merlin.dataloader.tf_utils.get_dataset_schema_from_feature_columns

merlin.dataloader.tf_utils.get_dataset_schema_from_feature_columns(feature_columns)

Maps from a list of TensorFlow feature_column to two lists. The first list provides the categorical feature names and the second list provides the continuous feature names for a dataset. This function is useful for constructing NVTabular Workflows from feature columns.

Returns

One sorted list of categorical feature names and one sorted list of continuous feature names. The lists can have zero length.

Return type

list[str], list[str]