Pandera

Pandera is an open-source Python library for validating dataframes. Its README describes “a flexible and expressive API for performing data validation on dataframe-like objects” and says the goal is to make data processing pipelines “more readable and robust with statistically typed dataframes”. It is a Union.ai open-source project.

Repository snapshot (GitHub API, 2026-11-18)

ItemDetail
Repounionai-oss/pandera, Python, MIT licence
Created2018-11-01
Activitylast push 2026-11-18; latest release v0.36.0 (2026-11-10)
Size4,530 stars, 524 forks

What it does

  • You declare a schema (column types plus checks) and validate a dataframe against it, so bad data fails inside the pipeline code.
  • Supports pandas, Polars and PySpark, among other dataframe libraries (see its supported-libraries docs page).
  • Installed with an extra per library, for example pip install 'pandera[pandas]'.

Where it fits against Soda Core

Pandera validates data that is already in memory in your Python code. Soda Core scans tables in a warehouse with YAML checks, and great-expectations sits between the two. Pandera has no hosted platform, alerting or dashboard in the README read; for those, pair it with an observability tool.

Sources