temporian.input_node #
input_node(
features: List[Tuple[str, DType]],
indexes: Optional[List[Tuple[str, IndexDType]]] = None,
is_unix_timestamp: bool = False,
same_sampling_as: Optional[EventSetNode] = None,
name: Optional[str] = None,
) -> EventSetNode
Creates an input EventSetNode
.
An input EventSetNode can be used to feed data into a graph.
Usage example:
```python
>>> # Without index
>>> a = tp.input_node(features=[("f1", tp.float64), ("f2", tp.str_)])
>>> # With an index
>>> a = tp.input_node(
... features=[("f1", tp.float64), ("f2", tp.str_)],
... indexes=["f2"],
... )
>>> # Two nodes with the same sampling
>>> a = tp.input_node(features=[("f1", tp.float64)])
>>> b = tp.input_node(features=[("f2", tp.float64)], same_sampling_as=a)
```
Parameters:
Name | Type | Description | Default |
---|---|---|---|
features |
List[Tuple[str, DType]]
|
List of names and dtypes of the features. |
required |
indexes |
Optional[List[Tuple[str, IndexDType]]]
|
List of names and dtypes of the index. If empty, the data is assumed not indexed. |
None
|
is_unix_timestamp |
bool
|
If true, the timestamps are interpreted as unix timestamps in seconds. |
False
|
same_sampling_as |
Optional[EventSetNode]
|
If set, the created EventSetNode is guaranteed to have the
same sampling as same_sampling_as |
None
|
name |
Optional[str]
|
Name for the EventSetNode. |
None
|
Returns:
Type | Description |
---|---|
EventSetNode
|
EventSetNode with the given specifications. |