None],
        'B': ['x', 'y', 'x', 'z', 'y', 'y', 'z'],
        'C': [True, True, False, False, True, None, None]}
df = pd.DataFrame(data)

# Count unique values per column
unique_counts = df.nunique()

# Count unique values in column 'B'
unique_in_B = df['B'].nunique()

unique_counts, unique_in_B

Output

(A    4
B    3
C    2
dtype: int64, 3)

Common Pitfalls

Common mistakes when using nunique() include:

• Not realizing dropna=True by default, so missing values are not counted unless you set dropna=False.
• Using nunique() on the whole DataFrame without specifying axis, which counts unique values per column, not overall.
• Expecting nunique() to return unique values themselves instead of counts.

python

import pandas as pd

s = pd.Series([1, 2, 2, None, None])

# Wrong: expecting NaN to be counted
count_with_nan = s.nunique(dropna=True)  # returns 2

# Right: count NaN as unique
count_with_nan_correct = s.nunique(dropna=False)  # returns 3

count_with_nan, count_with_nan_correct

Output

(2, 3)

Quick Reference

Parameter	Description	Default
dropna	Whether to exclude NaN values from count	True
axis	Axis to count unique values on (0=columns, 1=rows)	0 (DataFrame only)

Key Takeaways

Use nunique() to count distinct values in pandas Series or DataFrame columns.

By default, nunique() ignores missing (NaN) values unless dropna=False is set.

For DataFrames, nunique() counts unique values per column (axis=0) or per row (axis=1).

Remember nunique() returns counts, not the unique values themselves.

Use dropna=False to include NaN as a unique value in the count.