cBioPortal Public Database Dumps

Weekly exports of the cBioPortal public database, available for download and self-hosting. Please note that these exports originate from ClickHouse Cloud and are adapted for self-hosted ClickHouse. Differences between ClickHouse versions or environments may cause restore or compatibility issues. For help, contact cbioportal@googlegroups.com.

How to download and restore (ClickHouse)

Pick a dump below and download the Parquet files you need. Each table is a separate file, so you can take only the tables you care about.

Create a target database, then run schema.sql against it to create all tables and views:

clickhouse-client --query "CREATE DATABASE <database>"
clickhouse-client --database <database> --queries-file schema.sql

Load each Parquet file into the corresponding table:

clickhouse-client --database <database> --query "
  INSERT INTO <table>
  SELECT * FROM file('<table>.parquet', 'Parquet')
"

Parameterized views are included in schema.sql and will work automatically once the base tables are populated.
Use manifest.json to verify the exported Parquet row counts after restore.

For version-specific setup and further guidance about importing data, see the official ClickHouse documentation.

ClickHouse (Parquet)

One Parquet file per table. Expand a dump to list its tables. Dumps run weekly, and only the most recent exports are retained — older ones are removed automatically, so download anything you need to retain.

Hugging Face

We also publish the ClickHouse dumps in the cBioPortal Public Database dataset on Hugging Face . Each table is a separate dataset configuration that can be browsed, with additional Data Studio search and filtering features available when supported by Hugging Face.

Legacy MySQL

Archived mysqldump exports, kept for backwards compatibility. No longer updated.

Date	Schema Version	Format	Size	Download