Polars

DataFrames for the new era — lightning-fast DataFrame library built in Rust.

Polars is a blazingly fast DataFrame library implemented in Rust, leveraging Apache Arrow as its memory model to enable zero-copy data exchange across language boundaries. It features a powerful query optimizer that applies predicate pushdown, projection pushdown, and other optimizations to minimize unnecessary computation. Polars supports both eager and lazy execution modes, the latter allowing users to build complex query plans that are optimized and executed in a single pass. With built-in multi-threaded parallelism, out-of-core processing for datasets larger than memory, and a rich expression API for data manipulation, Polars delivers orders-of-magnitude performance improvements over traditional DataFrame libraries.