polars - Fast DataFrame Processing with Polars
In-memory DataFrame library built on Apache Arrow with lazy evaluation and parallel execution for efficient data processing of 1-100GB datasets.
Tags
Updated: 2026-06-30Capabilities
Typical Inputs
Typical Outputs
What this skill does
- read CSV/Parquet/JSON data files
- write data to multiple formats
- filter DataFrame rows
- select and transform columns
- add computed columns
- group and aggregate data
- join multiple DataFrames
- pivot and unpivot data
- execute lazy evaluation queries
- apply window functions
- concatenate DataFrames vertically or horizontally
Inputs
- Data files (CSV, Parquet, JSON, Excel)
- Database connections
- Cloud storage credentials (S3, Azure, GCS)
- BigQuery dataset
Outputs
- Processed DataFrames
- Output data files
- Query results
Requirements
- Python or Rust runtime
- polars package installed
- Sufficient RAM for dataset (1-100GB)