polars-runtime-64

What it is and what it does

This package is a DataFrame query engine written in Rust that prioritizes speed and memory efficiency. It supports both lazy (optimized) and eager execution modes, allowing you to compose complex analytical queries using a chainable expression API. The lazy execution model optimizes your query plan before running it, and the streaming engine can process datasets that exceed available RAM by processing data in chunks.

You install it via pip and import it as `polars`. It has no runtime dependencies beyond the compiled Rust library bundled in the wheel. It's designed for analytical workloads—filtering, grouping, aggregating, and sorting large datasets—and integrates with Apache Arrow for zero-copy data sharing. The package is actively maintained, widely used (top 15000 PyPI packages by downloads), and carries no known security vulnerabilities.

Use it for:

• Process multi-gigabyte CSV or Parquet files on a machine with limited RAM using the streaming engine.
• Build analytical pipelines that filter, group, and aggregate time-series or transactional data with lazy query optimization.
• Compose complex multi-step queries (joins, window functions, aggregations) that benefit from automatic query optimization.
• Interoperate with Arrow-based tools and libraries by leveraging zero-copy columnar data sharing.
• Execute vectorized data transformations where compiled Rust execution delivers measurable performance gains.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

A Rust-based analytical query engine for DataFrames that executes lazy and eager queries with multi-threaded, vectorized performance and can process datasets larger than available RAM through streaming.

Yes. This package is production-stable, actively maintained, permissively licensed, and carries no known vulnerabilities. Install it if you need fast analytical queries on large datasets or want performance-critical data transformation work. The medium install friction (compiled wheels, Python 3.10+ requirement) is standard for Rust-based Python packages and not a barrier for most modern environments.

Install

polars-runtime-64 on PyPI

pip install polars-runtime-64

uv add polars-runtime-64

poetry add polars-runtime-64

Installing polars-runtime-64

Before you install

Medium install friction due to compiled Rust binaries; wheels are pre-built for common platforms (macOS, Linux, Windows on x86_64 and ARM64) but require Python 3.10+. Active maintenance with recent release (13 days old) and strong repository signals (39355 stars, current commit 2026-08-14).

License in practice

MIT license (permissive) allows commercial and private use with minimal restrictions—suitable for most projects without licensing concerns.

Quickstart

pip install polars-runtime-64

import polars as pl

df = (
    pl.scan_parquet("orders.parquet")
    .filter(pl.col("status") == "shipped")
    .group_by("customer_id")
    .agg(pl.col("amount").sum().alias("total"))
    .collect()
)

Requires Python 3.10 or later; pre-built wheels target modern CPUs (AVX2 support assumed unless LTS_CPU=1 flag used during source compilation).

Verify before relying

• Whether GPU acceleration (NVIDIA) is included in the standard PyPI wheel or requires separate installation.
• Memory overhead of lazy query optimization and expression compilation for typical workloads.
• Performance comparison to other dataframe libraries on small-to-medium datasets.

Package facts

Evidence: polars_runtime_64-1.43.2-cp310-abi3-macosx_10_12_x86_64.whl; polars_runtime_64-1.43.2-cp310-abi3-macosx_11_0_arm64.whl; polars_runtime_64-1.43.2-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl; polars_runtime_64-1.43.2-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl; polars_runtime_64-1.43.2-cp310-abi3-musllinux_1_2_aarch64.whl; polars_runtime_64-1.43.2-cp310-abi3-musllinux_1_2_x86_64.whl; polars_runtime_64-1.43.2-cp310-abi3-win_amd64.whl; polars_runtime_64-1.43.2-cp310-abi3-win_arm64.whl

Keywords: dataframe, arrow, out-of-core

Tags

More Scientific/Engineering packages