Skip to content

Pairity

Configurable data regression testing, starting with Databricks.

Pairity compares pairs of tables to show what changed between two versions of your data. It runs the comparisons in parallel on Spark and reports missing, extra and changed rows, with counts for each column.

Pairity on GitHub

Highlights

  • Primary keys come from your mapping or from Unity Catalog. A missing, null or duplicate key never produces a passing comparison.
  • Compares schemas and values, including nested structs, maps and arrays, with exact numeric tolerances.
  • Pins Delta inputs to recorded table versions, and never writes to the tables it compares.
  • Uses one JSON/YAML configuration model that you can validate, convert and bulk-patch.
  • Includes a Databricks workbench app for editing configurations, submitting Jobs and downloading evidence.
  • Renders CLI reports with rs-rich.