Feature Store Management MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Which of the following is a primary benefit of creating feature store tables at the account level in Unity Catalog compared to the workspace level?

   • AFaster query performance due to data replication across all workspaces.
   • BAutomatic backup of feature data to a separate cloud region.
   • CThe ability to use the feature store without any compute resources.
   • DCentralized governance and unified metadata management across all workspaces in the account.
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   When training a model with features from a feature store table, what is the typical workflow?

   • ALoad the feature DataFrame using the Feature Store API, then join it with the training labels and train the model.
   • BUse the Feature Store API to train the model directly without loading the data.
   • CCreate a new feature table for each training run to avoid data leakage.
   • DExport the feature table to a CSV file and then read it into the training script.
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A feature table is updated by multiple streaming jobs that write to the same primary key. The team wants to ensure that the latest value is always reflected in the table without duplicates. Which write mode should they use in the Feature Store API?

   • Aerror
   • Boverwrite
   • Cmerge
   • Dappend
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   When creating a feature store table in Unity Catalog, which of the following must be specified?

   • AThe catalog and table name, but the schema is automatically inferred from the data.
   • BThe primary key and a timestamp column for time-travel queries.
   • COnly the table name and the primary key.
   • DThe catalog, schema, and table name, along with the primary key and schema.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A multinational company has multiple Databricks workspaces across different regions. The data science teams in each region want to collaborate on a shared set of features for a global model. They need centralized access control and the ability to share features across workspaces without duplicating data. What should they implement?

   • ACreate separate feature tables in each workspace and periodically sync them.
   • BStore feature tables in a shared cloud storage bucket and mount it in each workspace.
   • CUse workspace-level tables and grant cross-workspace access via API keys.
   • DUse a single Unity Catalog metastore with account-level feature tables.
   Reveal Solution Discussion Select an answer first