Implementing Data Engineering Solutions Using Azure Databricks (DP-750) practice questions

Domain coverage 4 domains

• Set up and configure an Azure Databricks environment (15-20%)
• Secure and govern Unity Catalog objects (15-20%)
• Prepare and process data (30-35%)
• Deploy and maintain data pipelines and workloads (30-35%)

Sample questions 14 of 139 shown

1. Q1

   You need to deploy Declarative Automation Bundles to a development environment. The solution must support automated and repeatable deployments across environments. What should you use?

   1. the Databricks CLI
   2. the Databricks SDK for Python
   3. Git folders
   4. the Jobs UI

   The Databricks CLI contains the bundle command group (databricks bundle validate, databricks bundle deploy -t dev) for validating and deploying bundles, and it runs identically from a terminal or a CI/CD pipeline, making deployments automated and repeatable across targets. The Python SDK and Git folders do not deploy bundle-defined resources, and the Jobs UI is a manual, interactive tool rather than an infrastructure-as-code workflow.
2. Q2

   You have an Azure Databricks workspace that is enabled for Unity Catalog and contains a managed Delta table named Payments. Payments stores transaction data and contains a column named payment_amount of the Decimal data type. You must enforce the business rule that payment_amount must be between 0 and 10,000 inclusive. You need to ensure that records that violate the rule are rejected when data is written to Payments. What should you do?

   1. Add a CHECK constraint to payment_amount to enforce the allowed range
   2. Configure row-level security (RLS) to hide records that have invalid values
   3. Run SELECT statements to filter out records that have invalid values in payment_amount
   4. Create a table update trigger

   A CHECK constraint is an enforced Delta Lake constraint that fails any write transaction containing a value outside payment_amount >= 0 AND payment_amount <= 10000, so invalid rows are rejected at the storage layer regardless of which pipeline performs the write. Row-level security only controls which existing rows a user sees, SELECT filters never protect the table itself, and table update triggers are not the Delta Lake mechanism for this requirement.
3. Q3

   You have an Azure Databricks workspace named Workspace1 that contains a cluster named Cluster1. You need to enable a user named User1 to install a Maven package on all the nodes of Cluster1. The solution must follow the principle of least privilege. What should you do?

   1. Assign User1 the Contributor role for the resource group that contains Workspace1
   2. Grant User1 the CAN RESTART permission for Cluster1
   3. Grant User1 the CAN MANAGE permission for Cluster1
   4. Assign User1 the workspace admins role for Workspace1

   Installing a library on every node changes the cluster's library configuration, which requires CAN MANAGE on that specific cluster. CAN RESTART only permits restarting the cluster, while the Azure Contributor role and workspace admin role grant far more privilege than needed, violating least privilege.
4. Q4

   You have an Azure Databricks workspace that is enabled for Unity Catalog. You need to implement a daily batch data process that requires complex and highly customized Python transformations. The solution must minimize additional complexity. What should you include in the solution?

   1. a Databricks notebook triggered by a job
   2. an Azure Data Factory pipeline that has data flows
   3. a continuous job in Lakeflow Jobs
   4. a Lakeflow Spark Declarative Pipelines (SDP) pipeline

   A notebook gives full freedom for complex, highly customized Python/PySpark logic, and scheduling it as a Lakeflow Jobs task supplies native daily orchestration, retries and compute management without adding another service. ADF data flows add external orchestration, a continuous job is wrong for a once-daily workload, and SDP is more declarative than the procedural Python the requirement emphasizes.
5. Q5

   You have an Azure Databricks workspace and a remote Git repository named Repo1 that contains two branches, main and Branch1. You commit changes to Branch1 and must merge them into main. Before completing the merge you must ensure Branch1 includes the changes committed to main since Branch1 was created and that merge conflicts are detected and resolved. What should you do first?

   1. Merge main into Branch1
   2. Pull the latest changes from main
   3. Pull the latest changes from Branch1
   4. Create a pull request for Branch1

   The remote changes to main must first be retrieved so the local environment has the current main state; only then can main be merged into Branch1 and conflicts be detected and resolved. Merging a stale local copy of main could omit recent remote commits, pulling Branch1 does not bring in main's changes, and opening a pull request before updating defers conflict discovery.
6. Q6

   You have a Lakeflow Spark Declarative Pipelines (SDP) pipeline in Azure Databricks that ingests transaction data into a table named Table1. You need to ensure that in the event of an invalid record the pipeline continues to run, invalid records must NOT be written to Table1, invalid records must be preserved for review, and development effort must be minimized. What should you do?

   1. Implement advanced logic to quarantine the invalid records
   2. Define a pipeline expectation
   3. Run WHERE clauses in downstream queries to filter out invalid records
   4. Add a CHECK constraint to Table1

   An SDP expectation such as @dlt.expect_or_drop drops violating rows before they reach the target while the pipeline keeps running, and the built-in event log automatically records the dropped rows for review with no extra code. A CHECK constraint would fail the write and halt the pipeline, WHERE clauses only hide rows at read time, and custom quarantine logic requires extra tables and routing code.
7. Q7

   You have an Azure Databricks workspace that is enabled for Unity Catalog and contains two Delta tables, Table1 and Table2, of the same data type. Table1 has a column Column1 and Table2 has a column Column2. You run: SELECT Column1 FROM Table1 GROUP BY Column1 HAVING COUNT() > 1 INTERSECT SELECT Column2 FROM Table2 GROUP BY Column2 HAVING COUNT() > 1; What occurs when you run the query?

   1. Values appear in either table more than once
   2. Values appear in both tables more than once
   3. Values appear in Table1 more than once
   4. Values appear in Table2 but NOT Table1

   Each subquery's GROUP BY / HAVING COUNT(*) > 1 isolates values that are duplicated within its own table, and INTERSECT then returns only values present in both of those result sets — that is, values duplicated in both tables. Option A describes UNION, option C only the first subquery, and option D describes EXCEPT/MINUS rather than INTERSECT.
8. Q8

   You have an Azure Databricks workspace. You are creating a Lakeflow Spark Declarative Pipelines (SDP) pipeline that scales automatically. You need to configure compute for the pipeline. The solution must minimize operational costs and effort. What should you use?

   1. an all-purpose cluster that uses autoscaling
   2. the existing SQL warehouse
   3. a job cluster that uses autoscaling
   4. a single-node, all-purpose cluster

   A job cluster is created for the pipeline run and terminated when the run finishes, so there is no idle cost between runs, and autoscaling lets it expand and contract with the workload. All-purpose clusters bill at a higher DBU rate and persist, SQL warehouses target interactive BI/SQL, and a single-node cluster cannot scale.
9. Q9

   You have an Azure Databricks workspace that is attached to a Unity Catalog metastore named metastore1 that contains a catalog named catalog1. You need to create a new schema named schema2 that is contained in catalog1 and uses abfss://[email protected]/data as its managed location. Which SQL statement should you execute?

   1. CREATE SCHEMA catalog1.schema2 MANAGED LOCATION 'abfss://[email protected]/data';
   2. CREATE CATALOG schema2 MANAGED LOCATION 'abfss://[email protected]/data';
   3. CREATE SCHEMA catalog1.schema2 LOCATION 'abfss://[email protected]/data';
   4. CREATE SCHEMA catalog1.schema2 WITH DBPROPERTIES (LOCATION='abfss://[email protected]/data');

   The three-part name catalog1.schema2 places the schema in catalog1, and MANAGED LOCATION tells Unity Catalog where to store the data files of managed tables and volumes created in that schema. CREATE CATALOG creates a catalog rather than a schema, LOCATION is used for external locations, and WITH DBPROPERTIES only stores arbitrary metadata without changing storage.
10. Q10

    You have an Azure Databricks workspace that is enabled for Unity Catalog and contains a catalog named Catalog1 with a table named Transactions. The table has columns transaction_id, customer_name, email_address, credit_card_number and transaction_amount. Business analysts must query all rows but must not see full email_address or credit_card_number values; they must see only the text after the @ in each email and only the last four digits of each card number. They must query without errors and the solution must follow least privilege. What should you do?

    1. Grant the analysts SELECT on the Transactions table and implement row-level filters
    2. Grant the analysts SELECT on columns that do NOT contain sensitive data
    3. Grant the analysts SELECT on the Transactions table and apply column masks to email_address and credit_card_number
    4. Grant the analysts SELECT on the Transactions table and apply column-level encryption

    Unity Catalog column masks use SQL UDFs that dynamically transform column values at query time, so the analysts keep SELECT on the whole table while seeing only the permitted partial values. Row filters hide rows rather than redact columns, granting only non-sensitive columns would break queries that reference the masked columns, and column-level encryption is not the redaction mechanism used here.
11. Q11

    You have an Azure Databricks workspace that contains multiple all-purpose clusters. Some clusters remain idle for long periods after users finish their work. You need to reduce compute costs without affecting active workloads. What should you do?

    1. Convert the clusters into job clusters
    2. Use spot instances
    3. Enable autoscaling
    4. Configure automatic termination

    Auto-termination shuts down an all-purpose cluster after a configurable idle period, eliminating idle spend without interrupting running work. Autoscaling adjusts worker count during active use, spot instances reduce VM price but can be reclaimed, and converting to job clusters changes the workload model rather than addressing idle interactive clusters.
12. Q12

    You have an Azure Databricks workspace that uses Unity Catalog. A Lakeflow Spark Declarative Pipelines (SDP) pipeline ingests data into a managed Delta table named Table1. New columns are added to the source data, causing pipeline failures during writes to Table1. You need to prevent the failures and ensure schema changes are detected and handled. What should you do?

    1. Disable schema enforcement for Table1
    2. Use row filters to exclude records that have new columns
    3. Enable schema evolution
    4. Create a separate table for each schema version

    Schema evolution lets an SDP pipeline detect new source columns and add them to the target Delta table on the next run, so writes succeed without manual intervention. Disabling schema enforcement would admit mistyped or corrupt data, row filters would discard valid records, and a table per schema version creates an unmanageable proliferation of objects.
13. Q13

    You have an Azure Databricks workspace that is enabled for Unity Catalog and contains two managed Delta tables, sales.schema1.table1 (current year) and sales.schema1.table2 (historical). You need to load all rows from table1 into table2, preserving any existing data in table2 and minimizing processing effort. Which command should you run?

    1. INSERT INTO sales.schema1.table2 SELECT * FROM sales.schema1.table1;
    2. CREATE TABLE sales.schema1.table2 AS SELECT * FROM sales.schema1.table1;
    3. INSERT OVERWRITE sales.schema1.table2 SELECT * FROM sales.schema1.table1;
    4. CREATE OR REPLACE TABLE sales.schema1.table2 AS SELECT * FROM sales.schema1.table1;

    INSERT INTO appends the new rows to the existing table, preserving its current contents and history with a single statement. INSERT OVERWRITE and CREATE OR REPLACE replace the existing data, and CREATE TABLE would fail because the table already exists.
14. Q14

    You have an Azure Databricks workspace named Workspace1 that contains a lakehouse enabled for Unity Catalog and a connection to a Microsoft SQL Server database named DB1. You need to expose DB1's schemas and tables so they can be queried in Databricks, appear alongside other Unity Catalog objects, and the data is NOT copied into Databricks-managed storage. Solution: You create a Databricks access connector. Does this meet the goal?

    1. Yes
    2. No

    An access connector (with a managed identity) is used to give Unity Catalog access to Azure storage for external locations and storage credentials; it does not federate an external SQL Server database. Exposing DB1 without copying data requires Lakehouse Federation: create a connection, then a foreign catalog that mirrors the external database.

Answer all 139 questions with instant feedback →

Reader reviews 51 ratings

• DP-750 was harder than I expected but these questions prepared me well. The multi-select ones in particular were really close to the exam.

  ★★★★★★ Chloe D.

• Passed the DP-750 exam today! The scenario-based questions here were extremely similar to what I saw on the test.

  ★★★★★ Riley B.

• Got the DP-750 certification after three weeks of self-study with this question bank. Very happy with the purchase.

  ★★★★★★ Victoria K.

• Very realistic DP-750 exam simulation. The timer feature helped me practice pacing before the actual test.

  ★★★★★★ Ellie B.

• I was preparing for the DP-750 exam while juggling a newborn at home. The flexibility of this platform was a lifesaver.

  ★★★★★★ Cooper D.

• Great resource for DP-750. I liked that I could jump straight to specific domains instead of going through everything in order.

  ★★★★★★ Dominic S.

Frequently asked questions

Free study resources

Community-verified analysis of 25 questions from real test-taker discussions — 6 deep analyses and 0 FAQs.

More Microsoft exams