Can I use Narwhals for lazy execution?

Yes, Narwhals supports both eager and lazy execution modes, including backends like Dask, DuckDB, PySpark, and SQLFrame.

What is the main benefit of using Narwhals?

It allows you to write code once using a unified API that works across pandas, Polars, PyArrow, and more without changing your logic.

Does Narwhals add significant overhead?

No, Narwhals is designed to be lightweight with zero dependencies and negligible performance overhead during dataframe operations.

Is Narwhals type-safe?

Yes, Narwhals provides full static typing support with specialized type hints like DataFrameT and IntoDataFrame for robust, error-free development.

How does the @narwhalify decorator work?

The @narwhalify decorator automatically converts input dataframes to Narwhals objects and converts results back to the original native type upon function return.

Narwhals Dataframe Compatibility

Name: Narwhals Dataframe Compatibility
Author: anam-org

byanam-org

•

Data Science & ML

Writes dataframe-agnostic Python code that runs seamlessly across pandas, Polars, PyArrow, and other major backends.

Narwhals provides a lightweight, zero-dependency compatibility layer for Python dataframe libraries, allowing developers to write backend-independent code using a unified Polars-like API. It is particularly useful for library maintainers and data engineers who need to support multiple data processing engines without maintaining separate codebases, ensuring performance and type safety across both eager and lazy execution modes. By using Narwhals, you can resolve subtle differences between libraries while maintaining negligible performance overhead.

Key Features

01Zero-dependency, lightweight compatibility layer with negligible overhead

02Multi-backend support including pandas, Polars, PyArrow, and cuDF

03Full static typing support for robust, type-safe data pipelines

04Automatic dataframe conversion using the @narwhalify decorator

05Comprehensive expression API for complex data transformations

0612 GitHub stars

Use Cases

01Standardizing data processing logic across teams using different libraries

02Building portable Python libraries that work with any dataframe backend

03Migrating data pipelines from pandas to high-performance engines like Polars

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add anam-org/metaxy narwhals

For use in Claude.ai and ChatGPT

Download Skill