|View source on GitHub|
View of statistics for a dataset (slice).
tfdv.DatasetView( stats_proto: statistics_pb2.DatasetFeatureStatistics )
get_cross_feature( x_path: Union[str,
tfdv.FeaturePath, Iterable[str]], y_path: Union[str,
tfdv.FeaturePath, Iterable[str]] ) -> Optional['CrossFeatureView']
Retrieve a cross-feature if it exists, or None.
get_derived_feature( deriver_name: str, source_paths: Sequence[
tfdv.FeaturePath] ) -> Optional['FeatureView']
Retrieve a derived feature based on a deriver name and its inputs.
||The name of a deriver. Matches validation_derived_source deriver_name.|
||Source paths for derived features. Matches validation_derived_source.source_path.|
|FeatureView of derived feature.|
|ValueError if multiple derived features match.|
get_feature( feature_id: Union[str,
tfdv.FeaturePath, Iterable[str]] ) -> Optional['FeatureView']
Retrieve a feature if it exists.
Features specified within the underlying proto by name (instead of path) are normalized to a length 1 path, and can be referred to as such.
||A types.FeaturePath, Iterable[str] consisting of path steps, or a str, which is converted to a length one path.|
|A FeatureView, or None if feature_id is not present.|
list_cross_features() -> Iterable[Tuple[types.FeaturePath, types.FeaturePath]]
Lists cross-feature identifiers.
list_features() -> Iterable[
Lists feature identifiers.
proto() -> statistics_pb2.DatasetFeatureStatistics
Retrieve the underlying proto.