Skip to content
DatabricksPractice bank availableAssociate

Databricks Certified Data Engineer Associate: exam facts and practice questions

Verified exam facts from Databricks’s published guide, checked on 20 November 2026, and an original practice bank with 168 independently validated questions.

Verified exam facts

What Databricks publishes about this exam

Recorded on 20 November 2026. Confidence in the record: high. Check Databricks's own pages before you book.

Exam code
No exam code published
Exam version
Exam Published May 4, 2026; Exam Guide Updated August 6, 2026Effective 6 August 2026
Scored questions
45
Time limit
90 minutes
Unscored questions
Up to 10 unscored items may be included; they are not identified on the exam and do not affect the score. 90 minutes (additional time is already factored in to account for unscored items).
Question format
multiple-choice
Delivery
Online proctored or test center (delivered through Kryterion per the certification page FAQ)
Languages
English, Japanese, Brazilian Portuguese (pt-BR), Korean
Test aids
None allowed
Exam price
US$200
Certification validity
2 years
Prerequisites
None required; related training highly recommended.
Recommended experience
Six months of hands-on experience in the Databricks Platform, with the listed exam outline, are highly recommended (guide p.2). The certification page says: hands-on experience performing the data engineering tasks outlined in the exam guide.
Passing score
Databricks does not publish a passing score for this exam.CertExam does not state one. Neither the guide nor the certification page states a passing score. The certification page FAQ says Databricks does not publish passing scores; they are set through statistical analysis and can change as the exam is updated with new questions.
Facts checked on
20 November 2026
What we could not verify (5)
  • Exam code: none published in the guide or on the certification page.
  • Passing score: not published by the vendor; CertExam records none.
  • Question counts per section and per objective: the guide gives only section weights, not item counts.
  • Renewal exam details (format, length, objectives) were not read; only its existence is stated in the guide (p.2). Its page is listed in officialResources but was not fetched.
  • Kryterion delivery and the 14-day retake rule come from the certification page FAQ, not the guide.

Exam outline

Domains and weights

The published outline has 7 domains; the weights add up to 100 percent. Open a domain to read its objectives.

  1. 1. Databricks Intelligence Platform6 percent

    Objectives

    • Understand the core components of the Databricks Platform, such as its architecture, Delta Lake, and Unity Catalog.
    • Understand Databricks Platform's compute services, including their characteristics, limitations, and cost models, and select the most suitable option for each workload use case.
  2. 2. Data Ingestion and Loading21 percent

    Objectives

    • Enable and detail data ingestion patterns, including batch, streaming, and incremental loading, and import data from sources such as local files, Lakeflow Connect standard connectors, and Lakeflow Connect managed connectors.
    • Use the COPY INTO command to incrementally load files from cloud object storage (ADLS/S3/GCS) into Unity Catalog-governed tables.
    • Use Auto Loader with schema enforcement and schema evolution in batch modes (for example, directory listing or file notification) to land data into Unity Catalog-governed tables.
    • Configure Lakeflow Connect to reliably ingest data from diverse enterprise sources into Unity Catalog-governed tables.
    • Use JDBC/ODBC or REST clients in notebooks to land data into cloud storage or directly into Unity Catalog-governed tables, usually orchestrated and scheduled with Lakeflow Jobs.
    • Prioritize between Auto Loader, Lakeflow Connect (standard and managed connectors), partner connectors, and other ingestion methods based on technical requirements such as data volume, ingestion frequency, data types, and governance needs with Unity Catalog.
    • Ingest semi-structured and unstructured data (for example, JSON and nested data) via Lakeflow Connect and other managed connectors into Unity Catalog-governed Delta tables.
  3. 3. Data Transformation and Modeling22 percent

    Objectives

    • Implement data cleaning by reading bronze tables with PySpark/SQL, cleaning nulls, standardizing data types, and writing to new silver tables.
    • Combine DataFrames with operations such as Inner join, left join, broadcast join, multiple keys, cross join, union, and union all.
    • Manipulate columns, rows, and table structures by adding, dropping, splitting, renaming column names, applying filters, and exploding arrays.
    • Perform data deduplication operations and aggregate operations on DataFrames, such as count, approximate count distinct, and mean, summary.
    • Understand the basic tuning parameters (spark.sql.shuffle.partitions, spark.default.parallelism, spark.executor/driver.memory, spark.sql.autoBroadcastJoinThreshold) and re-measure the performance.
    • Understand the difference between, and how to build, Gold layer objects such as materialized views, views, streaming tables, and tables for BI and analytics teams in Unity Catalog.
    • Apply data quality checks and validation rules to ensure reliable Silver and Gold datasets.
  4. 4. Working with Lakeflow Jobs16 percent

    Objectives

    • Implement control flows (retries and conditional tasks such as branching and looping) using Lakeflow Jobs for pipeline orchestration
    • Configure common tasks (notebook, SQL query, dashboard, and pipeline tasks) and their dependencies using Lakeflow Jobs and its DAG-based task graph
    • Implement job schedules using Lakeflow Jobs with an understanding of trigger types (scheduled, file arrival, and table update)
    • Choose between time-based and data-driven triggers based on data availability and pipeline dependencies.
  5. 5. Implementing CI/CD10 percent

    Objectives

    • Manage your code development workflow within the Databricks workspace UI, including creating and switching between branches in Databricks Git folders (formerly Databricks Repos), committing and pushing changes, and creating pull requests using Databricks Git integration.
    • Understand environment-specific configuration using Declarative Automation Bundle (formerly Databricks Asset Bundles) variables and overrides while promoting the same codebase across dev, test, and prod targets.
    • Deploy Declarative Automation Bundles (formerly Databricks Asset Bundles) to package, configure, and promote Lakeflow Jobs, Lakeflow Declarative Pipelines, and other workspace assets across dev, test, and prod environments.
    • Understand the Databricks CLI to validate, deploy, and manage Declarative Automation Bundles (formerly Databricks Asset Bundles) and other workspace assets in automated CI/CD workflows.
  6. 6. Troubleshooting, Monitoring, and Optimization10 percent

    Objectives

    • Identify trends in job performance using the Lakeflow Jobs run history view to compare current execution times against historical baselines.
    • Use the Lakeflow Jobs UI to monitor pipeline health by interpreting job statuses, viewing DAG-based task graphs to spot upstream blockers, and tracking pipeline run times and failure rates.
    • Identify common performance bottlenecks such as data skew, shuffling, and disk spilling by interpreting stage-level metrics in the Spark UI.
    • Understand the features of Liquid Clustering and predictive optimization.
    • Diagnose cluster startup failures, library conflicts, and out-of-memory issues.
  7. 7. Governance and Security15 percent

    Objectives

    • Differentiate between managed and external tables in Unity Catalog and perform basic operations (create, modify, delete, and convert between managed and external tables) on them.
    • Configure access controls using the UI and SQL by applying GRANT, REVOKE, and DENY privileges to principals (users, groups, and service principals) at appropriate levels of the security hierarchy.
    • Understand column-level masking and row-level security to restrict data visibility based on user groups.
    • Understand Unity Catalog ABAC policies to centrally control row-level filtering and column masking for sensitive data.

What CertExam offers

Practice bank

168 questions are published after passing automated checks and an independent blind verification by someone other than the author; 40 of them are free to try without a card. Drafted-but-unverified questions are never counted or shown.

This bank meets CertExam's own launch criteria: at least 150 independently validated questions, every exam-guide objective covered by at least two questions, enough questions for two non-overlapping full-length mock exams, and every option explained with a citation to official public documentation that a second, independent check opened. That is CertExam's own bar, not the vendor's. It does not claim complete coverage of the exam, it says nothing about your result, and it does not mean the questions resemble real exam items. All questions are original.

Which modes you can use depends on your plan. See plans and pricing.

Practice modes

  • Quick practice
  • Diagnostic
  • Topic practice
  • Full mock exam
  • Weak areas
  • Review incorrect
  • Scenario practice
  • Code and SQL
  • Adaptive practice
  • Spaced repetition

Free samples

Sample questions with answers

7 sample questions from the bank, shown with their answers and explanations. Think it through, then open the answer. Create a free account to keep your progress on the free sample set.

Sample question 1

intermediateDomain 1, objective 2

A team runs SQL-only ELT and BI queries on Databricks on AWS. Their network design keeps an event bus and several databases in their own cloud network, and queries must federate to an on-premises database over that network. The workspace is in a region that does not offer serverless SQL warehouses, and the team wants better than entry-level performance. Which SQL warehouse type fits best?

Choose one answer.

  1. Option A:

    A serverless SQL warehouse.

  2. Option B:

    A pro SQL warehouse.

  3. Option C:

    A classic SQL warehouse.

  4. Option D:

    A Lakehouse Real-Time SQL warehouse.

Show answer and explanation for sample question 1

Correct answer: B

Why

Serverless SQL warehouses are recommended when available. A pro SQL warehouse is the choice when serverless is not available in a region or when you need custom-defined networking for federation or hybrid setups, because its compute layer sits in your own cloud account. Classic warehouses give only entry-level performance.

Every option

  • A(incorrect)

    Incorrect. Serverless is the recommended type when it is available, but this region does not offer it. Its compute also runs in the Databricks account rather than in the team's own network.

  • B(correct)

    Correct. A pro SQL warehouse supports Photon and Predictive IO, and its compute layer is in your AWS account. It is the documented choice when serverless is not available in a region or when custom-defined networking is needed for federation or hybrid architectures.

  • C(incorrect)

    Incorrect. A classic warehouse also runs in your account but supports only Photon, so it gives entry-level performance. It is meant for basic interactive exploration when serverless or pro are not options.

  • D(incorrect)

    Incorrect. Lakehouse Real-Time is a serverless warehouse type, so it is not available in this region. It is also built for sub-second read queries and supports only SELECT and CREATE TABLE AS SELECT, which does not fit ELT.

Sample question 2

foundationalDomain 2, objective 1

An analyst has a zip archive on a laptop that contains four CSV exports, about 300 MB in total. The analyst needs the data in a managed Delta table in Unity Catalog, has permission to create tables in the target schema, and can use a running SQL warehouse. No code should be written.

Which approach loads the data?

Choose one answer.

  1. Option A:

    Upload the zip archive on the Create or modify a table using file upload page, which unpacks it during import

  2. Option B:

    Extract the four CSV files and upload them together on the Create or modify a table using file upload page

  3. Option C:

    Upload each CSV file in its own upload, because the page accepts only one file at a time

  4. Option D:

    Run COPY INTO from a SQL warehouse with the laptop folder as the source location of the files

Show answer and explanation for sample question 2

Correct answer: B

Why

The Create or modify a table using file upload page is the Databricks UI route for small local files. It supports CSV, TSV, JSON, Avro, Parquet, Excel and text files, up to 10 files at a time with a total size under 2 GB, and it does not accept compressed archives. Extracting the CSV files first meets those limits, and rows from files uploaded together are appended to one table.

Every option

  • A(incorrect)

    Incorrect. Compressed files such as zip and tar files are not supported by the file upload page, even though the total size is well under the limit.

  • B(correct)

    Correct. The page accepts up to 10 files at a time with a total size under 2 GB, and CSV is a supported format. Files uploaded together are appended as rows into the one target table.

  • C(incorrect)

    Incorrect. The one-file limit applies to Excel files. The page accepts up to 10 CSV, TSV, JSON, Avro, Parquet, or text files at a time.

  • D(incorrect)

    Incorrect. The documented sources for COPY INTO are cloud object storage and Unity Catalog volumes, not a folder on a user's local machine. It is also a code-based approach, while the analyst wants the UI.

Sample question 3

foundationalDomain 3, objective 6

A retail team's Gold layer needs a daily_revenue result that aggregates roughly two billion Silver order rows. Fifty BI analysts open the same dashboard every morning, and each query re-runs the full aggregation, which is slow. The team wants the result stored in Unity Catalog so that dashboard queries read precomputed data, and the Silver source is updated with inserts, updates and deletes. Which object should the engineer create?

Choose one answer.

  1. Option A:

    A standard view defined over the Silver orders table

  2. Option B:

    A temporary view defined over the Silver orders table

  3. Option C:

    A materialized view defined over the Silver orders table

  4. Option D:

    A streaming table defined over the Silver orders table

Show answer and explanation for sample question 3

Correct answer: C

Why

A materialized view stores the precomputed result of its defining query and is refreshed by Databricks, so many consumers read the cached result instead of recomputing it. Standard and temporary views store no data, and streaming tables target append-only, process-once workloads rather than aggregations over changing sources.

Every option

  • A(incorrect)

    Incorrect. A view does not store data; only its query text is registered, so every dashboard query would still recompute the aggregation.

  • B(incorrect)

    Incorrect. A temporary view is not registered to a schema or catalog and is scoped to a notebook or query, so it cannot serve a shared dashboard and stores nothing.

  • C(correct)

    Correct. A materialized view precomputes and stores the query result in a Unity Catalog table and refreshes it, so dashboards read cached results and the aggregation stays consistent with a source that changes.

  • D(incorrect)

    Incorrect. Streaming tables are designed for append-only sources and process each record once, so they are a poor fit for an aggregate over a source that receives updates and deletes.

Sample question 4

foundationalDomain 4, objective 4

A partner uploads CSV files to a Unity Catalog volume at irregular times, anywhere from zero to six times a day. An hourly scheduled job processes them. Most runs find nothing new and still pay for compute startup, while files that arrive just after a run wait almost an hour. Which change best fits this arrival pattern?

Choose one answer.

  1. Option A:

    Replace the hourly schedule with a file arrival trigger on the volume path.

  2. Option B:

    Switch the job to a continuous trigger so it is always running.

  3. Option C:

    Shorten the schedule to every 10 minutes so files wait less time.

  4. Option D:

    Add a table update trigger on the table the job writes to.

Show answer and explanation for sample question 4

Correct answer: A

Why

When data arrives at irregular times, a time-based schedule either wastes runs or delays processing. A file arrival trigger starts runs from the data itself. The documentation presents it as useful when a scheduled job's efficiency is compromised by irregular new data arrivals.

Every option

  • A(correct)

    Correct. File arrival triggers start a run when new files land, which suits irregular arrivals where a fixed schedule wastes runs and adds delay.

  • B(incorrect)

    Incorrect. A continuous job keeps running all the time, which suits always-on streaming workloads and would cost more here than running only when files arrive.

  • C(incorrect)

    Incorrect. This still polls on the clock, so it would start even more runs that find nothing and use compute each time.

  • D(incorrect)

    Incorrect. The job's own output table changes only after a run, so a trigger on it cannot start the run that processes the new files.

Sample question 5

intermediateDomain 5, objective 1

Marcus has feature-ingest checked out in his Git folder and has edited config.py without committing. He opens the Git dialog and uses the branch dropdown to check out main. His uncommitted edit does not conflict with any code on main. What happens to the edit?

Choose one answer.

  1. Option A:

    Databricks stashes the edit and reapplies it automatically the next time he checks out feature-ingest.

  2. Option B:

    Databricks discards the edit during the checkout, so he must redo it on whichever branch he wants.

  3. Option C:

    The edit carries over and shows as an uncommitted change on main, where he can commit it or discard it.

  4. Option D:

    Databricks blocks the checkout until he commits and pushes the edit to feature-ingest.

Show answer and explanation for sample question 5

Correct answer: C

Why

When you switch branches in a Git folder, uncommitted changes that do not conflict with the new branch come along and appear as uncommitted changes there. If you do not want them to travel, discard them before or after the switch.

Every option

  • A(incorrect)

    Incorrect. The documentation describes uncommitted changes carrying over to the new branch, not an automatic stash tied to the old branch. Stashing is something you run yourself with the Git CLI.

  • B(incorrect)

    Incorrect. Uncommitted changes are not discarded by a branch switch. They are discarded only if the user chooses to discard them before or after switching.

  • C(correct)

    Correct. Uncommitted changes carry over to the new branch when they do not conflict with code on that branch.

  • D(incorrect)

    Incorrect. A non-conflicting uncommitted change does not block switching. Committing first is a good habit but the dialog does not require it.

Sample question 6

intermediateDomain 6, objective 1

A job normally completes in about 40 minutes, based on its run history. The team wants an email whenever a run takes longer than 40 minutes, but late runs must be allowed to finish and keep their normal Succeeded status. Which configuration meets this requirement?

Choose one answer.

  1. Option A:

    Set the job's Timeout to 40 minutes and add an email notification for Failure

  2. Option B:

    Set a streaming backlog threshold of 40 minutes and add an email notification for Streaming backlog

  3. Option C:

    Add a task-level Failure notification on the final task with Mute notifications until the last retry selected

  4. Option D:

    Set the Run duration Warning threshold to 40 minutes and add an email notification for Duration warning

Show answer and explanation for sample question 6

Correct answer: D

Why

An expected completion time is configured with the Run duration Warning threshold, and a Duration warning notification sends the alert when a run exceeds it. The separate Timeout field is a maximum completion time and marks the run Timed Out, and failure or backlog notifications respond to other events. Only the duration warning lets slow runs finish normally while still alerting the team.

Every option

  • A(incorrect)

    Incorrect. The Timeout field sets the maximum completion time, and a run that exceeds it gets the Timed Out status. That conflicts with the requirement that late runs keep their Succeeded status.

  • B(incorrect)

    Incorrect. Streaming backlog thresholds apply to backlog metrics of streaming sources, not to how long a run takes. They do not detect a slow batch run.

  • C(incorrect)

    Incorrect. Failure notifications are sent when a run or task fails, and the muting option only filters notifications until the last retry. A slow run that finishes successfully triggers neither.

  • D(correct)

    Correct. The Warning field defines the job's expected completion time and triggers an event when the run exceeds it. A Duration warning notification turns that event into an email while the run continues and ends with its normal status.

Sample question 7

foundationalDomain 7, objective 1

A data engineer registered main.raw.clickstream as a Unity Catalog external table whose LOCATION is s3://acme-landing/clickstream/. The table is retired and the engineer runs DROP TABLE main.raw.clickstream;. What happens to the table and to the files in that location?

Choose one answer.

  1. Option A:

    Unity Catalog removes the table metadata from the metastore, and the files in s3://acme-landing/clickstream/ stay where they are.

  2. Option B:

    Unity Catalog removes the table metadata, and the files in s3://acme-landing/clickstream/ are deleted once the 7-day recovery period ends.

  3. Option C:

    Unity Catalog removes the table metadata, and the files in s3://acme-landing/clickstream/ are deleted immediately because the engineer owns the table.

  4. Option D:

    The statement fails, because the files in s3://acme-landing/clickstream/ must be deleted from cloud storage before an external table can be dropped.

Show answer and explanation for sample question 7

Correct answer: A

Why

The managed versus external distinction is about who controls the files. For a managed table, Unity Catalog deletes the data files after the recovery period when the table is dropped. For an external table, it removes only the metadata and leaves the files in the location you chose.

Every option

  • A(correct)

    Correct. Dropping an external table removes only the metadata from the metastore. The underlying files remain in place, and anyone who wants them removed must delete them directly in cloud storage.

  • B(incorrect)

    Incorrect. The deletion of data files after a recovery period (7 days by default) is the documented behavior for managed tables. External table files are not deleted by Unity Catalog on drop.

  • C(incorrect)

    Incorrect. Ownership allows the engineer to drop the table, but it does not make Unity Catalog delete files for an external table. Unity Catalog does not manage the lifecycle of external table files.

  • D(incorrect)

    Incorrect. DROP TABLE succeeds for an external table whose owner or MANAGE holder runs it. Deleting the files is a separate manual step that is needed only if the data must be removed.

FAQ

Data Engineer Associate: common questions

How many questions and how much time does the Data Engineer Associate exam have?

Under the guide we checked (Exam Published May 4, 2026; Exam Guide Updated August 6, 2026), Databricks describes 45 scored questions and a 90-minute time limit. Up to 10 unscored items may be included; they are not identified on the exam and do not affect the score. 90 minutes (additional time is already factored in to account for unscored items). We checked these facts on 20 November 2026; confirm them with Databricks before you book.

What is the passing score for the Data Engineer Associate exam?

Databricks does not publish a passing score for this exam, so CertExam does not state one. Figures quoted on forums or by other providers are guesses unless Databricks publishes them.

Does CertExam offer practice questions for the Data Engineer Associate exam?

Yes. The bank has 168 independently validated questions, 40 of them free to try. The bank meets CertExam's own launch criteria: at least 150 independently validated questions, every exam-guide objective covered by at least two questions, enough questions for two non-overlapping full-length mock exams, and every option explained with a citation to official public documentation that a second, independent check opened. That is CertExam's own bar, not the vendor's. It does not claim complete coverage of the exam, it says nothing about your result, and it does not mean the questions resemble real exam items.

Where do these exam facts come from?

From Databricks's own published pages and exam guide, read and recorded on 20 November 2026. Links to the official sources are on this page. Vendors revise exams, so treat Databricks's pages as the final word.

Is CertExam affiliated with Databricks?

No. CertExam is an independent provider, not affiliated with or endorsed by Databricks. Databricks and the certification name are trademarks of their owner and are used only to describe the exam.

Independence. CertExam is not affiliated with or endorsed by Databricks. Databricks and the name of this certification are trademarks of their owner, used here only to describe the exam. Read our editorial policy.