Home / Databricks Data Engineer Associate

Databricks · Data Engineer Associate

Databricks Certified Data Engineer Associate Practice Exams

Free Databricks Certified Data Engineer Associate Practice Exams from all 7 official domains, with every answer option explained.

Start practicing free Take a mock examSee plans

Databricks Certified Data Engineer Associate

Questions
45

Multiple-choice.

Time
90 min
Exam Price
$200

What is on the Data Engineer Associate exam?

Open a domain for its objectives and their practice questions.

Try free Data Engineer Associate practice questions

Pick an answer to see why each option is right or wrong. A preview of 4 of the 14 free questions, which cover every domain.

Domain 1: Databricks Intelligence Platform

1.2 Understand Databricks Platform's compute services, including their characteristics, limitations, and cost models, and select the most suitable option for each workload use case.easy

A team is moving a nightly Python ETL notebook into a Lakeflow Jobs task. The task needs no custom cluster settings. Which compute does Databricks recommend for it?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

For automated workloads, serverless is the recommended default: fewer settings, fast start-up, automatic scaling and lower cost. Nothing here calls for custom cluster settings.

All-purpose compute suits interactive work; for scheduled jobs Databricks advises avoiding it because it is not optimized for automated workloads.

Classic jobs compute is the choice when a non-SQL job needs cluster settings serverless doesn't offer; this task needs none, so it adds configuration for no benefit.

SQL warehouses serve SQL tasks; a Python notebook task needs general-purpose compute instead.

1.1 Understand the core components of the Databricks Platform, such as its architecture, Delta Lake, and Unity Catalog.easy

A data engineer writes a DataFrame to a new table on Databricks with default settings and never names a storage format. What kind of table results?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Delta Lake is the default format on Databricks: Parquet data files plus a file-based transaction log, so each write becomes a new table version with ACID guarantees.

CSV is not the default format, and Databricks describes storing data in DBFS root as a deprecated pattern rather than the default for new tables.

Tables written with default settings are Delta Lake tables, not Hive SerDe tables, so there is nothing to upgrade.

Delta does store data in Parquet files, which makes this tempting, but by default Databricks adds the transaction log on top, so no conversion step is needed.

Domain 2: Data Ingestion and Loading

2.1 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.easy

A team must land Salesforce objects in Unity Catalog tables and is deciding which ingestion layer to start from. What does Databricks guidance recommend?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Databricks recommends starting with the most managed layer and dropping down only if it does not meet requirements, for example when the source is unsupported. Salesforce has a fully-managed connector, so that is the starting point.

It is true that SQL and Spark APIs in Lakeflow pipelines give more customization, which is why this is tempting. But customization is the reason to drop down a layer, not where to start, and nothing here says the managed connector falls short.

DIY connectors are possible on Databricks, but the premise is wrong: fully-managed connectors include automated schema evolution along with authentication, CDC and retries.

Federation avoids copying data, but SaaS data can be ingested into tables; Lakeflow Connect has SaaS connectors for Salesforce specifically.

2.2 Use the COPY INTO command to incrementally load files from cloud object storage (ADLS/S3/GCS) into Unity Catalog-governed tables.easy

A user runs COPY INTO from a path under /Volumes/ and is denied access to the source files. Which privilege on that volume should an admin grant?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Reading source files from a Unity Catalog volume with COPY INTO requires READ VOLUME on that volume, and it applies to all nested directories under it.

CREATE TABLE concerns creating the target table, not reading the source files, so it would not clear a denial on the volume.

READ FILES is the matching privilege when the source path is a Unity Catalog external location, which is why it looks right. The source here is a volume, so the volume privilege is the one needed.

COPY INTO only reads from the source volume; writing to the volume is not what the load needs.

The other 10 free questions, by domain:

Domain 3: Data Transformation and Modeling · 2 questions
3.1 Implement data cleaning by reading bronze tables with PySpark/SQL, cleaning nulls, standardizing data types, and writing to new silver tables.easy

A new bronze table ingests JSON events from cloud storage. An engineer proposes casting every field to its final numeric or date type during bronze ingestion so downstream code is simpler. What does Databricks medallion guidance recommend instead?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Bronze keeps data close to its raw form, mostly as string, VARIANT or binary, so a source schema change does not drop data. Type casting is one of the operations performed when silver tables are built from bronze.

Silver exists to deliver validated, typed records; pushing casting to every analyst query skips the cleanup the silver layer is responsible for.

Gold holds aggregated, business-aligned data. Casting has to happen before aggregation, in silver, or the aggregates would be built on untyped values.

Schema enforcement and type casting are silver-layer operations; bronze is meant to do minimal validation, so moving casting there and dropping enforcement in silver inverts the pattern.

3.2 Combine DataFrames with operations such as Inner join, left join, broadcast join, multiple keys, cross join, union, and union all.easy

A silver job joins customers (left) to orders (right) on customer_id. The output must list every customer, with the order columns set to NULL for customers who have not ordered yet. Which join type meets this?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

A semi join sounds close because it is driven by the left table, but it returns only left rows that do have a match and no columns from the right side.

An inner join keeps only customers that have a matching order, so customers without orders would disappear.

A cross join pairs every customer with every order, ignoring customer_id entirely.

A left outer join keeps every row of the left table and fills the right-side columns with NULL when no order matches.

Domain 4: Working with Lakeflow Jobs · 2 questions
4.2 Configure common tasks (notebook, SQL query, dashboard, and pipeline tasks) and their dependencies using Lakeflow Jobs and its DAG-based task grapheasy

A data engineer adds a SQL task to a Lakeflow Jobs job to run a saved query after the nightly load. Which compute can this task use?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

SQL tasks run on Databricks SQL and need a serverless or pro SQL warehouse; the query, alert or file runs against the warehouse selected in the task.

Job clusters serve tasks such as notebooks and scripts; a SQL task needs a SQL warehouse instead of a cluster.

Any SQL warehouse may seem acceptable, but the SQL task requirement names only serverless or pro warehouses, so a classic one does not qualify.

All-purpose clusters run notebooks and other code tasks, but a SQL task is configured against a SQL warehouse.

4.1 Implement control flows (retries and conditional tasks such as branching and looping) using Lakeflow Jobs for pipeline orchestrationeasy

A data engineer must run one ingestion notebook 30 times inside a single Lakeflow Jobs job, once per source system, with each run receiving a different source system name. Which task setup fits this requirement?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Schedules decide when a job starts, not how a task iterates over values inside one run. The requirement is a loop within a single job run.

Retries rerun a task only after it fails and reuse the same parameters, so they cannot hand each attempt a different source system name.

The For each task loops over an input array and runs its nested task once per element, passing that element as the iteration's parameters, which is exactly one notebook run per source system.

If/else is tempting because it reads a parameter, but it evaluates one boolean expression and picks a branch. It does not repeat a task for a list of values.

Domain 5: Implementing CI/CD · 2 questions
5.1 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.easy

An engineer who is new to Git wants to bring the latest commits from main into a feature branch in a Databricks Git folder. Teammates also work on that feature branch. Which Git folders operation does Databricks recommend?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Merge combines main's history into the feature branch without rewriting commits or force pushing, so teammates' copies of the branch stay consistent.

Reset replaces the branch with main's state and force pushes it, discarding the feature work rather than bringing main into it.

Sparse checkout limits which directories are cloned, for large repositories; it doesn't integrate commits between branches.

Rebase also integrates main and gives a linear history, so it is tempting; but Git folders then force push and the rewritten history can cause versioning issues for the teammates on that branch.

5.2 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.easy

A bundle declares a custom variable named warehouse_id in its top-level variables mapping. Which expression references that variable inside a job task definition?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

The bundle prefix reaches bundle settings such as ${bundle.name} or ${bundle.target}, not custom variables.

variables is the mapping where custom variables are declared, but references use the var prefix, not the mapping name.

The workspace prefix resolves workspace settings such as ${workspace.host} or ${workspace.root_path}, not custom variables.

Custom variables are referenced with the var prefix; the value is resolved from --var, BUNDLE_VAR_ environment variables, overrides, the target or the default.

Domain 6: Troubleshooting, Monitoring, and Optimization · 2 questions
6.2 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.easy

A platform lead wants one place in the Lakeflow Jobs UI that shows which error types caused the most failed job runs across the workspace over the last two days. Where is this summarized?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

The Timeline view is useful for finding slow tasks and overlap inside one run, but it covers a single run, not error trends across all jobs.

The Runs tab under Jobs & Pipelines includes a Top 5 error types table for the selected time range, listing the most frequent error types across runs, which is the workspace-wide summary the lead needs.

Exporting to HTML preserves one notebook run's source and outputs. It produces no cross-job error summary.

Lineage shows upstream and downstream Unity Catalog tables for a job. It does not aggregate run errors.

6.1 Identify trends in job performance using the Lakeflow Jobs run history view to compare current execution times against historical baselines.easy

A compliance team asks a data engineer to keep a record of every run of a nightly job for the past six months, so run durations can later be compared against older baselines. What should the engineer do?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Tempting because the Runs tab is where history is browsed, but it does not keep runs forever: the history window is 60 days, so most of a six-month record would be lost.

Maximum concurrent runs controls how many runs of the job may execute at the same time (a Queued-phase cause), not how long finished runs are retained.

Filtering by Run ID helps find a specific run that is still retained, such as a one-time runs/submit run, but it cannot bring back runs that have aged out of history.

Databricks keeps job run history for up to 60 days, so a six-month baseline only exists if results are exported before older runs expire. Exporting is the documented way to persist runs.

Domain 7: Governance and Security · 2 questions
7.2 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.easy

In Unity Catalog, the group contractors holds a direct SELECT grant on the table hr.payroll.salaries. A security engineer must remove that group's access to the table. Which statement does this?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

DENY is a real SQL statement, but it applies only to the legacy hive_metastore catalog and is not supported by Unity Catalog. Denial in Unity Catalog uses ABAC DENY policies (Beta), not this command.

BROWSE adds metadata visibility; granting it does not replace or remove the existing SELECT grant.

Transferring ownership would give the group all capabilities on the table, the opposite of removing access.

Unity Catalog access exists only through grants, so removing the direct SELECT grant with REVOKE removes the group's access to the table.

7.1 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.easy

A data engineer runs DROP TABLE on a Unity Catalog external table whose files sit in a storage path the team chose. What happens to the underlying data files?

Included in Sample: sample every domain.See plans

Choose an answer to check it. Afterward, select any option to review its explanation.

Dropping a table never relocates its files; moving data to managed storage is what a managed conversion does.

That is what happens when a managed table is dropped, because Unity Catalog controls its file lifecycle. This table is external.

VACUUM removes unreferenced files of an existing table; it is not what governs files after an external table is dropped.

For an external table, Unity Catalog governs the metadata but you own the file lifecycle, so dropping it removes only the metadata and leaves the files where they are.

Practice all 14 free questionsGet 100+ questions

Data Engineer Associate question bank coverage

DomainExam weightFree sampleMaster (full bank)
1. Databricks Intelligence Platform6%210
2. Data Ingestion and Loading21%228
3. Data Transformation and Modeling22%230
4. Working with Lakeflow Jobs16%222
5. Implementing CI/CD10%214
6. Troubleshooting, Monitoring, and Optimization10%214
7. Governance and Security15%220
Total100%14138

Looking for a Data Engineer Associate practice exam? Start with the free questions above, then take the timed mock exam to rehearse the Data Engineer Associate format under exam conditions.

How these Data Engineer Associate questions are made

The domains, weights and objectives come from Databricks's official Data Engineer Associate exam guide (vAugust 2026 (exam published 2026-05-04, guide updated 2026-08-06)). Questions are original and written only from the vendor's official prep material, and each one cites the passage it is based on. Every question is then answered independently, without seeing its key, and every answer option's explanation is checked for accuracy before it is published.