make-pythonic

Native Python types replace lit() wrappers in DataFusion Python APIs through signature coercion. Functions in python/datafusion/functions.py accept native literals like integers or strings directly, updating aliases and doctests to reflect Pythonic calling conventions while maintaining backward compatibility.

Passing raw Python literals directly to DataFusion expressions requires auditing function signatures against upstream Rust constraints. The API applies type coercion rules to accept native types for literal-only arguments, enabling calls like split_part(col("a"), ",", 2) without explicit wrappers.

Refactoring DataFusion Python bindings for native types preserves backward compatibility. Existing lit() and col() wrappers continue functioning alongside new coercion rules, ensuring prior code remains operational while doctests refresh to demonstrate updated Pythonic calling conventions.

Updating doctests for Pythonic literal coercion involves validating argument categories and consulting upstream Rust bindings for type constraints. Refresh test cases to demonstrate native literal usage, ensuring new coercion rules align with expected DataFusion API behavior.

DataFusion historically required lit() for simple literal arguments due to Rust binding type constraints. Auditing function signatures identifies literal-only patterns, enabling automatic coercion of native Python types and reducing boilerplate while respecting upstream type limitations.

Coercing native Python types in DataFusion function signatures is limited by upstream Rust binding constraints. The API validates argument categories to determine eligible literal-only patterns, restricting automatic coercion where type safety or complex argument handling prevents direct native type acceptance.