View Full Microsoft DP-750 Exam Dumps and Practice Test Dumps.

 

Question 21

A data engineer needs to grant a group permission to create tables in an existing Unity Catalog schema without allowing the group to create new schemas. Which privilege should be granted?

1. Create Schema
2. Create Table
3. Use Catalog
4. Ownership

Correct Answer: 2

Explanation

The Create Table privilege allows a principal to create tables within a Unity Catalog schema when the required parent permissions are available. Create Schema is broader because it allows the principal to create schemas within a catalog. Use Catalog provides access to the catalog namespace but does not by itself allow table creation. Ownership provides extensive administrative control and is unnecessary for this requirement. Granting only Create Table follows the principle of least privilege by providing the specific capability required by the group without giving additional permissions. This approach helps maintain stronger governance and reduces the possibility of unauthorized structural changes within the catalog.

Question 22

A company stores data in Azure Data Lake Storage that must remain outside the Databricks-managed storage location. Which table type should the data engineer use?

1. Managed table
2. Temporary view
3. External table
4. Materialized view

Correct Answer: 3

Explanation

An external table is appropriate when the underlying data must remain in a storage location managed outside the default Databricks-managed storage area. The table metadata can be registered and governed through Unity Catalog while the actual data remains in the specified external storage path. Managed tables are generally associated with storage locations managed by Databricks. Temporary views do not provide persistent physical table storage, while materialized views are designed to maintain query results for analytical workloads. External tables are therefore useful when organizations need centralized metadata and governance while retaining control over the physical location and lifecycle of the underlying cloud data.

Question 23

A data engineer needs to allow analysts to query tables in a Unity Catalog schema but prevent them from modifying the data. Which permission model should be used?

1. Grant SELECT on the required tables or appropriate schema
2. Grant MODIFY on the schema
3. Grant OWNERSHIP on the catalog
4. Grant CREATE SCHEMA on the catalog

Correct Answer: 1

Explanation

The SELECT privilege provides read access to table data without granting users the ability to modify the underlying records. This is appropriate when analysts need to perform queries, reporting, and analysis but should not insert, update, or delete production data. MODIFY provides write capabilities and would give more access than required. OWNERSHIP provides extensive control over the object and should not be granted simply for analytical access. CREATE SCHEMA allows users to create schemas and is unrelated to reading table data. Using SELECT supports least-privilege security and helps separate analytical responsibilities from data-management and administration tasks.

Question 24

A data engineer wants to reference a table in another Unity Catalog schema while writing SQL. Which object naming convention should be used?

1. table
2. schema.table
3. catalog.schema.table
4. workspace.catalog.table

Correct Answer: 3

Explanation

Unity Catalog uses a three-level namespace for identifying securable data objects. The standard fully qualified naming structure is catalog.schema.object, such as catalog.schema.table. Using the complete three-level name makes the target object explicit and is especially useful when queries access multiple schemas or catalogs. Referencing only a table name depends on the current namespace and may create ambiguity. The workspace is not included as a standard level in a Unity Catalog table identifier. Fully qualified names therefore make SQL code easier to understand, maintain, and deploy across environments where the active catalog or schema context may differ.

Question 25

A data engineer wants a table to automatically reflect changes to its source query result while providing faster access for repeated analytical queries. Which Databricks object is most appropriate?

1. Temporary view
2. Materialized view
3. External volume
4. Secret scope

Correct Answer: 2

Explanation

A materialized view maintains the results of a query so that downstream users can access derived data without recalculating the complete query each time. This can improve performance for repeated analytical workloads, particularly when the underlying transformation involves expensive joins, aggregations, or filtering. A temporary view only provides a logical query representation and does not maintain persisted query results. An external volume is intended for governed file access, while a secret scope is used to manage sensitive credentials. Materialized views therefore provide a useful option when the goal is to maintain derived results and improve performance for recurring analytical queries.

Question 26

A data engineer needs to ingest newly arriving files from cloud object storage continuously without repeatedly listing the entire directory. Which Databricks feature is designed for this scenario?

1. Auto Loader
2. VACUUM
3. OPTIMIZE
4. DESCRIBE HISTORY

Correct Answer: 1

Explanation

Auto Loader is designed for incremental ingestion of newly arriving files from supported cloud storage systems. It can efficiently identify and process new files without requiring the pipeline to repeatedly scan the entire directory for every execution. This makes it useful for large-scale ingestion workloads where files arrive continuously. VACUUM removes obsolete Delta files according to retention settings, while OPTIMIZE improves the physical organization of Delta data. DESCRIBE HISTORY provides information about previous Delta table operations. Auto Loader is therefore the appropriate feature when the primary requirement is reliable and scalable ingestion of newly arriving cloud files.

Question 27

A streaming pipeline must process records continuously as new data arrives and write the results to a Delta table. Which Databricks capability should be used?

1. Batch SQL only
2. Unity Catalog volumes
3. Structured Streaming
4. Cluster policies

Correct Answer: 3

Explanation

Structured Streaming provides a framework for processing continuously arriving data using Spark DataFrame and SQL-based operations. A streaming query can read from supported streaming sources, apply transformations, and write processed results to destinations such as Delta tables. Batch SQL is primarily designed for finite datasets and scheduled processing. Unity Catalog volumes provide governed access to files but do not themselves perform continuous stream processing. Cluster policies control compute configuration and governance rather than data processing. Structured Streaming is therefore appropriate when the pipeline must continuously process new records and maintain results as data arrives.

Question 28

A data engineering team organizes its platform into Bronze, Silver, and Gold layers. What is the primary purpose of this architecture?

1. To separate raw, refined, and business-ready data
2. To replace Unity Catalog
3. To eliminate all data validation
4. To ensure every dataset uses CSV format

Correct Answer: 1

Explanation

The Bronze, Silver, and Gold architecture organizes data according to different levels of refinement. Bronze commonly contains raw or lightly processed source data. Silver contains cleaned, validated, and transformed data that is suitable for further processing. Gold contains curated datasets designed for reporting, analytics, and business consumption. This architecture does not replace Unity Catalog, eliminate data validation, or require CSV as the storage format. Instead, it provides a structured approach for managing data throughout its lifecycle. Data engineers can apply different quality rules, transformations, and access requirements at each layer according to organizational needs.

Question 29

A data engineer receives duplicate records during ingestion and needs to remove duplicate rows based on a business key. Which Spark operation is appropriate?

1. dropDuplicates
2. repartition
3. cache
4. coalesce

Correct Answer: 1

Explanation

The dropDuplicates operation removes duplicate records from a Spark DataFrame. It can operate across all columns or use a specified subset of columns to determine which records are considered duplicates. When a business key uniquely identifies an entity, specifying that key can help remove repeated records according to the required business logic. Repartition changes how records are distributed across Spark partitions but does not remove duplicates. Cache stores data for reuse, while coalesce can reduce the number of partitions. Therefore, dropDuplicates is the operation that directly addresses the requirement to eliminate duplicate records based on a business key.

Question 30

A pipeline receives records repeatedly and must ensure that the same business event is not processed more than once. Which design principle is most important?

1. Random partitioning
2. Idempotent processing
3. Larger driver memory
4. Disabling checkpoints

Correct Answer: 2

Explanation

Idempotent processing means that processing the same input multiple times does not produce an incorrect final result. This principle is important in production data pipelines because retries, duplicate deliveries, and recovery processes can cause the same event to be encountered more than once. Engineers can support idempotency through stable business keys, deduplication, merge operations, and carefully designed transaction logic. Random partitioning does not prevent duplicate processing. Increasing driver memory addresses resource capacity rather than data correctness. Disabling checkpoints can make streaming recovery more difficult. Idempotent pipeline design therefore helps maintain consistent results even when input records are delivered repeatedly.

Question 31

A data engineer needs a pipeline to run every morning automatically. Which Databricks capability should be configured?

1. Unity Catalog volume
2. DataFrame cache
3. Databricks Workflow job schedule
4. SQL temporary view

Correct Answer: 3

Explanation

A Databricks job can be configured with a schedule to execute automatically at a specified time or recurring interval. This capability is useful for recurring ingestion, transformation, validation, and reporting workloads. A Unity Catalog volume provides governed access to files but does not schedule pipelines. DataFrame caching is a performance technique and does not initiate job execution. A temporary view provides a logical SQL representation and is not an orchestration mechanism. Scheduled jobs can also include task dependencies, retries, notifications, and monitoring, making them suitable for production workloads that must execute automatically according to a defined timetable.

Question 32

A Databricks job contains several tasks, and the second task should run only after the first task completes successfully. What should the engineer configure?

1. A task dependency
2. A temporary table
3. A volume mount
4. A cluster policy

Correct Answer: 1

Explanation

A task dependency defines the relationship between tasks in a Databricks workflow and can ensure that one task runs after another has completed successfully. Dependencies are important when downstream processing requires data or output generated by an upstream task. A temporary table can store intermediate data but does not establish workflow execution order. A volume mount provides access to files, while a cluster policy controls compute configuration and governance. By configuring a dependency, engineers can create a predictable execution sequence and reduce the risk of downstream tasks starting before the required upstream processing has completed.

Question 33

A production job occasionally fails because of temporary network or service issues. Which job configuration can help automatically recover from these transient failures?

1. Removing all dependencies
2. Increasing table permissions
3. Configuring task retries
4. Disabling job monitoring

Correct Answer: 3

Explanation

Task retries allow a Databricks workflow to automatically attempt a failed task again. This can help recover from temporary failures such as short-lived connectivity problems, service interruptions, or other transient conditions. Retries should be configured carefully because they cannot correct deterministic application errors and excessive retries may delay detection of a genuine failure. Removing task dependencies can create incorrect execution order, while increasing table permissions does not address transient infrastructure issues. Disabling monitoring would reduce operational visibility. Appropriate retry settings can improve pipeline resilience while allowing temporary failures to recover without requiring immediate manual intervention.

Question 34

A development team wants to track notebook and source-code changes and collaborate using a Git-based workflow. Which Databricks capability should be used?

1. Git folders
2. VACUUM
3. Photon
4. Delta time travel

Correct Answer: 1

Explanation

Git folders provide integration between Databricks development environments and supported Git repositories. They allow teams to synchronize notebooks, source files, and other supported content with version-control workflows. This makes it easier to collaborate, review changes, maintain branches, and manage source code across development activities. VACUUM is used for removing obsolete Delta data files, Photon is an execution engine designed to accelerate supported workloads, and Delta time travel provides access to historical table versions. None of those features provides source-code version control. Git folders are therefore appropriate for teams that want collaborative development and version tracking.

Question 35

A data engineer wants to identify which rows changed between two versions of a Delta table before performing downstream processing. Which capability can provide historical table versions for comparison?

1. Cluster pools
2. Delta time travel
3. Secret scopes
4. SQL warehouses

Correct Answer: 2

Explanation

Delta time travel allows users to access previous versions of a Delta table using supported version or timestamp references, provided the required historical data has not been removed by retention and cleanup processes. This capability can support auditing, troubleshooting, validation, and comparison of historical data states. Cluster pools help reduce compute startup delays, secret scopes help manage sensitive credentials, and SQL warehouses provide compute resources for SQL workloads. Time travel is particularly useful when engineers need to understand how a dataset changed over time. Retention settings should always be considered when historical versions are required for operational or compliance purposes.

Question 36

A data engineer wants to improve query performance by reorganizing small files and optimizing the physical layout of a Delta table. Which operation is designed for this purpose?

1. GRANT
2. SHOW GRANTS
3. DESCRIBE DETAIL
4. OPTIMIZE

Correct Answer: 4

Explanation

The OPTIMIZE operation is designed to improve the physical organization of Delta table data. One important use is compacting small files so that query engines can process data more efficiently. Depending on the table design and supported features, optimization can also improve physical data layout for common access patterns. GRANT changes permissions, SHOW GRANTS displays existing access assignments, and DESCRIBE DETAIL returns metadata about the table. OPTIMIZE should be used based on actual workload characteristics because optimization consumes compute resources. Engineers should evaluate table size, file distribution, query patterns, and maintenance requirements before scheduling optimization operations.

Question 37

A pipeline should prevent unauthorized users from accessing sensitive columns while still allowing them to query the rest of a table. Which approach is appropriate?

1. Apply column-level access controls where supported
2. Grant ownership to every user
3. Store the table as CSV
4. Disable Unity Catalog

Correct Answer: 1

Explanation

Column-level access controls can restrict access to sensitive attributes while allowing authorized users to query other permitted data. This supports least-privilege security and is useful for datasets containing confidential or regulated information. Granting ownership to every user would provide excessive permissions and weaken governance. Changing the storage format to CSV does not provide access-control functionality, and disabling Unity Catalog would remove centralized governance capabilities rather than improve security. Data engineers should define access according to business and security requirements and combine appropriate Unity Catalog controls with organizational policies to protect sensitive information while maintaining useful access for authorized analytical users.

Question 38

A data engineer wants to store application credentials securely rather than placing passwords directly in notebooks. Which capability is intended for managing secrets?

1. Delta Live Tables
2. Secret management
3. Photon
4. Auto Loader

Correct Answer: 2

Explanation

Secret management provides a mechanism for storing sensitive values such as passwords, tokens, and other credentials separately from notebook source code. Applications and notebooks can retrieve these values through supported secret-management mechanisms instead of embedding them directly in scripts. This reduces the risk of exposing credentials through source control, shared notebooks, or copied code. Delta Live Tables focuses on declarative data pipeline processing, Photon focuses on query execution performance, and Auto Loader provides incremental file ingestion. Separating credentials from application logic is an important security practice and should be combined with appropriate access controls and secret rotation procedures.

Question 39

A data engineer needs to inspect the metadata and configuration details of a Delta table, including information about its storage and properties. Which SQL command is appropriate?

1. DESCRIBE DETAIL
2. DELETE
3. INSERT
4. MERGE

Correct Answer: 1

Explanation

DESCRIBE DETAIL provides detailed metadata about a Delta table, including information such as its location, format, properties, and other table-level characteristics. This makes it useful when engineers need to inspect table configuration or investigate storage-related details. DELETE removes records that meet a specified condition, INSERT adds records to a table, and MERGE performs conditional synchronization between source and target datasets. These commands are primarily related to data manipulation rather than metadata inspection. Using DESCRIBE DETAIL can therefore help data engineers understand the current configuration and physical characteristics of a Delta table during troubleshooting and operational analysis.

Question 40

A data engineer wants to combine records from a customer table with matching records from an orders table while retaining only customers that have at least one matching order. Which JOIN should be used?

1. LEFT JOIN
2. INNER JOIN
3. FULL OUTER JOIN
4. CROSS JOIN

Correct Answer: 2

Explanation

An INNER JOIN returns records where the specified join condition has a matching row in both tables. In this scenario, the requirement is to retain only customers who have at least one corresponding order, so an INNER JOIN satisfies the requirement. A LEFT JOIN would also retain customers without matching orders, which is not desired. A FULL OUTER JOIN would preserve unmatched records from both datasets, while a CROSS JOIN creates combinations between rows and is not appropriate for matching customer and order records. Selecting the correct join type helps ensure that the resulting dataset contains exactly the records required by the business logic.