GCP Professional Data Engineer Practice Question

A media company plans to build a nightly batch job that merges three existing datasets before loading results into BigQuery: (1) terabytes of raw video metadata currently stored in an object store, (2) historical analytics already persisted in the enterprise data warehouse, and (3) transactional user-profile data that must remain globally consistent across regions and support SQL queries. Which Google Cloud services should be configured as the batch sources for this pipeline?

Cloud Storage, BigQuery, and Cloud Spanner
Cloud Storage, Pub/Sub, and Cloud SQL
Cloud Storage, Bigtable, and Dataproc
Datastore, BigQuery, and Cloud SQL

Report Issue

Answer Description

For a daily batch workflow you should choose batch-oriented, durable storage systems that already hold the required data:

Raw video metadata in an object store maps directly to Cloud Storage, Google Cloud's highly durable object storage service.
Historical analytics data kept in the existing warehouse can be read from BigQuery, which supports batch exports and queries.
Globally consistent, strongly consistent relational user-profile data is best stored in Cloud Spanner, Google Cloud's multi-region, horizontally scalable relational database.

The combination of Cloud Storage, BigQuery, and Spanner therefore covers all three data sources and is fully supported as batch inputs. The other options are unsuitable:

Pub/Sub is primarily a streaming ingestion service, not intended as a batch data store.
Datastore is a document NoSQL database and does not provide strong SQL semantics or multi-region consistency for relational workloads.
Bigtable is a wide-column NoSQL store optimized for low-latency access, not relational queries, and Dataproc is a processing engine, not a data source. Hence, these alternatives do not satisfy the stated requirements.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.