Pandera

Pandera is an open source framework for validating dataframe-like objects. It provides a flexible and expressive API to define schemas that describe the expected columns, data types and checks on values, so that data processing pipelines are more readable and robust with statistically typed dataframes.

Schemas can be written as objects or as class-based models, used to validate data at runtime or in tests, and to generate synthetic data for property-based testing. Pandera supports several dataframe libraries, including pandas, Polars and PySpark.

Pandera is a Union.ai open source project under the MIT licence, installed with pip or conda.