Use the apply() method to register objects to the repository.
- If the object is an Entity or FeatureGroup and the same entity or feature group is already registered in the repository, it is not updated.
- If the entity or feature group is associated with any feature process, an error is raised while modifying these objects.
Optional Parameter
- object
- Specifies the object to update the repository.
Example setup
>>> from teradataml import FeatureStore, DataFrame, load_example_data
Create a DataFrame on sales_data.
>>> from teradataml import FeatureStore, DataFrame, load_example_data
Create a FeatureStore for repo 'vfs_v1'.
>>> fs = FeatureStore("vfs_v1")
Repo vfs_v1 does not exist. Run FeatureStore.setup() to create the repo and setup FeatureStore.
Set up FeatureStore for this repository.
>>> fs.setup()
True
Example 1: Create a Feature for column 'Feb' from 'sales' DataFrame and register with repo 'vfs_v1'
Create a Feature.
>>> from teradataml import Feature
>>> feature = Feature('sales:Feb', df.Feb)
Register the Feature with the repo.
>>> fs.apply(feature)
True
Example 2: create Entity for 'sales' DataFrame and register with repo 'vfs_v1'
Create an Entity.
>>> from teradataml import Entity
>>> entity = Entity('sales:accounts', df.accounts)
Register the Entity with the repo.
>>> fs.apply(entity)
True
Example 3: create DataSource for 'sales' DataFrame and register with repo 'vfs_v1'
Create a DataSource.
>>> from teradataml import DataSource
>>> ds = DataSource('Sales_Data', df)
Register the DataSource with repo.
>>> fs.apply(ds)
True
Example 4: create FeatureStore with all the objects created in above examples and register with repo 'vfs_v1'
Create a FeatureGroup.
>>> from teradataml import FeatureGroup
>>> fg = FeatureGroup('Sales',
... features=feature,
... entity=entity,
... data_source=data_source)
Register the FeatureStore with the repo.
>>> fs.apply(fg)
True