DuckDB
In-process analytical database for fast local queries
DuckDB is profiled here as a Backend tool for engineering teams. Read about features, pricing, and how it compares to related options in the tools directory.
Description
DuckDB is an open-source analytical database created by Mark Raasveldt and Hannes Mühleisen at CWI in the Netherlands. It runs in-process inside an application with no server to manage, and its columnar vectorized engine executes analytical SQL over millions of rows on a laptop in milliseconds. DuckDB reads Parquet, CSV, and JSON directly and queries data in cloud storage, which has made it a staple of local analytics and data pipelines. An active extension ecosystem adds capabilities such as full-text search and geospatial queries on top of the core engine. Because it embeds in the host process, it carries no network overhead and needs no separate service to maintain.
Key Capabilities:
In-process execution with zero server setup
Columnar vectorized engine for fast analytical SQL
Direct querying of Parquet, CSV, and JSON files
Queries over data in S3 and other object storage
Larger-than-memory processing that spills to disk
MIT license with extensions and bindings for many languages
Alternative tools
- WorkOS
Enterprise-ready authentication and SSO for SaaS apps
- pgvector
Vector similarity search as a Postgres extension
- Convex
Reactive backend with a database and serverless functions
- Appwrite
Open-source backend platform for web and mobile apps
- ClickHouse
Open-source columnar database for real-time analytics
- Auth0
Managed identity platform for applications and APIs