Using pandas fillna() to Handle Missing Data: 10 Practical Examples

Dear developers, have you ever struggled with NaN or None values when processing data? In data science and analysis, these missing values can easily lead to biased results.

Fortunately, Python's pandas library offers the powerful fillna() method. Below are ten practical examples that illustrate how to master missing‑data handling.

Example 1: Basic Filling

import pandas as pd
import numpy as np
df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
print(df.fillna(0))

Example 2: Column‑Specific Filling

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
values = {'A': 0, 'B': 5}
print(df.fillna(value=values))

Example 3: Forward Fill

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
print(df.fillna(method='ffill'))

Example 4: Backward Fill

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
print(df.fillna(method='bfill'))

Example 5: Limit Number of Fills

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4, np.nan],
    'B': [5, np.nan, np.nan, 8, 9],
    'C': [9, 10, 11, 12, 13]
})
print(df.fillna(method='ffill', limit=1))

Example 6: Fill Using Another DataFrame

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
other_df = pd.DataFrame({
    'A': [0],
    'B': [5],
    'C': [9]
})
print(df.fillna(other=other_df))

Example 7: Fill with Computed Values (Mean)

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
mean_values = df.mean()
print(df.fillna(mean_values))

Example 8: Conditional Fill per Column

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
print(df.fillna({col: df[col].mean() for col in df.columns}))

Example 9: Dictionary Fill for Specific Columns

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
values_dict = {'A': 0, 'B': df['B'].mean()}
print(df.fillna(value=values_dict))

Example 10: Fill Using a Series

df = pd.DataFrame({
    'A': [1, 2, np.nan, 4],
    'B': [5, np.nan, np.nan, 8],
    'C': [9, 10, 11, 12]
})
fill_series = pd.Series([0, 5, 9], index=['A', 'B', 'C'])
print(df.fillna(fill_series))

Through these examples, we learn both the basic usage of fillna() and how to choose the most appropriate filling strategy for different scenarios. Mastering this method is crucial for real‑world datasets and building reliable machine‑learning models.

Properly handling missing values improves data quality, prevents bias, and ensures accurate analysis results.

We’ll continue exploring more Python and data‑analysis techniques in future articles. Remember to select filling strategies carefully, as inappropriate choices can introduce bias, and consider performance optimizations for large datasets.

If you encounter any issues or want to share your own data‑cleaning tips, feel free to comment and join the discussion.