DuckDB
In-process analytical database for fast SQL on local data.
DuckDB is an in-process columnar SQL engine in the same architectural lineage as SQLite (no separate server, runs inside the application) but optimised for analytical queries: vectorised execution, columnar storage, and excellent performance on Parquet, CSV, and JSON files directly from disk.
It has become the standard tool for ad-hoc analysis of local data on a Linux workstation, often replacing pandas pipelines for larger datasets.
Install
pip install duckdb # or download the standalone CLI from https://duckdb.org/
Authors
- DuckDB Labs
- CWI (original authors)
