Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Cloud Digital Leader practice bank has an additional 3,070 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
An enterprise is modernizing its data value chain on Google Cloud to transform raw information into actionable business insights. The organization has the following architectural requirements:
Which architectural approach best satisfies these requirements?
A data lakehouse is a unified data architecture that combines the cost efficiency and flexible scale of a data lake with the structured data management, governance, and analytical speed of an enterprise data warehouse. Within this architecture, the medallion architecture serves as a logical design pattern that progresses data along the data value chain through three distinct tiers of increasing structure and quality: Bronze (raw), Silver (cleansed), and Gold (curated).
Apache Iceberg or Parquet at a low cost.
Apache Spark, and Apache Flink—can operate on a single copy of data without creating isolated data silos.Unlike maintaining separate, disconnected data lakes and data warehouses that require complex duplication pipelines, a data lakehouse using a medallion architecture unifies storage and analytics while clearly structuring the progression from raw data to actionable business insights.
In a cloud analytics architecture, external tables are table definitions that reference data files and metadata residing outside of a managed lakehouse runtime catalog or native data warehouse storage. They allow query engines like BigQuery to read structured or semi-structured files stored directly in Cloud Storage or third-party cloud object stores without importing the data into native warehouse storage.
Relying exclusively on self-managed external tables outside of a centralized catalog prevents BigQuery from performing write operations or automated storage optimization. Because BigQuery has read-only access to external tables, it cannot write cleansed silver-layer transformations or curated gold-layer aggregations back to those tables, nor can it execute automated maintenance tasks such as data compaction and garbage collection.
Apache Iceberg tables for gold-layer reporting.The organization only needed ad-hoc, read-only querying over existing files stored in Cloud Storage or another cloud provider's object storage without requiring BigQuery to write transformations or manage table compaction.
To achieve progressive data refinement across the data value chain, organizations use managed catalog tables (such as Apache Iceberg tables in the Lakehouse runtime catalog) and BigQuery tables, which support full read-write interoperability, schema evolution, and automated storage optimization.
BigQuery is Google Cloud's fully managed, serverless enterprise data warehouse designed to store and analyze structured and semi-structured relational data at petabyte scale using standard SQL. Cloud Storage is a scalable object storage service commonly used as the foundation for a data lake, capable of storing any file format—including raw unstructured media, backups, and open analytical file formats (Parquet, Avro, ORC).
This option reverses the fundamental architectural roles of a data lake and a data warehouse, as well as the layers of the medallion architecture. Using BigQuery native tables as an object store for raw, unstructured bronze-layer files is architecturally inappropriate because relational warehouse storage requires defined schemas. Conversely, Cloud Storage is an object storage system rather than a SQL query engine, meaning it cannot independently execute high-concurrency analytical queries for the gold consumption layer.
The architectural mapping were inverted so that Cloud Storage served as the bronze-layer data lake for raw and unstructured data ingestion, while BigQuery served as the gold-layer data warehouse for curated, high-concurrency business reporting.
Understanding the distinction between object storage (data lakes) and relational analytical engines (data warehouses) is critical in the data value chain. Raw data starts in flexible storage and moves toward structured warehouse consumption as business value is extracted.
The Apache Hive metastore (HMS) endpoint is a metadata management interface within the Lakehouse runtime catalog designed for backward compatibility with open-source Apache Spark and Apache Hive workloads. Rather than acting as a physical storage engine, a metastore is a catalog directory that tracks database names, table definitions, partitions, and pointers to where actual data files reside.
First, an Apache Hive metastore cannot replace Cloud Storage because a metastore only stores metadata definitions—it still relies on Cloud Storage to physically hold the underlying data files. Second, Hive metastore tables are designed for structured and semi-structured analytical file formats (such as Parquet, ORC, and Avro), not raw unstructured files. Third, for tables managed through the Hive metastore endpoint, BigQuery is restricted to read-only access and cannot perform SQL write transformations.
Spark and Hive) have read and write access, whereas BigQuery is limited strictly to read-only queries.Apache Spark and Apache Hive extract, transform, and load (ETL) pipelines without maintaining a self-hosted metastore.The scenario asked how to migrate legacy Apache Spark and Apache Hive workloads to a managed metadata service where open-source engines handle all ETL writes to Cloud Storage and BigQuery only requires read-only access.
For modern lakehouse workloads requiring read and write interoperability between open-source engines and BigQuery, organizations should use Apache Iceberg tables managed through the Apache Iceberg REST catalog endpoint backed by Cloud Storage.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated