Polars Optimization
To maximize Polars performance:
- Always use Lazy API (
scan_parquetinstead ofread_parquet) when reading large datasets. - Avoid
apply: Custom Python functions break the Rust optimizations. Stick to native Polars expressions. - Use Categorical types for string columns with low cardinality (like
airline_codeorweather_condition) to save memory and speed up joins.