Free Databricks Certified Data Engineer Associate Practice Questions

Sample Practice Questions

Try these 10 sample questions from the Databricks Certified Data Engineer Associate practice bank. Each includes the answer and explanation. Continue with the interactive practice above to review the full question set.

1. 1Which architecture component hosts the Databricks web application?

   1. A. The classic compute plane
   2. B. The control plane
   3. C. The serverless compute plane
   4. D. The workspace storage bucket

   Answer: B

   Explanation: The documentation states, "The web application is in the control plane." Compute planes process workloads; the workspace storage bucket stores data rather than hosting the application.

   Reference: High-level architecture — Workspace architecture

2. 2How does triggered processing support nightly batch ingestion in a Lakeflow pipeline?

   1. A. It continuously polls and processes arrivals between nightly runs.
   2. B. It processes new data when scheduled or manually triggered.
   3. C. It processes source data only once over the pipeline's lifetime.
   4. D. It requires a user to start every update and cannot be scheduled.

   Answer: B

   Explanation: Triggered processing is described as "Triggered: Processes new data on a schedule or when manually triggered." A nightly schedule invokes batch processing; it does not require continuous execution or a manual start for every update.

   Reference: Ingest with SQL and Spark APIs — Pipeline modes

3. 3A bronze DataFrame has nullable customer_id and optional comment columns. Which operation drops rows with missing customer_id while retaining rows whose only null is comment?

   1. A. df.na.drop("any")
   2. B. df.na.drop(subset=["customer_id"])
   3. C. df.na.drop("all")
   4. D. df.na.fill({"customer_id": 0})

   Answer: B

   Explanation: For dropna, subset is the "optional list of column names to consider." Restricting it to customer_id preserves optional null comments. Dropping across all columns is too broad; filling identifiers changes data rather than removing invalid rows.

   Reference: pyspark.sql.DataFrame.dropna — Parameters — subset

4. 4A task occasionally fails because of a transient source outage. Which task setting controls automatic re-execution after failure?

   1. A. Run-duration warning
   2. B. Maximum concurrent job runs
   3. C. Task timeout
   4. D. Retry policy

   Answer: D

   Explanation: The retry policy controls "when and how many times failed task runs are retried". A warning reports duration, concurrency governs overlapping runs, and timeout terminates an overlong attempt rather than configuring re-execution.

   Reference: Configure and edit tasks in Lakeflow Jobs — Advanced task settings — Set a retry policy

5. 5When creating a new branch in the Databricks Git dialog, which information is selected?

   1. A. A branch name and a commit message
   2. B. A branch name and its base branch
   3. C. A base branch and a pull request title
   4. D. A commit message and an author email

   Answer: B

   Explanation: The workflow requires "Enter a name for the new branch and select the base branch." Commit messages and author email describe commits; a pull request title describes review, not the new branch’s starting point.

   Reference: Create and manage Git folders — Create a new branch

6. 6Which Lakeflow Jobs view shows historical runs together with their task results so current execution durations can be compared with earlier runs?

   1. A. The job's run-history matrix
   2. B. The Graph view for one job run
   3. C. The Timeline view for one job run
   4. D. The task details page for one job run

   Answer: A

   Explanation: The guide states, "The matrix view shows a history of runs for the job, including each job task." A single-run graph, timeline, or task detail does not provide this multi-run comparison matrix.

   Reference: Monitor Lakeflow Jobs — View runs for a job

7. 7How do managed and external tables differ in responsibility for underlying storage and maintenance?

   1. A. Unity Catalog manages storage and maintenance for both table types.
   2. B. Users manage storage and maintenance for both table types.
   3. C. Unity Catalog manages storage for external tables; users manage it for managed tables.
   4. D. Unity Catalog manages storage for managed tables; users manage it for external tables.

   Answer: D

   Explanation: For managed tables, "Unity Catalog manages all read, write, storage, and optimization responsibilities." External tables retain Unity Catalog governance but leave the underlying data lifecycle and layout to the user.

   Reference: Unity Catalog managed tables — Introduction

8. 8For classic Databricks compute on AWS, where do the compute resources run?

   1. A. In the Databricks control-plane account
   2. B. In an account shared by all customers
   3. C. In the customer's AWS account
   4. D. In the serverless compute-plane account

   Answer: C

   Explanation: For classic compute, "Databricks compute resources run in your AWS account." This distinguishes classic resources from the Databricks-managed serverless plane and control plane.

   Reference: High-level architecture — Classic compute plane

9. 9How does continuous Lakeflow pipeline processing handle new source data?

   1. A. It processes only the snapshot present when the pipeline is created.
   2. B. It waits for the next configured batch schedule before processing.
   3. C. It processes new data as that data arrives in the source.
   4. D. It requires an operator to start a separate update for each file.

   Answer: C

   Explanation: The documented behavior is "Continuous: Processes new data as it arrives in the source." Continuous processing does not wait for a later batch schedule, stop after an initial snapshot, or require per-file manual updates.

   Reference: Ingest with SQL and Spark APIs — Pipeline modes

10. 10The DataFrame has numeric amount and STRING category columns. Which operation replaces null amount values with 0 and null category values with Unknown while preserving rows?

    1. A. df.na.drop(subset=["amount", "category"])
    2. B. df.na.fill("Unknown", subset=["amount", "category"])
    3. C. df.na.fill({"amount": 0, "category": "Unknown"})
    4. D. df.na.fill(0, subset=["amount", "category"])

    Answer: C

    Explanation: A fill dictionary is "a mapping from column name (string) to replacement value." It supplies type-appropriate defaults per column; a single numeric or string default does not replace both types, and drop removes rows.

    Reference: pyspark.sql.DataFrame.fillna — Parameters — value