Features can be read in two modes latest and for target
df_latest = spark.read.table("odap_features.features_client_latest")
%run ../init/target_store
dbutils.widgets.text("target", "")
Insert target name into the created widget, e.g. First mortgage payment
feature_lookups = [
FeatureLookup(
table_name="odap_features_customer.product_features",
feature_names=[
"investice_web_visits_count_in_last_14d",
"pujcky_web_visits_count_in_last_14d",
],
lookup_key="customer_id",
timestamp_lookup_key="timestamp",
),
FeatureLookup(
table_name="odap_features_customer.simple_features",
feature_names=[
"customer_email",
],
lookup_key="customer_id",
timestamp_lookup_key="timestamp",
)
]
training_set = fs.create_training_set(
spark.table("target_store"),
feature_lookups=feature_lookups,
label="target",
)
training_df = training_set.load_df()
training_df.display()