DQBench v1.0.0
The standard benchmark for data quality and validation tools.
Three Tiers
- Tier 1 (Basics): 5K rows, 20 columns — obvious errors
- Tier 2 (Realistic): 50K rows, 30 columns — subtle issues + false positive traps
- Tier 3 (Adversarial): 100K rows, 50 columns — encoding traps, semantic errors, cross-column logic
Features
- Tool-agnostic adapter interface (20 lines to integrate)
- Recall + Precision + F1 scoring per tier
- DQBench Score (0-100) composite
- Built-in GoldenCheck adapter
- Rich CLI scorecard output
- Deterministic, reproducible datasets
Install
pip install dqbench