Configuration

On this page

SQL config files loaded at startup, each owning a piece of the runtime.

  • catalog.sql — DuckLake catalog attachment (required)
  • state.sql — state catalog attachment for push queue, fetch staging, OAuth tokens (required)
  • extensions.sql — load DuckDB extensions before execution
  • secrets.sql — DuckDB secrets for cloud storage, databases, APIs
  • settings.sql — memory, parallelism, spill behavior
  • sources.sql — attach external databases (Postgres, MySQL, …)
  • macros/ — reusable SQL macros for validation, helpers, masking
  • variables/ — constant/global/per-model variables
settings.sql

Configure DuckDB execution settings in OndatraSQL. Control memory limits, parallelism, and disk spill behavior for your SQL data pipeline runtime.

secrets.sql

OndatraSQL uses DuckDB's secrets manager for external credentials. Configure access to cloud storage, databases, and APIs for your data pipeline.

extensions.sql

Load optional DuckDB extensions in OndatraSQL. Add cloud storage access, external database connectors, geospatial functions, and more to your pipeline.

catalog.sql

Configure the DuckLake catalog in OndatraSQL. The only required config file. Tells the runtime where to store metadata and Parquet data files.

state.sql

Configure the operational state backend in OndatraSQL. Holds the push queue, fetch staging buffer, and encrypted OAuth refresh tokens.

macros/

OndatraSQL loads reusable SQL macros from the config/macros/ directory. Define DuckDB functions for validation, helpers, and masking in your pipeline.

variables/

OndatraSQL loads runtime variables from the config/variables/ directory. Define pipeline configuration as SQL across constant, global, and per-model scopes.

sources.sql

Query external databases directly in OndatraSQL via DuckDB's attach mechanism. Connect PostgreSQL, MySQL, and other sources for use in SQL models.