pola-rs/polars

Health v2 · maintenance only

Adoption not part of the score

39505 stars · 3052 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Polars is an extremely fast analytical query engine for DataFrames written in Rust, with multi-threaded vectorized execution, lazy query optimization, and a streaming engine for larger-than-RAM datasets. It offers bindings for Python, Rust, Node.js, R, and SQL, uses the Apache Arrow columnar format for zero-copy interoperability, and optionally supports NVIDIA GPU acceleration.

Use cases

• analyze large csv and parquet files faster than pandas
• process datasets larger than RAM with streaming
• replace pandas with a faster dataframe library
• run optimized lazy queries on tabular data
• read and write data from cloud storage and databases
• accelerate dataframe workloads on NVIDIA GPUs
• build etl pipelines in python or rust

When to choose

• you need high-performance dataframe processing that uses all CPU cores
• your data doesn't fit in memory and you need out-of-core streaming execution
• you want query optimization with a lazy/expressive API
• you need zero-copy interop with Apache Arrow ecosystems
• you work across Python, Rust, Node.js, or R with a consistent API

When to avoid

• you depend on the full pandas/NumPy ecosystem of third-party libraries
• you need row-level iterative mutation rather than columnar batch operations
• your team relies on pandas-specific idioms and APIs without migration effort
• you need a distributed cluster engine rather than single-machine parallelism

Facets

library · maturity stable

data-science etl database parser concurrency data-science analytics big-data developer-tools python rust cross-platform cli dataframe query-engine apache-arrow parquet streaming-engine lazy-evaluation out-of-core pandas-alternative sql simd data-engineering nodejs gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="pola-rs/polars")