CertifyCloudxAce every cloud certification.

DP-750 sample questions with answers

10 free practice questions for the Exam DP-750: Implementing Data Engineering Solutions Using Azure Databricks exam. Try each one, then open the answer to see why the right option wins and every other option loses.

Question 1Set up and configure an Azure Databricks environment

Tailwind Traders triggers roughly 300 short job tasks an hour. Each task waits four to six minutes while VMs are acquired. The workspace must stay on classic compute because the tasks need a custom container image and an init script that reaches a private endpoint. Every run must remain isolated and billed at the jobs rate. You need to cut the start-up wait. What should you do?

  1. A.

    Point every task at a single all-purpose cluster that is kept running continuously.

  2. B.

    Enable Photon acceleration on every job cluster and switch the clusters to a compute-optimized node type so that each short task finishes its work sooner.

  3. C.

    Enable autoscaling on each job cluster and raise the maximum worker count so that additional worker nodes are requested from Azure sooner.

  4. D.

    Create an instance pool with the required node type and a minimum idle instance count, then configure the job clusters to draw nodes from the pool.

Show answer

Answer: D

An instance pool keeps warm, pre-acquired VMs on standby so job clusters can claim them in seconds while still being separate, jobs-rate clusters.

  • A. A shared always-on all-purpose cluster removes isolation between runs and is billed at the higher interactive DBU rate, violating two stated constraints.
  • B. Photon and node family affect execution speed after the cluster is running, not the time spent waiting for Azure to allocate virtual machines.
  • C. Autoscaling adjusts worker count on a cluster that is already up; it cannot shorten the initial VM acquisition and a higher ceiling can lengthen it.
  • D. Pools hold pre-acquired, optionally pre-loaded VMs, so job clusters start in seconds while keeping per-run isolation and jobs-rate billing.
Question 2Set up and configure an Azure Databricks environment

At Trey Research, a data platform engineer must create the catalog researchprod with MANAGED LOCATION set to a subpath of the external location elresearch. The engineer is not a metastore admin. Which two privileges must the engineer hold? (Choose TWO.)

Choose 2.

  1. A.

    CREATE MANAGED STORAGE on el_research

  2. B.

    READ FILES and WRITE FILES on el_research

  3. C.

    CREATE EXTERNAL TABLE on el_research

  4. D.

    CREATE CATALOG on the metastore

  5. E.

    CREATE CONNECTION on the metastore

Show answer

Answer: A, D

Creating any catalog requires CREATE CATALOG on the metastore, and pointing its managed location at an external location requires CREATE MANAGED STORAGE on that location.

  • A. CREATE MANAGED STORAGE on the external location is required to use it as a catalog's managed location.
  • B. READ FILES and WRITE FILES grant direct file access and are not required for managed storage.
  • C. CREATE EXTERNAL TABLE registers external tables in the location; it does not allow managed storage.
  • D. CREATE CATALOG on the metastore is required to create any catalog.
  • E. CREATE CONNECTION is for Lakehouse Federation connections, not for catalogs with managed storage.
Question 3Set up and configure an Azure Databricks environment

Analysts at Alpine Ski House want to save a query that joins lift_scans to passes so that anyone with access can reuse it by name from notebooks and dashboards. Results must always reflect the current table contents, and no data may be stored for it. Which object should they create?

  1. A.

    A streaming table that reads the two tables

  2. B.

    A materialized view refreshed on a schedule

  3. C.

    A view registered in a Unity Catalog schema

  4. D.

    A temporary view created in the notebook

Show answer

Answer: C

A Unity Catalog view stores only the query text, so it processes no data at creation and always returns current results when queried.

  • A. A streaming table stores ingested results; it is for incremental processing, not saved query logic.
  • B. A materialized view stores precomputed results that are only as fresh as the last refresh.
  • C. A view registers only query text, stores no data, and returns current results each time it is queried.
  • D. A temporary view is not registered in a schema and vanishes with the notebook session.
Question 4Set up and configure an Azure Databricks environment

An analyst at Fabrikam holds USE CATALOG, USE SCHEMA, and SELECT on tables in the foreign catalog crmfed but does not own its connection. Which two compute resources can the analyst use to query crmfed? (Choose TWO.)

Choose 2.

  1. A.

    An all-purpose cluster in standard access mode (formerly shared) on Databricks Runtime 15.4 LTS

  2. B.

    A classic SQL warehouse on a current channel version

  3. C.

    A pro SQL warehouse on a current channel version

  4. D.

    An all-purpose cluster in standard access mode on Databricks Runtime 12.2 LTS

  5. E.

    An all-purpose cluster in dedicated access mode (formerly single user), assigned to the analyst, on Databricks Runtime 15.4 LTS

Show answer

Answer: A, C

Federation needs Databricks Runtime 13.3 LTS or above in standard or dedicated mode, or a pro or serverless warehouse; dedicated mode works only for the connection owner.

  • A. Standard access mode on Databricks Runtime 13.3 LTS or above supports federation with normal catalog privileges.
  • B. Classic SQL warehouses are not supported; federation requires pro or serverless.
  • C. Pro (or serverless) SQL warehouses are supported for federated queries.
  • D. Databricks Runtime 12.2 LTS is below the 13.3 LTS minimum for federation.
  • E. On dedicated access mode, only the owner of the connection can query the foreign catalog.
Question 5Set up and configure an Azure Databricks environment

Litware federates an Oracle database whose schemas change most nights. Queries against the foreign catalog run slowly first thing each morning, because the first query after a change triggers a full metadata refresh. You need morning queries to use current, cached metadata. What should you do?

  1. A.

    Recreate the foreign catalog every night with CREATE OR REPLACE so that its metadata is always rebuilt from scratch

  2. B.

    Convert the foreign tables to managed tables with ALTER TABLE SET MANAGED so that no metadata refresh is needed

  3. C.

    Enable the result cache for the SQL warehouse so that federated queries reuse cached query results each morning

  4. D.

    Schedule a Lakeflow job that runs REFRESH FOREIGN CATALOG on the catalog after the nightly change window

Show answer

Answer: D

Learn recommends proactively refreshing foreign catalog metadata with a scheduled REFRESH FOREIGN statement so queries use cached metadata instead of refreshing at run time.

  • A. Recreating the catalog discards its grants and still leaves the first query to fetch metadata.
  • B. SET MANAGED for foreign tables supports only Hive metastore and Glue federation, and it abandons federation.
  • C. Result caching is not supported for federated queries.
  • D. A scheduled REFRESH FOREIGN CATALOG loads current metadata ahead of the morning queries, as Learn recommends.
Question 6Set up and configure an Azure Databricks environment

Yesterday an engineer at Lamna Healthcare accidentally ran DROP TABLE on the managed table claims.silver.encounters. No table with that name has been created since. The table, its privileges, and its properties must be restored. What should you run?

  1. A.

    UNDROP TABLE claims.silver.encounters;

  2. B.

    CREATE TABLE claims.silver.encounters DEEP CLONE claims.silver.encounters;

  3. C.

    ALTER TABLE claims.silver.encounters SET MANAGED;

  4. D.

    RESTORE TABLE claims.silver.encounters TO VERSION AS OF 0;

Show answer

Answer: A

UNDROP TABLE recovers a dropped Unity Catalog managed table within the recovery period, including its privileges and properties.

  • A. UNDROP TABLE restores the dropped managed table and its metadata within the recovery period.
  • B. A clone requires an existing source table, which no longer exists after the drop.
  • C. SET MANAGED converts external tables to managed tables and does not recover drops.
  • D. RESTORE operates on an existing table's history; it cannot recover a dropped table.
Question 7Set up and configure an Azure Databricks environment

Scanned delivery receipts at Wingtip Toys are PDF and JPEG files that only Azure Databricks notebooks and jobs will process. The files must be governed by Unity Catalog grants, and nobody wants to manage cloud paths or credentials for them. Where should the files be stored?

  1. A.

    In an external location granted READ FILES to all users

  2. B.

    In an external table that points at the image folder

  3. C.

    In a managed volume in the appropriate schema

  4. D.

    In the DBFS root, under a folder shared with the team

Show answer

Answer: C

Volumes govern non-tabular files, and a managed volume is the simplest governed option when only Databricks workloads use the files.

  • A. Best practices advise against granting READ FILES on external locations to end users; use volumes instead.
  • B. External tables register tabular data; PDFs and images belong in volumes, not tables.
  • C. A managed volume governs non-tabular files through Unity Catalog with no paths or credentials to manage.
  • D. The DBFS root bypasses Unity Catalog grants and is a legacy storage pattern.
Question 8Set up and configure an Azure Databricks environment

A supplier shares a set of inventory tables with Adventure Works through Databricks-to-Databricks OpenSharing (formerly Delta Sharing). Adventure Works analysts must query the tables in their own workspace with Unity Catalog grants, and no copy of the data may be made. The share already appears in Adventure Works' metastore under the supplier's provider. Which statement should an administrator run?

  1. A.

    CREATE FOREIGN CATALOG supplierinventory USING CONNECTION supplierconn OPTIONS (database 'inventory');

  2. B.

    CREATE CATALOG supplier_inventory MANAGED LOCATION 'abfss://inv@awdata.dfs.core.windows.net/supplier';

  3. C.

    CREATE EXTERNAL VOLUME supplier_inventory LOCATION 'abfss://inv@awdata.dfs.core.windows.net/supplier';

  4. D.

    CREATE CATALOG supplierinventory USING SHARE supplierprovider.inventory_share;

Show answer

Answer: D

A shared catalog created with USING SHARE makes the provider's shared assets readable in the recipient workspace without copying them.

  • A. Foreign catalogs mirror external databases through a connection; they are not created from shares.
  • B. A standard catalog would start empty, so the data would have to be copied in, which is forbidden.
  • C. An external volume governs files at a cloud path and does not expose the shared tables.
  • D. USING SHARE creates a shared catalog that exposes the provider's assets for reading, with no copy.
Question 9Set up and configure an Azure Databricks environment

A data engineer at Best For You Organics works in a notebook on serverless compute and needs a PyPI package that no other notebook should receive. Which two methods install the package for this notebook? Each correct answer presents a complete solution. (Choose TWO.)

Choose 2.

  1. A.

    Install the package as a compute-scoped library from the Libraries tab

  2. B.

    Add the package to the libraries of a compute policy the notebook uses

  3. C.

    Reference the package from a compute-scoped init script stored in a volume

  4. D.

    Run %pip install with the package name in a cell of the notebook

  5. E.

    Add the package as a dependency in the notebook's Environment side pane and apply it

Show answer

Answer: D, E

Serverless notebooks install dependencies through the Environment side pane or notebook-scoped %pip; compute-scoped libraries, init scripts and compute policies are not supported on serverless.

  • A. Serverless compute does not support compute-scoped libraries.
  • B. Serverless compute does not support compute policies.
  • C. Serverless compute does not support compute-scoped init scripts.
  • D. %pip creates a notebook-scoped library that other notebooks do not see.
  • E. The Environment side pane installs dependencies for this serverless notebook only.
Question 10Set up and configure an Azure Databricks environment

A data engineer at Northwind Traders must create a schema named logistics inside the existing catalog ops_prod. The schema will use the catalog's managed storage. Which privileges must the engineer hold?

  1. A.

    CREATE MANAGED STORAGE on the catalog's external location

  2. B.

    CREATE CATALOG on the metastore and USE SCHEMA on ops_prod

  3. C.

    CREATE TABLE and USE SCHEMA on ops_prod

  4. D.

    USE CATALOG and CREATE SCHEMA on ops_prod

Show answer

Answer: D

Creating a schema requires USE CATALOG and CREATE SCHEMA on the parent catalog.

  • A. CREATE MANAGED STORAGE is needed only when the schema gets its own managed location.
  • B. CREATE CATALOG creates catalogs, and USE SCHEMA traverses an existing schema; neither creates a schema.
  • C. CREATE TABLE and USE SCHEMA apply to working inside a schema that already exists.
  • D. USE CATALOG plus CREATE SCHEMA on the parent catalog is the documented requirement.

Keep going with 502 more DP-750 questions

Free papers every day, in the real exam formats, with progress by exam domain. Unlock every paper and timed mock exam when you are ready.

Practise freeSee all DP-750 papers
DP-750 sample questions with answers (10 free) · CertifyCloudx