chDB

chDB is a fast in-process SQL OLAP Engine powered by ClickHouse. You can use it when you want to get the power of ClickHouse in a programming language without needing to connect to a ClickHouse server.

Key features

• In-process SQL OLAP Engine - Powered by ClickHouse, no need to install ClickHouse server
• Multiple data formats - Input & Output support for Parquet, CSV, JSON, Arrow, ORC and 70+ more formats
• Minimized data copy - From C++ to Python with python memoryview
• Rich Python Ecosystem Integration - Native support for Pandas, Arrow, DB API 2.0, seamlessly fits into existing data science workflows
• Zero dependencies - No need for external database installations

What languages are supported by chDB?

chDB has the following language bindings:

• Python
• Go
• Rust
• NodeJS
• Bun
• C and C++

How do I get started?

• If you're using Go, Rust, NodeJS, Bun or C and C++, take a look at the corresponding language pages.
• If you're using Python, see the getting started developer guide. There are also guides showing how to do common tasks like:
  • JupySQL
  • Querying Pandas
  • Querying Apache Arrow
  • Querying data in S3
  • Querying Parquet files
  • Querying remote ClickHouse
  • Using clickhouse-local database

An introductory video

You can listen to a brief project introduction to chDB, courtesy of Alexey Milovidov, the original creator of ClickHouse:

Performance benchmarks

chDB delivers exceptional performance across different scenarios:

• ClickBench of embedded engines - Comprehensive performance comparison
• DataFrame Processing Performance - Comparative analysis with other DataFrame libraries
• DataFrame Benchmark

About chDB

• Read the full story about the birth of the chDB project on blog
• Read about chDB and its use cases on the Blog
• Discover chDB in your browser using codapi examples
• More examples see (https://github.com/chdb-io/chdb/tree/main/examples)

License

chDB is available under the Apache License, Version 2.0. See LICENSE for more information.