Polars
Rust-based DataFrame library built for speed, increasingly displacing pandas on large datasets.
https://pola.rsUpdate history
Polars 2.0 adds an out-of-core streaming sort and numerous engine-level speedups across joins, group-by, and Parquet/IPC/Iceberg scans, along with changes to how decimal arithmetic and runtime filters are computed. SQL window functions now operate on aggregated rows with QUALIFY evaluated before projection, SQL numeric literals are typed as Decimal with truncated %/DIV, Parquet ENUM columns are read as pl.String instead of their prior type, and cut/qcut are deprecated, so SQL queries and code relying on the old numeric, window, or ENUM behavior need to be updated.
Matters for: Teams running large-scale SQL analytics or ETL pipelines over Parquet, IPC, or Iceberg data lakes with Polars.
Get the badge
Show that Polars is tracked on StackFollow in your project's README.
[](https://stackfollow.xyz/tools/polars)