Polars DataFrames

Migrate from pandas to Polars for fast, lazy DataFrames. Rewrites groupby/window/join chains using expression API, streams large datasets that don't fit in RAM, and emits idiomatic LazyFrame pipelines.

Polars is a Rust-backed DataFrame library with lazy evaluation, multi-threaded execution, and a typed expression API. This skill rewrites pandas pipelines as Polars LazyFrames, streams larger-than-memory files, and explains query plans when something is slow.

When to use

Use when pandas is too slow or runs out of memory, when you need predictable types and null handling, or when you want a query plan you can actually optimize before running.

Examples

Pandas → Polars migration

Convert an existing analysis to lazy Polars

Rewrite this 100-line pandas notebook as a single Polars LazyFrame chain and explain where the speedups come from

Streaming a 50GB CSV

Process data that doesn't fit in memory

Use Polars streaming to aggregate hourly metrics from a 50GB CSV with groupby+agg and write Parquet output