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

 

Question 261

A data engineer needs to create a table from files stored in a Unity Catalog volume. Which statement best describes the benefit of using the volume?

1. It provides governed access to files through Unity Catalog.
2. It automatically converts every file into a managed table.
3. It replaces all SQL warehouse functionality.
4. It provides cluster autoscaling configuration.

Correct Answer: 1

Explanation

Unity Catalog volumes provide a governed way to work with files that do not necessarily fit the traditional table model. They can be used for unstructured and semi-structured data such as JSON, images, documents, and other files. Access to volumes can be controlled through Unity Catalog privileges, allowing organizations to apply centralized governance to file-based data. A volume does not automatically convert every file into a managed table, nor does it replace SQL warehouses or control compute autoscaling. When a data engineering solution needs controlled file access while maintaining centralized governance, volumes provide an appropriate abstraction. This makes them useful for workloads involving file-oriented data processing.

Question 262

A company has multiple Databricks workspaces that need to follow the same compute configuration requirements. Which feature can help standardize cluster creation?

1. Materialized views
2. Cluster policies
3. Delta constraints
4. Watermarks

Correct Answer: 2

Explanation

Cluster policies allow administrators to define approved compute configurations and restrict settings that users can select when creating clusters. Policies can be used to standardize runtime versions, node types, autoscaling limits, and other compute-related parameters. This helps organizations maintain consistent environments and reduce the risk of unsupported or unnecessarily expensive configurations. Materialized views concern query results, Delta constraints address data integrity, and watermarks are used for streaming state management. When multiple teams or workspaces need to follow defined compute standards, administrators can apply appropriate policies to enforce those requirements. Therefore, cluster policies are the relevant feature for standardizing compute configurations.

Question 263

A streaming pipeline receives records continuously and should execute processing as soon as possible whenever new data becomes available. Which trigger configuration is most appropriate?

1. Available-now trigger
2. Once-per-day schedule
3. Continuous manual execution only
4. Static batch execution

Correct Answer: 1

Explanation

An available-now style trigger is useful when a pipeline needs to process all data that is currently available and then stop after catching up. It can be particularly useful for incremental processing workloads where new files or records arrive between executions. Streaming trigger behavior should be selected according to the business requirement, expected arrival pattern, and desired latency. A daily schedule is appropriate for fixed recurring batch workloads but does not provide the same streaming-oriented behavior. Manual execution requires users to start the workload and static batch processing does not continuously monitor new data. Therefore, an available-now trigger can be appropriate when the pipeline should process currently available incremental data promptly.

Question 264

A data engineer wants to give a group permission to create tables inside a specific Unity Catalog schema without granting broad catalog administration rights. Which principle is being applied?

1. Data duplication
2. Least privilege
3. Full administrative access
4. Unrestricted inheritance

Correct Answer: 2

Explanation

The principle of least privilege means granting users or groups only the permissions necessary to perform their assigned tasks. In this scenario, the group needs the ability to create tables within a particular schema, but it does not need broad administrative control over the entire catalog or metastore. Restricting privileges to the required schema scope reduces unnecessary access and simplifies governance. Data duplication is unrelated to authorization, while full administrative access would provide considerably more permissions than required. Unrestricted inheritance is also not a substitute for carefully designed access controls. Therefore, granting only the required schema-level capability demonstrates the principle of least privilege.

Question 265

A pipeline must process records from a source that provides change events such as inserts, updates, and deletes. Which Delta Lake operation is commonly used to apply these changes to a target table?

1. CACHE
2. MERGE
3. COALESCE
4. DESCRIBE DETAIL

Correct Answer: 2

Explanation

The MERGE operation is commonly used to apply change data capture information to a Delta target table. It can match incoming source records against existing target records and perform different actions depending on whether a match exists. For example, an existing record can be updated, a new record can be inserted, and appropriate logic can handle records representing deletions. This makes MERGE useful for maintaining synchronized target datasets from systems that generate change events. CACHE improves repeated data access, COALESCE changes DataFrame partition counts, and DESCRIBE DETAIL returns table metadata. Therefore, MERGE is the relevant Delta Lake operation for applying insert, update, and delete changes.

Question 266

A team wants to store frequently accessed query results so that expensive calculations do not need to be repeated for every request. Which object can provide this behavior?

1. External location
2. Materialized view
3. Storage credential
4. Cluster policy

Correct Answer: 2

Explanation

A materialized view can store the results of a query so that subsequent queries can access the computed results without necessarily repeating the complete underlying computation each time. This can be useful for expensive aggregations or frequently requested analytical results. The exact refresh and maintenance behavior depends on the platform capabilities and configuration. An external location defines access to external cloud storage, while a storage credential provides authentication for that storage. A cluster policy controls compute configuration. These objects serve different purposes and do not primarily provide persisted query results. Therefore, when frequently accessed query-derived results need to be maintained for efficient consumption, a materialized view is an appropriate option.

Question 267

A data engineer wants to make sure that users cannot directly access a cloud storage path unless they have been granted access through Unity Catalog. Which configuration is most appropriate?

1. External location with appropriate privileges
2. Local notebook variables
3. Temporary DataFrame cache
4. Git branch permissions

Correct Answer: 1

Explanation

An external location provides a governed representation of a cloud storage path within Unity Catalog. Administrators can associate the external location with an appropriate storage credential and then grant privileges to users or groups. This allows access to be centrally controlled rather than relying on users embedding storage credentials in notebooks or applications. Local notebook variables do not provide centralized governance, DataFrame caching concerns processing performance, and Git branch permissions control source-code collaboration. External locations therefore provide an important governance boundary for cloud storage access. By combining them with appropriate privileges, organizations can control which identities are allowed to access specific external storage paths.

Question 268

A streaming query performs a stateful aggregation and needs to remove old state after events become sufficiently late. Which feature helps the engine determine when old state can be discarded?

1. Git integration
2. Watermarking
3. Cluster autoscaling
4. External volumes

Correct Answer: 2

Explanation

Watermarking helps Structured Streaming manage state for event-time operations. A watermark indicates how far the system believes event time has progressed and provides a threshold for handling late-arriving data. Once data becomes older than the configured watermark threshold, state associated with sufficiently old events can eventually be removed for supported stateful operations. This prevents state from growing indefinitely in long-running streaming workloads. Git integration manages source-code versioning, cluster autoscaling adjusts compute capacity, and external volumes provide governed file access. None of those capabilities determines the retention period of streaming state. Therefore, watermarking is the appropriate feature for managing old state in an event-time streaming aggregation.

Question 269

A data engineer needs to inspect the structure and metadata of a Delta table, including information about its files and properties. Which command is useful?

1. SHOW GRANTS
2. DESCRIBE DETAIL
3. VACUUM
4. CREATE SCHEMA

Correct Answer: 2

Explanation

DESCRIBE DETAIL provides detailed metadata about a Delta table. It can be useful when engineers need information about the table’s location, format, properties, file-related information, and other metadata supported by the platform. This makes it a useful diagnostic command when investigating table configuration or physical characteristics. SHOW GRANTS focuses on access permissions, VACUUM performs storage cleanup, and CREATE SCHEMA creates a new organizational namespace. These operations do not provide the same table metadata information. Therefore, when the requirement is to inspect detailed metadata associated with a Delta table, DESCRIBE DETAIL is the appropriate command.

Question 270

A team needs to run a notebook every night and automatically execute a downstream validation task after the notebook succeeds. Which Databricks feature should be used?

1. Delta time travel
2. Lakeflow Jobs
3. Column masking
4. External location

Correct Answer: 2

Explanation

Lakeflow Jobs provides workflow orchestration capabilities for scheduling and coordinating Databricks tasks. A job can contain multiple tasks, define dependencies between them, and execute according to a schedule. In this scenario, the notebook can be configured as an initial task and the validation task can depend on its successful completion. This allows the entire process to run automatically each night without manual intervention. Delta time travel addresses historical table versions, column masking protects sensitive values, and external locations govern cloud storage access. Therefore, Lakeflow Jobs is the appropriate capability for scheduling the notebook and coordinating the downstream validation task.

Question 271

A data engineer wants to prevent malformed source records from stopping an ingestion pipeline while preserving those records for later investigation. Which approach is appropriate?

1. Quarantine invalid records
2. Delete all invalid records immediately
3. Disable schema validation
4. Grant ownership to every user

Correct Answer: 1

Explanation

Quarantining invalid records allows a pipeline to continue processing valid data while separating records that fail defined quality or schema checks. The rejected records can then be stored in a dedicated location for investigation, correction, or later reprocessing. This approach improves pipeline resilience because a small number of malformed records does not necessarily stop the entire ingestion process. Simply deleting invalid records can result in data loss, while disabling validation can allow poor-quality information into downstream datasets. Granting ownership to users does not solve data-quality problems. Therefore, a quarantine pattern is appropriate when malformed records need to be isolated without interrupting processing of valid data.

Question 272

A data engineer needs to increase the number of partitions in a DataFrame and redistribute records across the new partitions. Which operation should be considered?

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

Correct Answer: 2

Explanation

The repartition operation can change the number of partitions in a DataFrame and redistribute data across those partitions. Because redistribution generally involves a shuffle, repartition can be more expensive than simply reducing partitions with coalesce. However, repartition can be useful when the existing distribution is unsuitable for a downstream operation or when more parallelism is required. COALESCE is generally used to reduce the number of partitions without a full shuffle. VACUUM manages obsolete Delta files, while MERGE synchronizes records between datasets. Therefore, when a DataFrame needs to be redistributed across a different number of partitions, repartition is the appropriate operation.

Question 273

A company wants to use a cloud storage identity without embedding long-lived secrets directly in Databricks notebooks. Which Unity Catalog feature represents the authentication configuration for external storage?

1. Storage credential
2. Materialized view
3. Row filter
4. Delta constraint

Correct Answer: 1

Explanation

A storage credential represents the authentication information used by Unity Catalog to access supported external cloud storage. It provides a centralized mechanism for managing access rather than requiring engineers to place storage secrets directly in notebooks or application code. External locations can reference storage credentials to connect governed Unity Catalog paths with the appropriate authentication mechanism. Materialized views store query-derived results, row filters restrict visible records, and Delta constraints enforce data conditions. These features do not provide cloud-storage authentication. Therefore, when an organization wants to centralize and govern the identity used for external storage access, a storage credential is the appropriate Unity Catalog component.

Question 274

A query repeatedly filters a large dataset using the same highly selective columns. Which data-layout strategy can help organize the data for efficient filtering?

1. Git folders
2. Liquid clustering
3. Task retries
4. Storage credentials

Correct Answer: 2

Explanation

Liquid clustering is a data-layout capability designed to organize table data around clustering columns and can be useful for workloads with selective query patterns. It can help improve data skipping and reduce unnecessary file scanning when queries frequently filter on suitable columns. The appropriate columns should be selected based on actual workload characteristics rather than arbitrarily clustering every available field. Git folders manage source-code collaboration, task retries improve workflow reliability, and storage credentials govern cloud-storage authentication. These features do not directly organize table data for selective query performance. Therefore, when a large table is repeatedly filtered using suitable columns, liquid clustering can be considered as a data-layout optimization strategy.

Question 275

A workflow task fails because a temporary downstream service becomes unavailable. The team wants the workflow to automatically try the task again before marking the run as failed. Which configuration is required?

1. Retry settings
2. Column masking
3. Delta time travel
4. Volume permissions

Correct Answer: 1

Explanation

Retry settings allow a workflow task to be attempted again after a failure. This is useful for transient problems such as temporary service interruptions, short-lived connectivity issues, or intermittent infrastructure failures. The number of retries and any associated delay should be selected carefully because retries cannot resolve persistent configuration or data-quality problems and may increase resource usage. Column masking controls sensitive data visibility, Delta time travel provides access to historical table versions, and volume permissions govern file access. None of these features controls workflow recovery behavior. Therefore, when a task should automatically be retried after a temporary failure, the workflow’s retry settings are the appropriate configuration.

Question 276

A data engineer wants to inspect which operations were performed on a Delta table and when those operations occurred. Which feature should be used?

1. DESCRIBE HISTORY
2. SHOW COLUMNS
3. CREATE VOLUME
4. COALESCE

Correct Answer: 1

Explanation

DESCRIBE HISTORY provides transaction history for a Delta table and can show information about operations performed against the table. This is useful for auditing, troubleshooting unexpected changes, and understanding the sequence of writes or modifications that produced the current table state. It can also support investigations where engineers need to identify when particular operations occurred. SHOW COLUMNS is used to inspect column information rather than transaction history. CREATE VOLUME creates a governed file-access object, while COALESCE changes DataFrame partitioning. Therefore, when the objective is to review historical Delta table operations and their associated metadata, DESCRIBE HISTORY is the appropriate feature.

Question 277

A data platform team wants analysts to access a table but prevent them from seeing sensitive rows based on their assigned region. Which capability should be configured?

1. Row filter
2. Cluster policy
3. Compute pool
4. Job schedule

Correct Answer: 1

Explanation

A row filter can restrict which records are returned to users based on defined conditions. This supports row-level access control where different users or groups need to see different subsets of the same dataset. For example, a regional organization could use a row-filtering policy so that analysts see records associated only with their authorized region. Cluster policies govern compute configuration, compute pools help reduce compute startup time, and job schedules control when workflows execute. None of these features directly determines which table rows a user can see. Therefore, a row filter is the appropriate governance mechanism for restricting access to specific records according to user context or authorization.

Question 278

A pipeline should write data repeatedly without creating duplicate results if the same processing request is accidentally executed again. Which property should the pipeline design aim to provide?

1. Randomization
2. Idempotency
3. Full duplication
4. Manual intervention

Correct Answer: 2

Explanation

Idempotency means that performing the same logical operation more than once produces the same intended final result rather than creating unintended duplicate effects. This is particularly important in data engineering because jobs can be retried, restarted, or accidentally executed more than once. Pipelines can use techniques such as deterministic keys, MERGE operations, checkpoints, transaction identifiers, or controlled write patterns to achieve idempotent behavior depending on the workload. Randomization does not prevent duplicates, while full duplication intentionally creates repeated results. Manual intervention is not a scalable reliability mechanism. Therefore, designing a pipeline to be idempotent helps ensure safe recovery and repeated execution without introducing unintended duplicate data.

Question 279

A data engineer wants to provide a SQL user with access to a curated dataset while hiding implementation details of several underlying tables. Which object is appropriate?

1. View
2. Storage credential
3. Compute pool
4. Cluster policy

Correct Answer: 1

Explanation

A view can provide users with a logical representation of data without requiring them to directly interact with all underlying tables. A curated view can combine multiple datasets, select required columns, apply business logic, or filter information before exposing the result to consumers. This can simplify reporting and provide a controlled interface between data producers and data consumers. Storage credentials handle authentication to external storage, compute pools provide ready-to-use compute resources, and cluster policies govern compute configuration. None of these objects provides a logical SQL abstraction over underlying datasets. Therefore, a view is appropriate when users should consume curated data without directly managing the underlying implementation.

Question 280

A data engineer needs to identify whether a user has the necessary permissions to access a specific Unity Catalog table. Which information is most relevant?

1. Cluster runtime version
2. Unity Catalog grants
3. DataFrame partition count
4. Streaming trigger interval

Correct Answer: 2

Explanation

Unity Catalog grants determine which users and groups have privileges on governed securable objects such as catalogs, schemas, tables, views, volumes, and other supported resources. When troubleshooting access to a table, administrators should review the applicable grants and inherited permissions to determine whether the user has the required privilege, such as SELECT. Cluster runtime versions, DataFrame partition counts, and streaming trigger intervals affect compute or processing behavior but do not determine whether a user is authorized to access a Unity Catalog table. Therefore, reviewing Unity Catalog grants is the appropriate way to investigate table-access permissions.