FeatureStore
A FeatureStore object manages the connections and location details for source tables in your data warehouse, and it also acts as the database where partial aggregations (tiles) and online feature values for historical and online serving are stored.
Registering a FeatureStore¶
To connect to your data warehouse, refer to the "Connecting to your Data Warehouse" section in the Get Started documentation. It summarizes the steps to connect to your:
- Snowflake data warehouse
- DataBricks data warehouse
- Spark data warehouse
Exploring the Feature Store's source tables¶
To explore the available source tables through the FeatureStore object, follow these steps:
-
List the registered feature stores using the
list()
class method: -
Retrieve the FeatureStore object by its name using the
get()
class method: -
Obtain the DataSource object associated with the Feature Store object using the
get_data_source()
method: -
Consult the DataSource and SourceTable documentation to learn how to explore each individual source table that the feature store can access.
Monitoring Deployments and Feature Job status¶
Once feature lists have been deployed and are being served, use the Deployment object class methods to list deployments across all catalogs:
list()
to list all deployments across catalogs.get()
to get an Deployment object by its name.get_by_id()
to get a Deployment object by its Object ID.
You can also get a report on the recent activity of scheduled feature jobs using the get_feature_jobs_status()
method of a given deployed feature list.