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

 

Question 241

A data engineer wants to create a schema inside an existing Unity Catalog catalog to organize tables for a specific business domain. Which statement is appropriate?

1. CREATE SCHEMA
2. CREATE VOLUME
3. CREATE WAREHOUSE
4. CREATE STORAGE CREDENTIAL

Correct Answer: 3

Explanation

A schema is a logical container for database objects such as tables, views, and functions within a Unity Catalog catalog. Creating a schema provides an additional organizational level that helps separate datasets by business domain, project, environment, or other logical requirements. The CREATE SCHEMA statement is used to create this container when the user has the required privileges on the parent catalog. CREATE VOLUME creates a file-oriented storage object, while a warehouse provides SQL compute resources. Storage credentials are used to authenticate access to external cloud storage. Therefore, when the requirement is to create a logical namespace inside an existing catalog, creating a schema is the appropriate operation.

Question 242

A pipeline receives JSON files whose schema can change as new fields are introduced. Which Auto Loader capability helps manage these schema changes?

1. Schema evolution
2. Time travel
3. Query caching
4. Cluster autoscaling

Correct Answer: 1

Explanation

Auto Loader supports schema inference and schema evolution capabilities that can help pipelines handle changes in incoming file structures. When source systems add new fields, appropriate schema-evolution settings can allow the ingestion process to recognize and incorporate those changes rather than requiring the pipeline to be completely redesigned. The exact behavior depends on the configured schema evolution mode and source data. Delta time travel is used to access historical table versions, query caching concerns repeated query execution, and cluster autoscaling changes compute capacity. Therefore, when the primary requirement is handling changes in the structure of incoming JSON data, Auto Loader schema evolution is the relevant capability.

Question 243

A company wants to provide a data scientist with access to files stored in a governed Unity Catalog volume. Which permission is primarily relevant for reading the files?

1. MODIFY
2. READ VOLUME
3. OWNERSHIP
4. CREATE CATALOG

Correct Answer: 4

Explanation

Unity Catalog provides privileges specifically for governed volume access. A user who needs to read files from a volume should receive the appropriate read-related volume privilege rather than broader administrative permissions. Granting MODIFY would provide additional capabilities that are unnecessary for a read-only requirement, while ownership provides extensive control over the object. CREATE CATALOG is unrelated to reading files from a volume and would grant an entirely different capability. Applying the appropriate volume-level read permission follows the principle of least privilege and helps organizations control access to unstructured and semi-structured data. Therefore, the read volume privilege is the relevant permission for this requirement.

Question 244

A data engineer needs to write the results of a streaming micro-batch to a target system using custom batch-oriented logic. Which Structured Streaming feature should be used?

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

Correct Answer: 2

Explanation

The foreachBatch feature allows a Structured Streaming query to apply custom logic to each micro-batch as a DataFrame. This is useful when the target system or processing requirement is easier to handle with batch-oriented operations while the source continues to operate as a stream. Engineers can use the provided micro-batch DataFrame to perform transformations, writes, or other supported processing logic. DESCRIBE DETAIL returns metadata about a Delta table, SHOW GRANTS displays permissions, and OPTIMIZE reorganizes Delta table files. Therefore, when custom processing needs to be executed separately for every streaming micro-batch, foreachBatch is the appropriate Structured Streaming capability.

Question 245

An administrator needs to determine who currently has privileges on a Unity Catalog object. Which command can provide this information?

1. SHOW GRANTS
2. DESCRIBE HISTORY
3. OPTIMIZE
4. VACUUM

Correct Answer: 1

Explanation

SHOW GRANTS can be used to inspect permissions associated with securable objects in Unity Catalog. This makes it useful for auditing access and determining which users or groups have been granted particular privileges. Administrators can use this information when troubleshooting access problems, reviewing governance configurations, or validating that permissions follow organizational policies. DESCRIBE HISTORY provides Delta table transaction history rather than permission information. OPTIMIZE reorganizes data files for performance, while VACUUM removes obsolete files according to retention rules. Therefore, when the objective is to inspect grants and privileges associated with a Unity Catalog object, SHOW GRANTS is the appropriate command.

Question 246

A Delta table contains obsolete data files that are no longer required according to the organization’s retention policy. Which operation is designed to remove these files?

1. MERGE
2. VACUUM
3. CACHE
4. REPARTITION

Correct Answer: 2

Explanation

VACUUM is a Delta Lake maintenance operation used to remove obsolete data files that are no longer needed according to the configured retention requirements. Removing unnecessary files can reduce storage consumption, but administrators must carefully consider retention settings because older files may be required for time travel or recovery scenarios. MERGE synchronizes records between source and target datasets, CACHE keeps data available for repeated access, and REPARTITION changes how records are distributed across partitions. None of these operations is designed specifically to remove obsolete Delta files. Therefore, when the requirement is to clean up files that have exceeded the applicable retention period, VACUUM is the appropriate operation.

Question 247

A data engineer wants to ensure that invalid values are rejected when data is written to a Delta table. Which feature can help enforce data integrity conditions?

1. Git folder
2. Delta constraint
3. Compute pool
4. External location

Correct Answer: 2

Explanation

Delta constraints can be used to define data integrity requirements that records must satisfy when being written to a table. They help enforce expected conditions and can prevent invalid data from being accepted when the configured constraint is violated. This is useful for maintaining consistent business rules and improving data quality within curated datasets. Git folders support source control, compute pools improve compute startup behavior, and external locations govern access to cloud storage paths. These features do not directly validate the values being written to a Delta table. Therefore, when the requirement is to enforce conditions on incoming table data, Delta constraints are the relevant capability.

Question 248

A workflow contains an ingestion task followed by a transformation task and then a reporting task. What is the main purpose of defining these task dependencies?

1. To control execution order
2. To change storage credentials
3. To mask columns
4. To create catalogs

Correct Answer: 4

Explanation

Task dependencies define relationships between workflow tasks and determine when downstream tasks are eligible to execute. In the described workflow, the transformation task should normally wait for ingestion to complete, and the reporting task should wait for the transformation task. Dependencies therefore establish the required execution order and help coordinate multi-step data pipelines. They can also support conditional execution based on task outcomes. Storage credentials control access to external storage, column masking protects sensitive values, and catalog creation is a governance and organizational operation. Therefore, the primary purpose of task dependencies in this workflow is to control how tasks execute relative to one another.

Question 249

A data engineer needs to reduce data movement when filtering a large dataset by a highly selective predicate. Which technique can help improve query efficiency?

1. Predicate pushdown
2. CREATE VOLUME
3. Job retry
4. Table ownership

Correct Answer: 3

Explanation

Predicate pushdown can improve query efficiency by allowing filtering conditions to be applied closer to the data source or scan operation. When a predicate is highly selective, reducing the amount of data read and processed can lower I/O and computation requirements. This is particularly valuable when working with large datasets where only a small subset of records is needed. CREATE VOLUME creates a governed file-access object, job retry improves workflow resilience, and table ownership controls administrative permissions. These features do not directly optimize filtering operations. Therefore, when the objective is to reduce unnecessary data processing for selective queries, predicate pushdown is an appropriate optimization technique.

Question 250

A data engineer needs to store a dataset in cloud storage outside the default managed storage location while still registering it in Unity Catalog. Which table type is appropriate?

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

Correct Answer: 2

Explanation

An external table references data stored at a location outside the default managed storage area. Unity Catalog can register and govern the table while the underlying data remains in an externally specified cloud storage path. This approach is useful when organizations already have data in Azure Data Lake Storage or another supported external location and want centralized governance without moving the data into managed storage. A managed table uses storage managed by the platform, while a temporary view is not a persistent physical table. A materialized view stores query-derived results for supported workloads. Therefore, when the underlying data must remain at an external storage location, an external table is appropriate.

Question 251

A streaming application maintains large amounts of state because records can arrive late. Which feature helps limit how long old event-time state is retained?

1. Watermark
2. Git branch
3. Storage credential
4. Cluster policy

Correct Answer: 1

Explanation

A watermark helps manage state in Structured Streaming by indicating how far event time has progressed and how long late-arriving records should continue to be considered. For stateful operations such as aggregations, this allows the engine to eventually remove state associated with sufficiently old data. Without appropriate state-management strategies, a long-running streaming application can accumulate increasing amounts of state and consume significant resources. Git branches manage source-code versions, storage credentials provide authentication for cloud storage, and cluster policies govern compute configuration. Therefore, when the requirement is to control the retention of old event-time state in a streaming application, watermarking is the relevant feature.

Question 252

A company wants to separate raw source data from cleaned and business-ready datasets. Which layer normally contains cleaned and validated data?

1. Gold
2. Bronze
3. Silver
4. Raw archive

Correct Answer: 3

Explanation

In the commonly used medallion architecture, the Silver layer contains cleaned, validated, standardized, and transformed data derived from the raw Bronze layer. Bronze generally preserves source information with minimal transformation, while Silver improves data quality and consistency for downstream consumers. Gold contains curated datasets that are typically optimized for business reporting, analytics, and specific use cases. This layered structure provides clear processing boundaries and helps teams trace how data moves from ingestion toward consumption. Therefore, when a pipeline needs a layer specifically for cleaned and validated datasets before business-oriented aggregation, the Silver layer is the appropriate choice.

Question 253

A data engineer needs to determine whether a Delta table has recently undergone schema or data changes. Which command can provide transaction history?

1. DESCRIBE HISTORY
2. SHOW GRANTS
3. CREATE TABLE
4. REPARTITION

Correct Answer: 4

Explanation

DESCRIBE HISTORY provides information about the transaction history of a Delta table. It can help engineers investigate operations that affected the table, including writes, updates, merges, and other supported transactions. This information is valuable for troubleshooting unexpected changes, auditing activity, and understanding how a table reached its current state. SHOW GRANTS is focused on permissions, CREATE TABLE creates a new table, and REPARTITION changes data distribution in a processing operation. Therefore, when the objective is to inspect the historical sequence of operations performed against a Delta table, DESCRIBE HISTORY is the appropriate command.

Question 254

A team wants to avoid repeatedly entering cloud storage credentials in notebooks when accessing governed external data. Which Unity Catalog feature should be used with external locations?

1. Storage credential
2. Temporary view
3. Query profile
4. Materialized view

Correct Answer: 1

Explanation

A storage credential provides a governed representation of the authentication information required to access external cloud storage. External locations can reference these credentials so that access to defined storage paths can be controlled centrally through Unity Catalog. This reduces the need to embed or repeatedly provide storage credentials in notebooks and application code. Temporary views provide session-level logical data representations, query profiles help analyze workload execution, and materialized views store query-derived results. By combining storage credentials with external locations, administrators can establish a structured and governed approach to external storage access. Therefore, storage credentials are the appropriate feature for centralized authentication in this scenario.

Question 255

A Databricks workflow should run every day at a specific time without requiring a user to start it manually. Which configuration should be used?

1. Job schedule
2. Delta constraint
3. Column mask
4. Volume permission

Correct Answer: 4

Explanation

A job schedule allows a Databricks workflow to execute automatically according to a defined timetable. Scheduled workflows are commonly used for recurring ingestion, transformation, validation, and reporting workloads. A schedule can define when a job should start and can be combined with other workflow settings such as task dependencies and retry policies. Delta constraints enforce data conditions, column masks protect sensitive values, and volume permissions govern access to files. These features do not determine when a workflow runs. Therefore, when a job needs to execute automatically every day at a specified time, configuring a job schedule is the appropriate solution.

Question 256

A query scans a large Delta table but only needs records from one specific partition value. Which storage design can help reduce the amount of data scanned?

1. Git integration
2. Partitioning
3. Job retries
4. Storage credentials

Correct Answer: 2

Explanation

Partitioning can organize table data according to the values of a selected column or columns. When a query includes a filter that corresponds to a partitioning column, the engine may be able to eliminate partitions that cannot contain matching records. This can reduce the amount of data that needs to be read and processed. Partitioning should be selected carefully because excessive or poorly chosen partitions can introduce overhead and may not improve performance. Git integration manages source control, job retries improve workflow resilience, and storage credentials provide cloud authentication. Therefore, when queries frequently filter on a suitable partition column, partitioning can help reduce unnecessary data scanning.

Question 257

A data engineer wants to identify expensive stages and operations in a Spark workload after observing slow query execution. Which tool or capability should be examined?

1. Query profile
2. Storage credential
3. Unity Catalog catalog
4. Volume permission

Correct Answer: 1

Explanation

The query profile provides execution information that can help engineers understand how a workload was processed. It can be useful for identifying expensive operations, stages, scans, joins, shuffles, and other execution characteristics that contribute to slow performance. This information can guide optimization decisions such as improving filtering, changing data layout, reviewing joins, or addressing excessive data movement. Storage credentials handle authentication and do not explain query execution. Unity Catalog catalogs organize governed objects, while volume permissions control file access. Therefore, when the goal is to investigate where query execution time is being spent, the query profile is a relevant diagnostic capability.

Question 258

A data engineer wants a pipeline to process only newly arriving files instead of repeatedly processing all files already present in a cloud storage directory. Which feature is designed for this scenario?

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

Correct Answer: 3

Explanation

Auto Loader is designed for incremental ingestion of files arriving in supported cloud storage locations. It maintains information about discovered files so that a pipeline can focus on new arrivals instead of repeatedly treating the entire directory as new input. Auto Loader is particularly useful for scalable ingestion pipelines where files arrive continuously or in frequent batches. VACUUM removes obsolete Delta files, OPTIMIZE improves Delta file organization, and DESCRIBE HISTORY provides transaction information. None of those operations provides the primary mechanism for discovering and incrementally ingesting newly arrived source files. Therefore, Auto Loader is the appropriate feature for this ingestion requirement.

Question 259

A data platform team needs to ensure that analysts cannot see the actual values of a sensitive national identification column unless they meet an authorization condition. Which capability is most appropriate?

1. Repartitioning
2. Column masking
3. Job scheduling
4. File compaction

Correct Answer: 2

Explanation

Column masking provides a mechanism for protecting sensitive values while still allowing users to access the broader dataset. A masking rule can return protected or transformed values for users who do not satisfy the required access conditions, while authorized users may receive the original value according to the defined policy. This is useful for sensitive identifiers, confidential business attributes, and other information that should not be exposed universally. Repartitioning affects data distribution, job scheduling controls when workflows execute, and file compaction addresses storage organization. These features do not directly control whether a user can see sensitive column values. Therefore, column masking is the appropriate capability for this requirement.

Question 260

A data engineer needs to make a DataFrame available for repeated operations during the same workload to avoid recomputing the same transformations. Which approach can help?

1. Cache or persist the DataFrame
2. Create a storage credential
3. Apply a row filter
4. Create an external location

Correct Answer: 1

Explanation

Caching or persisting a DataFrame can keep computed data available for reuse during a workload. This can be beneficial when the same transformed dataset is referenced repeatedly and recomputing the transformations would otherwise consume additional processing resources. The actual benefit depends on the dataset size, workload pattern, available memory, and persistence strategy. Storage credentials and external locations address governed access to cloud storage, while row filters control which records users can access. These features do not directly reduce repeated computation of an already-created DataFrame. Therefore, when the same transformed DataFrame will be reused multiple times during a workload, caching or persistence can be an appropriate optimization.