10 beginner questions on the Data Engineer Associate exam outline: Databricks Intelligence Platform, Data Ingestion and Loading, Data Transformation and Modeling, Working with Lakeflow Jobs, Implementing CI/CD, Troubleshooting, Monitoring, and Optimization, and Governance and Security. They cover Delta Lake time travel, Auto Loader, COPY INTO, the medallion architecture, deduplication, job triggers, bundles, liquid clustering and Unity Catalog tables and privileges.

Ready to test yourself? Take the graded Beginner test (free; sign in with Google, one click) to get your score.

  1. A data engineer loaded bad records into a Delta table this morning and needs to query the table as it looked before that load. Which Delta Lake feature allows this?

    • A. VACUUM with a retention period
    • B. Time travel with VERSION AS OF ✔ correct
    • C. OPTIMIZE with file compaction
    • D. Liquid clustering on the date column

    Answer: B. Delta Lake keeps table history, so time travel can query an earlier version with VERSION AS OF or TIMESTAMP AS OF, and RESTORE can roll the table back. VACUUM removes old files and limits time travel.

    Source: docs.databricks.com/aws/en/delta/history · Domain: Databricks Intelligence Platform

  2. A team wants to process new JSON files incrementally as they arrive in cloud object storage, without tracking file names themselves. Which Structured Streaming source does Auto Loader provide?

    • A. delta
    • B. jdbc
    • C. kafka
    • D. cloudFiles ✔ correct

    Answer: D. Auto Loader provides the cloudFiles source, used as spark.readStream.format("cloudFiles"). It processes new files as they arrive in cloud storage and keeps track of the files it has already ingested.

    Source: docs.databricks.com/aws/en/ingestion/cloud-object-storage/auto-loader/ · Domain: Data Ingestion and Loading

  3. Running the same COPY INTO statement again skips the files in the source location that were already loaded into the target table.

    • A. True ✔ correct
    • B. False

    Answer: A. True. COPY INTO is retriable and idempotent: files that were already loaded are skipped on later runs, so a scheduled rerun loads only new files.

    Source: docs.databricks.com/aws/en/ingestion/cloud-object-storage/copy-into/ · Domain: Data Ingestion and Loading

  4. In the medallion architecture, which layer typically holds validated, cleaned and deduplicated data that is ready for joins and further modeling?

    • A. Bronze
    • B. Silver ✔ correct
    • C. Gold
    • D. Landing

    Answer: B. Bronze stores raw, unvalidated data as ingested. Silver holds validated, cleaned and deduplicated data. Gold holds aggregated, business-level tables for analytics, dashboards and machine learning.

    Source: docs.databricks.com/aws/en/lakehouse/medallion · Domain: Data Transformation and Modeling

  5. Which two PySpark DataFrame methods return a new DataFrame with duplicate rows removed? (Select TWO)

    • A. distinct() ✔ correct
    • B. dropDuplicates() ✔ correct
    • C. union()
    • D. explode()
    • E. fillna()

    Answer: A, B. distinct() removes duplicate rows, and dropDuplicates() does the same, optionally considering only selected columns. union() appends rows, explode() turns array elements into rows and fillna() replaces nulls.

    Source: spark.apache.org/docs/latest/api/python/reference/pyspark.sql/api/pyspark.sql.Da · Domain: Data Transformation and Modeling

  6. A data team needs a job to start as soon as new files land in a Unity Catalog external location, rather than at a fixed time. Which Lakeflow Jobs trigger fits best?

    • A. A scheduled trigger with a daily cron expression
    • B. A file arrival trigger on the storage location ✔ correct
    • C. A manual run started from the Jobs UI
    • D. A table update trigger on the job's output table

    Answer: B. A file arrival trigger starts a job run when new files arrive in an external location. A schedule is time-based, and a table update trigger reacts to changes in tables, not to raw files landing in storage.

    Source: docs.databricks.com/aws/en/jobs/file-arrival-triggers · Domain: Working with Lakeflow Jobs

  7. A team packages a job and its pipeline as a Declarative Automation Bundle (formerly Databricks Asset Bundles). Which Databricks CLI command deploys the bundle to a target workspace?

    • A. databricks bundle validate
    • B. databricks bundle init
    • C. databricks bundle deploy ✔ correct
    • D. databricks bundle destroy

    Answer: C. databricks bundle deploy deploys the bundle's resources to the target workspace (pick a target with -t). validate only checks the configuration, init creates a project from a template and destroy removes deployed resources.

    Source: docs.databricks.com/aws/en/dev-tools/bundles/work-tasks · Domain: Implementing CI/CD

  8. A data engineer wants a data layout optimization that replaces table partitioning and ZORDER, and lets clustering keys be changed later without rewriting existing data. Which feature should they use?

    • A. Liquid clustering ✔ correct
    • B. Delta change data feed
    • C. Deletion vectors
    • D. Time travel

    Answer: A. Liquid clustering replaces partitioning and ZORDER, and clustering keys can be redefined without rewriting existing data. Change data feed records row changes, deletion vectors speed up deletes and time travel reads old versions.

    Source: docs.databricks.com/aws/en/delta/clustering · Domain: Troubleshooting, Monitoring, and Optimization

  9. Dropping an external table in Unity Catalog also deletes its data files from cloud storage.

    • A. True
    • B. False ✔ correct

    Answer: B. False. Unity Catalog does not delete the underlying data in cloud storage when you drop an external table; only the table metadata is removed. For managed tables, Unity Catalog manages the data lifecycle.

    Source: docs.databricks.com/aws/en/tables/external · Domain: Governance and Security

  10. An analyst group already has USE CATALOG and USE SCHEMA on the parent objects. Which privilege on the table main.sales.orders lets the group read its rows?

    • A. MODIFY
    • B. SELECT ✔ correct
    • C. READ VOLUME
    • D. CREATE TABLE

    Answer: B. SELECT on a table allows querying it, together with USE CATALOG and USE SCHEMA on its parents. MODIFY allows inserts, updates and deletes, and READ VOLUME applies to files in volumes.

    Source: docs.databricks.com/aws/en/data-governance/unity-catalog/manage-privileges/privi · Domain: Governance and Security

Original practice questions written by xOps School (DevOpsSchool Private Limited) from the public Databricks exam objectives (Exam guide updated 2026-08-06 (exam version published 2026-05-04)), checked 2026-10-25. They are not actual exam questions. xOps School is not affiliated with or endorsed by Databricks; exam names are trademarks of their owners. Official exam page: www.databricks.com/learn/certification/data-engineer-associate.

Last modified: Sunday, 25 October 2026, 9:57 PM