ingest_features() | FeatureGroup Method | Teradata Package for Python - ingest_features() - Teradata Package for Python

Use the ingest_features() method to ingest the features from a feature group. ingest_features() considers the associated DataSource as the data source for the feature process and ingests the feature values in the feature catalog.

Required Parameter

repo

Specifies the name of the repository to ingest the features.

Optional Parameter

data_domain

Specifies the name of the data domain to ingest the features for.

Example setup

>>> load_example_data('dataframe', ['sales'])
>>> df = DataFrame("sales")

>>> from teradataml import FeatureStore
>>> fs = FeatureStore("vfs_v1", data_domain='sales')

Repo vfs_v1 does not exist. Run FeatureStore.setup() to create the repo and setup FeatureStore.

>>> fs.setup()

True

Example 1: create a FeatureGroup 'sales_data_fg' for above mentioned DataFrame and publish it to 'vfs_v1'

Create the features.

>>> jan_feature = Feature("sales:Jan", df.Jan)
>>> feb_feature = Feature("sales:Fan", df.Feb)
>>> mar_feature = Feature("sales:Mar", df.Mar)
>>> apr_feature = Feature("sales:Apr", df.Apr)

Create an entity.

>>> entity = Entity("sales:accounts", df.accounts)

Create a data source.

>>> data_source = DataSource("sales_source", df.show_query())

Create a feature group.

>>> fg = FeatureGroup('Sales',
...                   features=[jan_feature, feb_feature, mar_feature, apr_feature],
...                   entity=entity,
...                   data_source=data_source)

Ingest the features.

>>> fp = fg.ingest_features()

Process 'e04fd157-6c23-11f0-8bd4-f020ffe7fe09' started.
Process 'e04fd157-6c23-11f0-8bd4-f020ffe7fe09' completed.

>>> fp

FeatureProcess(repo=vfs_v1, data_domain=sales, process_id=e04fd157-6c23-11f0-8bd4-f020ffe7fe09)