What problem does it solve? Data pipelines built on Polars often lack explicit runtime contracts, so dtype drift, unexpected nulls, extra columns, and silent coercion bugs reach production undetected. This Skill guides the creation of executable Pandera schemas for the Polars backend so validation rules, coercion policy, and failure handling are explicit and testable. ## Core Features & Use Cases - Schema authoring guidance: Choose between DataFrameSchema for programmatic composition and DataFrameModel for declarative named contracts, with explicit strict, coerce, nullable, and unique policies. - Validation depth control: Distinguish schema-only LazyFrame validation from full data-level checks on eager DataFrames, and aggregate failures with lazy=True. - Native custom checks: Write vectorized Polars checks via PolarsData returning Boolean LazyFrames instead of slow element-wise Python callbacks. - Use Case: You ingest CSV order data into a Polars pipeline. Define a strict DataFrameSchema requiring order_id as unique non-null Int64, country in an allowed set, and amount non-negative, then validate at ingress with lazy=True to collect all failures at once. ## Quick Start Ask the AI to write a strict Pandera Polars DataFrameSchema that validates your orders DataFrame with unique integer IDs, an allowed country set, and aggregated failure reporting.
