How to Use count() in pandas: Syntax and Examples

In pandas, use the count() method on a DataFrame or Series to count the number of non-missing (non-NaN) values. It returns counts per column by default for DataFrames, or a single count for Series.

Syntax

The count() method syntax is:

• DataFrame.count(axis=0, level=None, numeric_only=False)
• Series.count()

Parameters:

• axis=0: Count along columns (0) or rows (1).
• level=None: For MultiIndex, count along a level.
• numeric_only=False: Count only numeric data if True.

It returns the count of non-NaN values.

python

df.count(axis=0, level=None, numeric_only=False)

Example

This example shows how to count non-missing values in each column of a DataFrame.

python

import pandas as pd
import numpy as np

data = {'A': [1, 2, np.nan, 4], 'B': [np.nan, 2, 3, 4], 'C': [1, np.nan, np.nan, 4]}
df = pd.DataFrame(data)

counts = df.count()
print(counts)

Output

A    3
B    3
C    2
dtype: int64

Common Pitfalls

Common mistakes include:

• Expecting count() to count all rows including NaNs (it only counts non-NaN values).
• Using count() on a Series without realizing it returns a single integer.
• Confusing count() with len() which counts all rows regardless of missing data.

python

import pandas as pd
import numpy as np

data = {'A': [1, 2, np.nan, 4]}
df = pd.DataFrame(data)

# Wrong: expecting count to include NaNs
wrong_count = len(df['A'])  # counts all rows including NaN
right_count = df['A'].count()  # counts only non-NaN

print(f"len(): {wrong_count}")
print(f"count(): {right_count}")

Output

len(): 4
count(): 3

Quick Reference

Summary tips for using count() in pandas:

• Use count() to count non-missing values only.
• Default axis=0 counts per column; use axis=1 to count per row.
• For Series, count() returns a single integer.
• Use len() to count total rows including missing values.

Key Takeaways

Use pandas count() to count non-NaN values in DataFrames or Series.

By default, count() counts non-missing values per column in DataFrames.

For Series, count() returns a single integer count of non-NaN values.

Do not confuse count() with len(); the latter counts all rows including NaNs.

Use axis=1 in count() to count non-NaN values per row.