Setup & Installation
npx skills add https://github.com/clickhouse/agent-skills --skill chdb-datastore
https://github.com/clickhouse/agent-skills/tree/main/skills/chdb-datastore
View on GitHubWhat This Skill Does
chDB DataStore exposes the ClickHouse engine behind a pandas-compatible API so existing pandas code runs 10-100x faster on large datasets, with cross-source joins across MySQL, PostgreSQL, S3, MongoDB, ClickHouse, Iceberg, Delta Lake, and file formats like Parquet, CSV, JSON, Arrow, and ORC.
Instead of choosing between pandas ergonomics and ClickHouse performance, chdb.datastore lets you keep pandas code while getting columnar execution and native connectors for a wide range of sources.
When to use it
- Speeding up slow pandas pipelines on large datasets without rewriting them
- Joining a Parquet file on S3 with a Postgres table using pandas-style syntax
- Loading remote MySQL or MongoDB data as a DataFrame for analysis
- Prototyping analytics against cloud storage without standing up a server
- Running notebooks that mix file-based and database-backed sources