temporian.IndexData dataclass

Features and timestamps data for a single index group.

Note: The schema constructor argument is only used for checking. If schema=None, no checking is done. Checking can be done manually with index_data.check_schema(...).

Attributes:

Name	Type	Description
features	List[ndarray]	List of one-dimensional NumPy arrays representing the features.
timestamps	ndarray	One-dimensional NumPy array representing the timestamps.

Usage example

>>> features = [np.array([1, 2, 3]), np.array([4, 5, 6])]
>>> timestamps = np.array([0, 1, 2])
>>> index_data = tp.IndexData(features, timestamps)
>>> len(index_data)
3

__init__

__init__(
    features: List[np.ndarray],
    timestamps: np.ndarray,
    schema: Optional[Schema] = None,
) -> None

Initializes the IndexData object by checking and setting the features and timestamps.

Raises:

Type	Description
ValueError	If features are not one-dimensional arrays.
ValueError	If the number of elements in features and timestamps do not match.

__len__

__len__() -> int

Number of events / timestamps.