Skip to content

featurebyte.Target.compute_target_table

compute_target_table(
observation_table: Union[ObservationTable, DataFrame],
observation_table_name: str,
serving_names_mapping: Optional[Dict[str, str]]=None,
skip_entity_validation_checks: bool=False,
context_id: Optional[ObjectId]=None
) -> ObservationTable

Description

Materialize feature list using an observation table asynchronously. The targets will be materialized into a target table.

Parameters

  • observation_table: Union[ObservationTable, DataFrame]
    Observation set with POINT_IN_TIME and serving names columns. This can be either an ObservationTable of a pandas DataFrame.

  • observation_table_name: str
    Name of the observation table to be created with the target values

  • serving_names_mapping: Optional[Dict[str, str]]
    Optional serving names mapping if the training events table has different serving name.

  • skip_entity_validation_checks: bool
    default: False
    Whether to skip entity validation checks

  • context_id: Optional[ObjectId]
    Context ID for DataFrame observation sets. When provided, the context's forecast point schema will be used to compute the target using FORECAST_POINT instead of POINT_IN_TIME. Ignored when observation_table is an ObservationTable (its own context_id is used).

Returns

  • ObservationTable

Examples

>>> target = catalog.get_target("target")
>>> target.compute_target_table(observation_table, "target_table")