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

 

Question 101

Which Unity Catalog object provides a logical grouping of schemas and other securable objects?

1. Catalog
2. Volume
3. Storage credential
4. External location

Correct Answer: 1

Explanation

A catalog is the top-level container in the standard Unity Catalog namespace hierarchy. It groups schemas and the objects contained within those schemas, such as tables, views, and volumes. Organizations can use catalogs to separate data according to departments, environments, business domains, or governance boundaries. Access to catalogs is controlled through Unity Catalog privileges, and users typically need appropriate permissions on the catalog before they can access objects underneath it. Volumes provide governed file access, storage credentials provide authentication to cloud storage, and external locations define governed storage paths. Therefore, a catalog is the appropriate object for logically grouping schemas and their related data assets.

Question 102

Which feature allows a data engineer to automatically discover and ingest newly arriving files from cloud storage?

1. VACUUM
2. Auto Loader
3. OPTIMIZE
4. Time travel

Correct Answer: 2

Explanation

Auto Loader provides scalable and incremental ingestion of files arriving in cloud storage. It is designed to detect new files and process them without repeatedly scanning the entire storage directory. This makes it especially useful for data engineering pipelines where files arrive continuously or in large numbers. Auto Loader can also support schema inference and schema evolution, helping pipelines adapt to changes in incoming data. VACUUM removes obsolete Delta files, OPTIMIZE improves the physical layout of Delta data, and time travel provides access to historical table versions. Therefore, Auto Loader is the appropriate capability for automatically discovering and ingesting newly arriving cloud-storage files.

Question 103

Which Unity Catalog object is used to define a governed path in cloud object storage?

1. Schema
2. View
3. External location
4. Catalog

Correct Answer: 3

Explanation

An external location defines a governed cloud storage path in Unity Catalog. It associates a specific storage path with a storage credential, allowing administrators to control access to external data through Unity Catalog permissions. External locations are commonly used when working with external tables and other data assets stored outside managed table storage. A catalog groups schemas, a schema organizes objects within a catalog, and a view stores a reusable query definition. External locations are therefore important when an organization needs centralized governance over cloud storage paths while keeping authentication and authorization under controlled management rather than distributing storage credentials directly to individual users.

Question 104

Which Delta Lake capability can be used to maintain historical versions of table data for auditing or recovery purposes?

1. Auto Loader
2. Time travel
3. Cluster policy
4. Materialized view

Correct Answer: 2

Explanation

Delta Lake time travel allows users to access previous versions of a Delta table by using historical versions or timestamps, subject to the availability of the required transaction history and data files. This capability is useful for auditing changes, investigating unexpected results, reproducing historical analyses, and recovering from certain accidental modifications. Delta Lake records table changes in its transaction log, which supports this historical access. Auto Loader handles incremental file ingestion, cluster policies govern compute configuration, and materialized views maintain reusable query results. Therefore, time travel is the Delta Lake feature specifically designed for accessing historical table states.

Question 105

Which operation is commonly used to synchronize a target Delta table with records from a source dataset?

1. MERGE
2. VACUUM
3. DESCRIBE
4. OPTIMIZE

Correct Answer: 1

Explanation

MERGE is commonly used to synchronize a target Delta table with records from a source dataset. It evaluates matching conditions and can perform different actions depending on whether a source record already exists in the target. For example, matching records can be updated while unmatched records can be inserted. MERGE can also support conditional deletes when required. This makes it useful for change data capture, incremental processing, and maintaining current-state tables. VACUUM manages obsolete files, DESCRIBE provides metadata information, and OPTIMIZE improves physical data layout. Therefore, MERGE is the appropriate operation when source records must be conditionally inserted, updated, or deleted in a Delta table.

Question 106

Which Structured Streaming feature helps a pipeline recover its processing progress after a failure or restart?

1. View
2. Catalog
3. Checkpoint
4. Volume

Correct Answer: 3

Explanation

A checkpoint stores progress and state information for a Structured Streaming query. When a streaming workload stops unexpectedly or is restarted, the checkpoint allows the query to recover from its previously recorded processing state instead of treating all input as completely new. This is important for building reliable streaming pipelines, particularly when stateful operations or incremental processing are involved. Checkpoint locations should be configured carefully and maintained consistently for the streaming workload. Views, catalogs, and volumes serve different purposes within the Databricks environment. Therefore, checkpointing is the feature that provides the persistent progress information required for streaming recovery.

Question 107

Which architecture organizes data into Bronze, Silver, and Gold layers?

1. Dimensional architecture
2. Medallion architecture
3. Three-tier application architecture
4. Hub-and-spoke architecture

Correct Answer: 2

Explanation

The medallion architecture organizes data into progressively refined layers, commonly known as Bronze, Silver, and Gold. Bronze generally contains raw or minimally processed data collected from source systems. Silver applies cleaning, validation, deduplication, and transformation to create reliable datasets. Gold contains curated datasets designed for business analytics, reporting, and downstream consumption. This layered approach helps separate ingestion from transformation and presentation while improving data lineage and maintainability. Other architecture patterns may be used for applications or analytical modeling, but they do not specifically define the Bronze, Silver, and Gold data-processing layers. Therefore, the medallion architecture is the correct choice for this pattern.

Question 108

Which Unity Catalog privilege allows a user to access a schema within a catalog?

1. SELECT
2. MODIFY
3. USE SCHEMA
4. CREATE TABLE

Correct Answer: 3

Explanation

USE SCHEMA allows a principal to access a schema namespace in Unity Catalog. It is generally needed when working with objects contained within that schema, together with the appropriate permissions on the parent catalog and specific object. SELECT grants read access to data, MODIFY permits supported data changes, and CREATE TABLE allows table creation when the required parent permissions are also present. Unity Catalog uses a hierarchical permission model, meaning a user may require privileges at multiple levels before an operation succeeds. Therefore, when the requirement is specifically to access a schema namespace, USE SCHEMA is the relevant Unity Catalog privilege.

Question 109

Which feature is most appropriate for protecting sensitive column values by displaying a transformed value to unauthorized users?

1. Column masking
2. Partition pruning
3. Auto Loader
4. Checkpointing

Correct Answer: 1

Explanation

Column masking can protect sensitive information by controlling how column values are presented to different users. Depending on the configured policy, authorized users can receive the original value while other users receive a masked or transformed representation. This is useful for protecting confidential information such as personal identifiers, financial values, or other restricted fields while allowing the same table to remain useful for authorized analysis. Partition pruning improves query performance, Auto Loader handles incremental file ingestion, and checkpointing supports streaming recovery. Therefore, column masking directly addresses the requirement to prevent unauthorized users from seeing sensitive column values in their original form.

Question 110

Which Spark operation can increase the number of partitions and redistribute records across the cluster?

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

Correct Answer: 2

Explanation

The repartition operation redistributes records across a specified number of Spark partitions. It can be useful when a dataset has an unsuitable partition distribution for the next stage of processing or when additional parallelism is needed. Repartitioning generally involves a shuffle, meaning data may move between executors, so it should be used when the expected processing benefits justify the additional cost. CACHE stores computed data for reuse, GRANT manages permissions, and VACUUM removes obsolete files from Delta storage. Therefore, REPARTITION is the Spark operation that can redistribute data and change the number of partitions used for subsequent processing.

Question 111

Which feature can help handle newly added columns in incoming files without requiring the ingestion process to be completely redesigned?

1. Schema evolution
2. File deletion
3. Time travel
4. Data masking

Correct Answer: 1

Explanation

Schema evolution allows supported data pipelines to accommodate changes in the structure of incoming data. For example, when a source system adds a new column, a pipeline configured to support schema evolution may be able to incorporate that change rather than failing because the incoming schema no longer exactly matches the existing schema. This is particularly useful in Auto Loader and Delta-based ingestion scenarios where source schemas can change over time. Time travel provides historical table access, data masking controls sensitive-value visibility, and file deletion does not provide schema management. Schema evolution therefore helps data engineers build ingestion pipelines that are more resilient to supported structural changes in source data.

Question 112

Which Delta Lake maintenance operation removes obsolete files that are no longer needed according to retention settings?

1. MERGE
2. OPTIMIZE
3. VACUUM
4. CREATE VIEW

Correct Answer: 3

Explanation

VACUUM is used to remove obsolete data files that are no longer required by a Delta table according to its retention configuration. It helps manage storage consumption by cleaning up files left behind after table modifications. Because historical versions can depend on older files, VACUUM should be used with an understanding of the table’s time-travel requirements and retention policies. MERGE changes table records, OPTIMIZE reorganizes files and can compact small files, and CREATE VIEW creates reusable query logic. Therefore, VACUUM is the appropriate maintenance operation when the objective is to clean up obsolete Delta data files while respecting configured retention requirements.

Question 113

Which feature provides centralized permissions and governance for tables, schemas, catalogs, and volumes?

1. Unity Catalog
2. Git folders
3. Spark UI
4. Cluster logs

Correct Answer: 4

Explanation

Unity Catalog provides centralized governance and access control for data assets across Azure Databricks. It supports permissions for catalogs, schemas, tables, views, volumes, external locations, and other governed objects. This centralized model helps organizations manage data access consistently and provides capabilities related to auditing, discovery, and governance. Git folders are used for source-code and notebook version control, while Spark UI and cluster logs are primarily used for monitoring and troubleshooting workloads. Therefore, Unity Catalog is the component responsible for centralized data governance and permissions. It provides the organizational framework required to control access to governed data assets across Databricks environments.

Question 114

Which Structured Streaming feature is used to limit how long state is maintained for events that arrive late?

1. Watermark
2. Catalog
3. External location
4. View

Correct Answer: 1

Explanation

A watermark establishes an event-time threshold that allows Structured Streaming to determine how long it should retain state for late-arriving events. This is especially important for stateful operations such as aggregations, stream-stream joins, and deduplication. By defining an acceptable lateness period, the streaming engine can eventually remove older state and reduce resource consumption. Watermarks do not simply discard every late record immediately; their behavior depends on the streaming operation and event-time semantics. Catalogs organize data namespaces, external locations govern cloud storage paths, and views provide reusable query definitions. Therefore, watermarking is the appropriate mechanism for controlling state retention in late-arriving event scenarios.

Question 115

Which Databricks capability allows multiple tasks to be connected through dependencies and executed according to a defined workflow?

1. Databricks Jobs
2. Unity Catalog
3. Delta Sharing
4. External locations

Correct Answer: 2

Explanation

Databricks Jobs, currently referred to in current Databricks terminology as Lakeflow Jobs, provides workflow orchestration capabilities. A job can contain multiple tasks and define dependencies that determine the order in which those tasks execute. Jobs can also include schedules, triggers, retries, notifications, and other operational settings depending on the workload. Unity Catalog is focused on data governance, Delta Sharing is designed for data sharing, and external locations govern access to cloud storage paths. Therefore, Databricks Jobs is the appropriate capability for coordinating multiple data engineering tasks into an automated workflow with defined dependencies and execution behavior.

Question 116

Which technique can reduce the amount of data read when a query contains a filter on a partitioning column?

1. Schema evolution
2. Partition pruning
3. Time travel
4. Column masking

Correct Answer: 2

Explanation

Partition pruning allows the query engine to avoid reading partitions that cannot contain records matching the query’s filter. For example, if a table is partitioned by date and a query requests records for one date, unrelated date partitions can potentially be skipped. This reduces data scanning and can improve query performance by lowering I/O and processing requirements. Partition pruning is different from schema evolution, which handles changes to data structures. Time travel provides access to historical Delta versions, while column masking controls visibility of sensitive values. Therefore, partition pruning is the appropriate optimization technique when filters can eliminate unnecessary table partitions from the query.

Question 117

Which approach provides a reliable way to prevent repeated processing from creating duplicate target records?

1. Use idempotent processing
2. Disable transaction logging
3. Remove all validation
4. Increase file sizes only

Correct Answer: 4

Explanation

Idempotent processing ensures that running the same input more than once does not produce unintended additional effects. In data engineering, this can be achieved through stable business keys, deduplication logic, MERGE operations, and appropriate checkpoint management. Idempotent pipelines are particularly important when workloads may be retried after failures or when source systems can deliver duplicate events or files. Disabling transaction logging would reduce reliability, removing validation would weaken data quality, and changing file sizes alone would not address duplicate records. Therefore, designing processing logic to be idempotent is a fundamental approach for creating reliable pipelines that can safely handle retries and repeated inputs.

Question 118

Which Delta Lake operation can compact small files into a more efficient physical layout?

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

Correct Answer: 3

Explanation

OPTIMIZE is used to improve the physical organization of Delta table data and can compact many small files into fewer, larger files. Small-file accumulation can occur when data is written frequently or in small batches, creating additional file-management overhead during queries. Compacting files can reduce this overhead and may improve query performance. OPTIMIZE changes the physical organization rather than the logical meaning of the table data. GRANT manages access permissions, DESCRIBE provides metadata, and SHOW GRANTS displays privilege information. Therefore, OPTIMIZE is the appropriate Delta Lake operation when the objective is to improve file layout and address small-file performance issues.

Question 119

Which Unity Catalog object provides governed access to non-tabular files such as CSV, JSON, or images?

1. Catalog
2. Volume
3. Schema
4. Materialized view

Correct Answer: 2

Explanation

A volume provides governed access to files that are not necessarily represented as database tables. Examples can include CSV files, JSON documents, images, machine-learning artifacts, and other file-based data. Volumes are managed through Unity Catalog and can therefore benefit from centralized permissions and governance. A catalog groups schemas, a schema organizes tables and other objects, and a materialized view maintains query-derived results. Volumes are particularly useful when data engineers need to manage file-based assets while still applying Unity Catalog access controls. Therefore, a volume is the appropriate Unity Catalog object for governed access to non-tabular files.

Question 120

Which capability allows a Databricks workload to automatically retry a failed task?

1. Table constraint
2. External location
3. Job retry configuration
4. Column masking

Correct Answer: 3

Explanation

Job retry configuration allows a Databricks workload to automatically attempt a failed task again according to defined settings. Retries are useful when failures are temporary, such as intermittent infrastructure problems or transient service issues. Configuring retries can improve the resilience of scheduled data engineering workflows and reduce the need for manual intervention. The retry policy should still be designed carefully because repeated failures caused by invalid data or code will generally require investigation rather than unlimited retries. Table constraints enforce data rules, external locations govern cloud storage paths, and column masking protects sensitive values. Therefore, job retry configuration is the appropriate feature for automatically retrying failed tasks.