Polars & Columnar Dataframes

For medium-to-large in-memory data work, the performance and memory gains are compelling enough to make migration worthwhile for any new data engineering code.