Skip to content

OutputsConfig

relationalai.config.config_deployment_fields

Defaults and per-output overrides for the tables and relations your model produces.

Top-level keys set deployment-wide defaults; overrides is an ordered list of per-output adjustments. Output names follow Snowflake rules: unquoted names match case-insensitively (cancel_chain matches CANCEL_CHAIN); wrap a name in double quotes (e.g. '"My.Table"') to match a quoted identifier exactly.

Examples

outputs:
reasoner: sql
type: table
schedule: standard
overrides:
- objects: [TRANSACTION]
reasoner: logic
- objects: [recursive_closure]
schedule: fast_path
- objects: [SOURCE_DERIVED_TRANSACTIONS, DERIVED_TRANSACTIONS]
type: view

Attributes

.reasoner

OutputsConfig.reasoner: (str, optional)

Default reasoner — "sql" or "logic". When omitted (default), the router auto-decides per output. When set, every output not covered by an override is forced onto this reasoner, except outputs another reasoner owns outright — the staging a configured Problem writes its results into, which only the solver can compute. The tables you read solutions from are ordinary SQL over that staging and are still forced.

Advanced option. Only force outputs that the target reasoner can actually compute. The reasoner name itself is checked when the config loads; whether the reasoner can compute a given output is not, and an incompatible forcing may not fail until refresh.

.type

OutputsConfig.type: str

Default materialization type for output tables (Snowflake only): "view", "table", or "dynamic_table" (default). Ignored on DuckDB.

.schedule

OutputsConfig.schedule: (str, optional)

Default schedule name (under deployment.schedules) that outputs refresh on. Outputs without a schedule override use this one; if omitted, those outputs are not scheduled.

.overrides

OutputsConfig.overrides: list[OutputOverride]

Ordered per-output adjustments. Each entry names a group of outputs and sets one or more of reasoner, type, schedule for them.

Inheritance Hierarchy

OutputsConfigpydantic.BaseModel