daffy

About daffy

A decorator-first DataFrame contracts and validation library for Python. daffy lets you define data contracts as decorators on functions, supporting Pandas, Polars, PyArrow, and Modin DataFrames for lightweight pipeline validation.

Key Features

• 1Lightweight data quality framework for pandas-based pipelines
• 2Expectations defined as Python decorators on transformation functions
• 3Validates DataFrame shape, column types, value ranges, and null counts
• 4Integrates with pytest for unit-testing pipeline data quality logic
• 5Minimal dependencies — pure Python and pandas

How Python Data Engineers Use daffy

Python data engineers use Daffy to add inline data quality assertions to pandas transformation functions — decorating a function with `@expect_column_not_null('id')` raises an exception if the output DataFrame violates the expectation. This brings data quality checks directly into the transformation code rather than requiring a separate validation framework.