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

 

Question 361

A data engineer needs to ensure that a workflow retries a failed task automatically before marking the task as unsuccessful. Which configuration should be used?

1. Job retry policy
2. Column mask
3. External location
4. Table constraint

Correct Answer: 1

Explanation

A job retry policy allows a workflow task to be attempted again when an execution fails. Retries are useful for handling temporary failures such as transient infrastructure issues, network interruptions, or temporary service unavailability. The number of retries and related behavior should be selected according to the workload and failure characteristics. A column mask protects sensitive column values, an external location governs access to external storage paths, and a table constraint enforces data-quality rules. None of these features controls automatic task retries. Therefore, when a failed workflow task should automatically be attempted again, configuring a job retry policy is the appropriate approach.

Question 362

A data engineer is creating a table from files stored in cloud object storage and wants the resulting table to reference those files rather than fully manage the underlying data. Which table type is appropriate?

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

Correct Answer: 2

Explanation

An external table references data stored at a specified external storage location while allowing the table metadata to be governed through Unity Catalog. This approach is useful when the data lifecycle needs to remain associated with an existing cloud-storage location or when multiple systems may interact with the underlying files. A managed table generally has its storage lifecycle managed by the platform according to the table configuration. Temporary views are session-scoped logical objects, while materialized views contain persisted query results. Therefore, when the table should reference files that remain in an externally managed storage location, an external table is appropriate.

Question 363

A streaming pipeline should process records based on event time rather than the time at which the records reach the processing system. Which concept should the engineer use?

1. Cluster autoscaling
2. Event-time processing
3. Storage credentials
4. Git folders

Correct Answer: 2

Explanation

Event-time processing uses the timestamp associated with the actual event rather than relying only on the time when the event is received by the processing system. This is important when network delays, buffering, or source-system behavior can cause records to arrive out of order. Combined with watermarks and suitable stateful operations, event-time processing helps streaming applications handle late-arriving records more effectively. Cluster autoscaling manages compute capacity, storage credentials authenticate access to cloud storage, and Git folders support source-code management. Therefore, when business logic depends on when an event actually occurred, event-time processing is the appropriate concept.

Question 364

A team wants to make a data pipeline reusable across development, test, and production by changing configuration values without modifying the pipeline code. Which feature is most suitable?

1. Job parameters
2. VACUUM
3. Delta constraints
4. Data skipping

Correct Answer: 1

Explanation

Job parameters allow runtime values to be supplied to workflow tasks without changing the underlying pipeline logic. For example, the same pipeline can receive different environment names, input paths, or processing dates depending on where it is executed. This improves reusability and reduces the need to maintain separate copies of the same workflow for different environments. VACUUM manages obsolete Delta files, Delta constraints enforce data-quality conditions, and data skipping improves query efficiency by avoiding irrelevant files. Therefore, when configuration values need to change while the core pipeline code remains reusable, job parameters are an appropriate solution.

Question 365

A data engineer wants to identify when a Delta table was created, modified, or updated and inspect the operations performed against it. Which feature should be used?

1. Schema evolution
2. DESCRIBE HISTORY
3. Cluster policy
4. Volume

Correct Answer: 2

Explanation

DESCRIBE HISTORY provides information about operations performed on a Delta table. It can help engineers investigate changes such as inserts, updates, deletes, merges, and other table transactions. This historical information is useful for auditing, troubleshooting, and understanding the sequence of changes that produced the current table state. Schema evolution concerns changes to incoming or table schemas, cluster policies control compute configurations, and volumes provide governed file storage. None of these features is designed to display Delta transaction history. Therefore, when an engineer needs to investigate operations and changes recorded for a Delta table, DESCRIBE HISTORY is the appropriate feature.

Question 366

A pipeline receives malformed records that cannot be parsed according to the expected schema. The engineer wants to retain the problematic data for later investigation. Which approach is useful?

1. Increase cluster size
2. Use a rescued data column
3. Disable schema validation
4. Delete malformed files

Correct Answer: 2

Explanation

A rescued data column can preserve data that does not fit the expected schema during supported ingestion scenarios. Instead of simply losing unexpected fields or making the entire ingestion process unusable, the pipeline can retain problematic or unexpected information for later examination. This is particularly useful when source systems are inconsistent or their schemas evolve unexpectedly. Increasing cluster size does not address malformed data, disabling schema validation can allow quality problems to pass unnoticed, and deleting malformed files removes potentially useful information. Therefore, when unexpected source fields or parsing issues need to be retained for investigation, a rescued data column can be an effective ingestion strategy.

Question 367

A data engineer needs to give a user permission to query a table but does not want to grant permission to modify the table. Which privilege is appropriate?

1. SELECT
2. MODIFY
3. OWN
4. CREATE

Correct Answer: 1

Explanation

The SELECT privilege allows a user to read or query data from an object without granting the ability to modify its contents. This follows the principle of granting users only the permissions required for their responsibilities. MODIFY is associated with changing data, OWN provides broader administrative control, and CREATE privileges relate to creating supported objects. Granting unnecessary modification or ownership privileges could provide more access than required. Therefore, when a user only needs to query a table and should not modify its data, SELECT is the appropriate privilege to grant.

Question 368

A team wants to automatically discover and ingest new files arriving in cloud storage without manually maintaining a list of each file. Which technology is designed for this purpose?

1. Unity Catalog views
2. Auto Loader
3. SQL warehouse
4. Delta constraints

Correct Answer: 2

Explanation

Auto Loader is designed for incremental ingestion of files arriving in cloud object storage. It can automatically discover new files and process them as they become available, making it well suited for continuously arriving data. Auto Loader also provides capabilities related to schema inference and evolution, which can help simplify ingestion pipelines when source data changes over time. Unity Catalog views provide logical data representations, SQL warehouses provide SQL compute, and Delta constraints enforce data-quality conditions. These features do not specialize in discovering newly arriving files. Therefore, when a pipeline needs automated incremental file ingestion from cloud storage, Auto Loader is the appropriate technology.

Question 369

A data engineer wants to ensure that users cannot access sensitive values in a particular column while still allowing them to query the table. Which Unity Catalog capability is appropriate?

1. Row filtering
2. Column masking
3. Autoscaling
4. Checkpointing

Correct Answer: 2

Explanation

Column masking can protect sensitive values by applying rules that determine what users are allowed to see when querying a protected column. This allows the table to remain queryable while reducing exposure of sensitive information to users who should not see the original values. Row filtering serves a different purpose by restricting which rows are visible. Autoscaling adjusts compute resources based on workload, and checkpointing stores streaming progress and state. Therefore, when the requirement is to protect the values of a specific column while still allowing authorized querying of the table, column masking is the appropriate governance capability.

Question 370

A data engineer needs to make a SQL query run faster by reducing the amount of data that must be read from storage when filtering on selective conditions. Which technique can help?

1. Predicate pushdown
2. User provisioning
3. Job retries
4. Secret rotation

Correct Answer: 1

Explanation

Predicate pushdown improves query efficiency by applying filtering conditions as close to the data source as practical. This can reduce the amount of data that needs to be transferred or processed by later stages of the query. When only a subset of records satisfies a filter, avoiding unnecessary reads can improve performance and reduce resource consumption. User provisioning manages identities, job retries handle failed workflow executions, and secret rotation manages credentials. These features do not directly optimize query filtering. Therefore, when the goal is to reduce unnecessary data processing by applying filters efficiently, predicate pushdown is a relevant optimization technique.

Question 371

A data engineer wants to organize a lakehouse pipeline into raw, refined, and business-ready layers. Which architecture pattern represents this approach?

1. Star schema
2. Medallion architecture
3. Point-to-point architecture
4. Master-slave architecture

Correct Answer: 2

Explanation

Medallion architecture organizes data into progressive quality and refinement layers, commonly represented as bronze, silver, and gold. The bronze layer generally contains raw or minimally processed data, the silver layer contains cleaned and transformed data, and the gold layer provides business-ready datasets for analytics and reporting. This layered design helps separate ingestion from data-quality processing and business transformations. A star schema is a dimensional modeling pattern, while point-to-point and master-slave descriptions do not represent the standard lakehouse layering approach. Therefore, when a pipeline uses raw, refined, and business-ready data layers, the medallion architecture pattern is appropriate.

Question 372

A streaming application should stop considering very old late-arriving records after a defined amount of event-time delay. Which configuration is relevant?

1. Watermark threshold
2. Cluster policy
3. Storage credential
4. Git repository

Correct Answer: 1

Explanation

A watermark threshold defines the amount of event-time lateness that a streaming application is prepared to handle for supported stateful operations. As the watermark advances, records that are sufficiently older than the watermark can be considered too late for certain operations, which also allows related state to be cleaned up. This helps control state growth while still providing a defined tolerance for delayed events. Cluster policies govern compute settings, storage credentials provide authentication, and Git repositories manage source code. Therefore, when the streaming application needs a defined threshold for handling late-arriving event-time data, a watermark threshold is the relevant configuration.

Question 373

A data engineer wants to identify the physical storage location and other detailed properties of a Delta table. Which command should be used?

1. SHOW GRANTS
2. DESCRIBE DETAIL
3. DROP TABLE
4. CREATE SCHEMA

Correct Answer: 2

Explanation

DESCRIBE DETAIL provides detailed metadata about a Delta table, including information such as its storage location and other table properties. This is useful when engineers need to understand how a table is configured or troubleshoot storage and performance-related issues. SHOW GRANTS focuses on permissions, DROP TABLE removes a table according to the applicable behavior, and CREATE SCHEMA creates a namespace for database objects. These commands do not provide the same detailed table metadata. Therefore, when the physical location and detailed properties of a Delta table need to be inspected, DESCRIBE DETAIL is the appropriate command.

Question 374

A data engineer wants to store credentials used by a workload without hard-coding sensitive values directly into notebook source code. Which practice is appropriate?

1. Store secrets in a secret-management mechanism
2. Add credentials to a Git repository
3. Put credentials in table names
4. Add passwords to SQL comments

Correct Answer: 1

Explanation

Sensitive credentials should be managed through an appropriate secret-management mechanism rather than embedded directly in source code. Keeping credentials outside notebooks and application code reduces the risk of accidental exposure through source control, logs, or shared notebooks. Secret-management capabilities can allow workloads to retrieve required values securely at runtime according to the configured access model. Adding passwords to Git repositories, table names, or SQL comments creates unnecessary security risks and can expose credentials to unauthorized users. Therefore, when a workload requires sensitive credentials, using a supported secret-management mechanism instead of hard-coding values is the appropriate practice.

Question 375

A team needs a SQL-based compute resource for running interactive queries against governed lakehouse data. Which resource is appropriate?

1. SQL warehouse
2. Storage credential
3. External location
4. Volume

Correct Answer: 1

Explanation

A SQL warehouse provides compute resources designed for SQL workloads. It can be used for interactive queries, analytics, and other supported SQL-based operations against governed data. The warehouse provides the processing environment while Unity Catalog controls access to the underlying governed objects. A storage credential provides authentication for external storage, an external location defines a governed storage path, and a volume provides governed file storage. None of these objects is itself a SQL compute resource. Therefore, when users need a compute environment specifically for executing SQL queries against lakehouse data, a SQL warehouse is appropriate.

Question 376

A data engineer wants to change the number of partitions before writing a large DataFrame so that the output is distributed more appropriately across files. Which operation can be used?

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

Correct Answer: 1

Explanation

The repartition operation can change the number of partitions and redistribute records across them. This can be useful before writing a large DataFrame when the existing partition structure would produce an inefficient number or distribution of output files. Repartitioning involves data redistribution and therefore can introduce a shuffle, so it should be used based on the workload and expected benefits. DESCRIBE HISTORY examines Delta transaction history, VACUUM removes eligible obsolete files, and SHOW GRANTS displays permissions. Therefore, when a DataFrame needs to be redistributed into a different number of partitions before writing, repartition is the relevant operation.

Question 377

A pipeline needs to run every day at a defined time without requiring a user to start it manually. Which workflow capability should be configured?

1. Scheduled trigger
2. Column masking
3. Delta constraint
4. External volume

Correct Answer: 1

Explanation

A scheduled trigger allows a workflow or job to execute automatically according to a defined schedule. This is useful for recurring data engineering workloads such as daily ingestion, transformations, reporting preparation, or maintenance operations. Scheduling removes the need for a user to manually start each execution and provides a predictable processing pattern. Column masking protects sensitive values, Delta constraints enforce data-quality rules, and an external volume is used for governed file storage. These features do not control when a workflow starts. Therefore, when a pipeline needs to run automatically at a specified recurring time, a scheduled trigger is appropriate.

Question 378

A data engineer wants to examine which users or groups have access to a governed object before troubleshooting a permission issue. Which command is useful?

1. SHOW GRANTS
2. OPTIMIZE
3. MERGE
4. VACUUM

Correct Answer: 1

Explanation

SHOW GRANTS is useful for inspecting privileges assigned to supported Unity Catalog objects. During access troubleshooting, engineers can use the command to determine which users or groups have been granted permissions and identify whether the required privilege is present. This supports governance and helps administrators understand access configuration. OPTIMIZE improves Delta file organization, MERGE performs record synchronization, and VACUUM removes eligible obsolete files. These commands do not provide a view of object permissions. Therefore, when investigating whether a user or group has the necessary access to a governed object, SHOW GRANTS is an appropriate command.

Question 379

A data engineer wants to prevent a pipeline from repeatedly processing the same source records after successful completion of a previous run. Which design is most appropriate?

1. Full refresh on every run
2. Incremental and idempotent processing
3. Delete the source after every run
4. Disable checkpoints and retries

Correct Answer: 2

Explanation

Incremental and idempotent processing together provide a strong design for avoiding unnecessary duplicate processing. Incremental processing identifies and handles only new or changed source data, while idempotent logic ensures that repeated processing of the same input does not create incorrect duplicate results. This is especially important when jobs can be retried or rerun after operational failures. A full refresh repeatedly processes all source data, deleting source data can create data-loss risks, and disabling checkpoints or retries does not solve duplicate-processing problems. Therefore, combining incremental ingestion with idempotent processing is an appropriate design for reliable recurring pipelines.

Question 380

A team wants to review code changes made to a Databricks pipeline and determine who changed the code and when the changes were introduced. Which practice provides this capability?

1. Git version control
2. Delta VACUUM
3. Table partitioning
4. Data masking

Correct Answer: 1

Explanation

Git version control maintains a history of source-code changes and can provide information about commits, authors, timestamps, branches, and changes introduced over time. This makes it useful for reviewing Databricks pipeline development, troubleshooting regressions, and maintaining controlled deployment processes. Version control also supports collaboration by allowing developers to work on branches and review changes before merging them. Delta VACUUM manages obsolete table files, table partitioning organizes data for storage and query processing, and data masking protects sensitive values. Therefore, when a team needs traceability for pipeline code changes, Git version control is the appropriate practice.