Pointblank is a Python data-validation library that checks Polars or pandas DataFrames against declared rules, accumulating failures until they cross configurable thresholds. The tutorial covers the full workflow: declaring validation plans with col_vals_*() methods, setting warning/error/critical thresholds with Thresholds(), triggering automated Actions(), splitting clean from failing rows using get_sundered_data() (data sundering), and storing plans as YAML to run via the pb CLI with --fail-on critical for CI integration. It also compares Pointblank with Pandera and Great Expectations, and outlines known limitations like runtime-only checks and partial sundering coverage.
Table of contents
Get Started With Pointblank in PythonSet Thresholds and Trigger Automated ActionsRoute Clean and Dirty Rows With Data SunderingReuse Validation Plans With YAML and the CLICompare Pointblank With Other Python Validation ToolsLimitations to Know About PointblankConclusionFrequently Asked Questions384 Impressions