---
id: pyjanitor
version: "0.32.23"
license: MIT
license_treatment: permissive
maintenance: active
---

# pyjanitor — Tools for cleaning pandas DataFrames

License: permissive · Maintenance: active · Downloads: 383.2K/mo

## What it is and what it does

pyjanitor is a pandas extension that adds a collection of data-cleaning methods designed to work as part of method chains. Instead of writing imperative pandas code with intermediate variable assignments, you can express a sequence of cleaning steps as a readable chain of method calls—each with an explicit verb name like clean_names(), remove_empty(), or rename_column(). The package is inspired by the R janitor package and the dplyr paradigm, bringing that fluent, declarative style to Python.

Under the hood, pyjanitor registers its functions as pandas DataFrame methods via pandas-flavor and depends on scipy, natsort, multipledispatch, and janitor-rs for its operations. It handles common preprocessing tasks: standardizing column names, removing null or empty rows and columns, identifying duplicates, encoding categorical data, splitting features and targets, coalescing columns, date conversions, and expanding delimited values into dummy variables. The package is actively maintained, supports Python 3.8 and later, and carries no known security vulnerabilities.

Use it for:

- Clean messy column names (spaces, mixed case, special characters) in a single method call before analysis.
- Chain multiple DataFrame transformations (drop nulls, rename columns, add computed fields) in a single readable expression.
- Preprocess raw data exports (Excel, CSV) by removing empty rows and columns and standardizing formats in one pipeline.
- Prepare machine learning datasets by splitting features and targets and encoding categorical variables declaratively.
- Expand delimited or categorical columns into dummy-encoded variables for statistical modeling.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Extends pandas DataFrames with method-chainable data-cleaning functions, enabling readable, verb-based operations for common preprocessing tasks like column renaming, null removal, and data transformation.

Yes. pyjanitor solves a real readability and usability gap in pandas workflows. Low install friction, active maintenance, permissive MIT license, no known vulnerabilities, and strong adoption make it a safe, practical choice for teams that value readable data-cleaning code. Install it if you work with pandas and want method chains to replace imperative preprocessing logic.

## Install

pip install pyjanitor
uv add pyjanitor
poetry add pyjanitor

## Installing pyjanitor

Before you install: Low friction install with five runtime dependencies. Actively maintained—last commit 2026-08-11, 1500 repository stars. Supports Python 3.8 and later.

License in practice: MIT license (permissive) places no restrictions on use, modification, or distribution in proprietary or open-source projects.

Quickstart:

import pandas as pd
import janitor

df = pd.DataFrame({'Name': ['Alice', 'Bob'], 'Age': [None]})
df = df.clean_names().remove_empty()

Requires Python 3.8 or later; janitor-rs (a Rust-based dependency) must compile on your platform.

Verify before relying:

- Performance characteristics when chaining many operations on large DataFrames.
- Compatibility of janitor-rs compilation on all target platforms (Windows, macOS, Linux variants).
- Extent of experimental submodules (finance, biology, chemistry, engineering, pyspark) and their stability.

## Package facts

- License: MIT (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 383.2K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

pandas data cleaning methods, dataframe preprocessing chaining, column name cleaning, remove empty rows columns, data wrangling pandas, janitor-style data prep, pandas method chaining verbs, data-cleaning, pandas-extension, method-chaining

[View on SkillFed](https://skillfed.io/packages/pyjanitor) · [View on PyPI](https://pypi.org/project/pyjanitor/)