Snowflake

Supported via database-sql-snowflake. The Snowflake JDBC driver ships on the platform classpath.

As the default database

DIRIGIBLE_DATASOURCE_DEFAULT_DRIVER=net.snowflake.client.jdbc.SnowflakeDriver
DIRIGIBLE_DATASOURCE_DEFAULT_URL=jdbc:snowflake://<account>.snowflakecomputing.com/?warehouse=COMPUTE_WH&db=DIRIGIBLE&schema=PUBLIC
DIRIGIBLE_DATASOURCE_DEFAULT_USERNAME=DIRIGIBLE
DIRIGIBLE_DATASOURCE_DEFAULT_PASSWORD=<secret>

As a named pool

{
    "name": "AnalyticsDB",
    "driver": "net.snowflake.client.jdbc.SnowflakeDriver",
    "url": "jdbc:snowflake://acct.snowflakecomputing.com/?warehouse=COMPUTE_WH&db=ANALYTICS",
    "username": "${env.SNOWFLAKE_USER}",
    "password": "${env.SNOWFLAKE_PASSWORD}"
}

Snowpark integration

The data-source-snowpark component provides a Snowpark session per Snowflake data source. Useful when you want to push compute into Snowflake instead of pulling rows over JDBC.

Authentication

Username + password is the simplest path. For production prefer key-pair authentication:

DIRIGIBLE_DATASOURCE_DEFAULT_URL=jdbc:snowflake://acct.snowflakecomputing.com/?private_key_file=/etc/secrets/sf_rsa_key.p8&private_key_file_pwd=<pwd>

Notes

• Default schema is PUBLIC. Override via the URL.
• Identifiers default to upper-case unquoted. Use upper-case names in .schema / .table artefacts.
• Snowflake supports IDENTITY and SEQUENCE. The platform's dialect emits both correctly.

See also

• Snowflake authentication
• Working with data