BigQuery
Table sources are used directly via a view. Featureform will never write to a primary source.
SQL transformations are used to create a view. By default, those views are materialized and updated according to the schedule parameter. Deprecated transformations are converted to un-materialized views to save storage space.
When a feature is registered, Featureform creates an internal transformation to get the newest value of every feature and its associated entity. A Kubernetes job is then kicked off to sync this up with the Inference store.
Every registered feature and label is associated with a view table. That view contains three columns, the entity, value, and timestamp. When a training set is registered, it is created as a materialized view via a JOIN on the corresponding label and feature views.
First we have to add a declarative BigQuery configuration in Python.
bigquery_config.py
import featureform as ff
bigquery = ff.register_bigquery(
name="bigquery-quickstart",
description="A BigQuery deployment we created for the Featureform quickstart",
project_id="<your-gcp-project-id>",
dataset_id="<your-big-query-dataset>",
credentials_path="<path-to-bigquery-credentials-file>"
)
Once our config file is complete, we can apply it to our Featureform deployment
featureform apply bigquery_config.py --host $FEATUREFORM_HOST
We can re-verify that the provider is created by checking the Providers tab of the Feature Registry.
descriptioncredentials