You want to schedule a number of sequential load and transformation jobs. Data files will be added to a Cloud Storage bucket by an upstream process. There is no fixed schedule for when the new data…

Question

Options

• A
  1. Create an Apache Airflow directed acyclic graph (DAG) in Cloud Composer with sequential tasks by using the Cloud Storage, Dataproc, and BigQuery
• B
  1. Create an Apache Airflow directed acyclic graph (DAG) in Cloud Composer with sequential tasks by using the Cloud Storage, Dataproc, and BigQuery
• C
  1. Create an Apache Airflow directed acyclic graph (DAG) in Cloud Composer with sequential tasks by using the Dataproc and BigQuery operators.
• D
  1. Create an Apache Airflow directed acyclic graph (DAG) in Cloud Composer with sequential tasks by using the Dataproc and BigQuery operators.

• A
  3% (1)
• B
  7% (2)
• C
  17% (5)
• D
  72% (21)

Explanation

The choices in your question appear truncated, but based on this well-known GCP exam scenario, the key differentiators are: how the DAG is triggered (Cloud Storage sensor polling vs. Cloud Functions event trigger) and how tables are handled (one DAG per table vs. a single parameterized DAG).

D is correct because it combines event-driven triggering via Cloud Functions (which fires immediately when a file lands in Cloud Storage) with a single parameterized DAG that handles all tables dynamically - this delivers the freshest data and remains maintainable at scale across hundreds of tables.

A is wrong because it typically relies on a Cloud Storage sensor, which polls on a schedule rather than reacting instantly to new files, introducing latency and wasted compute cycles when data arrival is irregular.

B is wrong for the same polling reason, and/or because it creates a separate DAG per table - managing hundreds of individual DAGs becomes an operational nightmare.

C is wrong because while it may use event-driven triggering, it still uses separate DAGs per table, which doesn't scale maintainably to hundreds of tables.

Memory tip: Think "fresh + manageable" - Cloud Functions = instant reaction (freshest data), single parameterized DAG = one DAG to rule them all (most maintainable). Whenever the exam asks about irregular data arrivals, favor event triggers over scheduled sensors.

No community discussion yet for this question.