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

 

Question 61

In Azure Databricks, which object is used to define a logical namespace above schemas in Unity Catalog?

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

Correct Answer: 1

Explanation

A catalog is the highest-level namespace in the standard Unity Catalog three-level object hierarchy. The hierarchy is catalog, schema, and object, where the object can be a table, view, volume, or another supported securable object. Catalogs help organizations separate datasets according to business domains, environments, or organizational boundaries. Permissions can also be applied at the catalog level and inherited by lower-level objects when appropriate. A volume is used for file-based data, while an external location defines governed access to cloud storage paths. Storage credentials provide authentication for accessing external storage but do not create the namespace hierarchy.

Question 62

Which Azure Databricks feature is primarily designed to automatically ingest files arriving in cloud storage while maintaining scalable incremental processing?

1. Delta Sharing
2. Auto Loader
3. Unity Catalog
4. Databricks SQL

Correct Answer: 3

Explanation

Auto Loader is designed for incremental and scalable ingestion of new files arriving in cloud storage. Instead of repeatedly scanning the entire directory, Auto Loader tracks discovered files and processes new arrivals efficiently. It supports common cloud storage environments and integrates with Structured Streaming. Auto Loader can also help handle evolving schemas and provides options for managing malformed or unexpected data. Delta Sharing focuses on data sharing, Unity Catalog provides governance and access control, and Databricks SQL is primarily intended for SQL-based analytics. Auto Loader is therefore the appropriate feature when a data engineering pipeline needs continuous or incremental file ingestion.

Question 63

A data engineer needs to provide governed access to a specific cloud storage path in Unity Catalog. Which object should be configured to represent that path?

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

Correct Answer: 4

Explanation

An external location in Unity Catalog represents a governed cloud storage path and associates that path with a storage credential. This allows administrators to control access to external storage through Unity Catalog permissions rather than relying only on direct cloud-storage permissions. External locations are commonly used when creating external tables or managing external data assets. A catalog organizes schemas and objects, while a schema organizes tables, views, and other objects. A volume provides governed file access but does not replace the external-location concept for defining a cloud storage path. Therefore, an external location is the appropriate object for this requirement.

Question 64

Which statement best describes a managed table in Unity Catalog?

1. Its data must always remain outside the metastore
2. Its data lifecycle is managed by Azure Databricks
3. It can only contain streaming data
4. It cannot use Delta Lake

Correct Answer: 2

Explanation

A managed table is a table for which Azure Databricks manages the data lifecycle and storage location according to the configured catalog and schema settings. This simplifies administration because users generally do not need to manage the underlying storage path manually. Managed tables commonly use Delta Lake, which provides transactional capabilities, schema enforcement, and reliable data processing. External tables differ because their underlying data resides at a location explicitly managed outside the table’s managed storage lifecycle. Managed tables are therefore useful when an organization wants Databricks and Unity Catalog to control the table’s storage and lifecycle while providing governed access to users.

Question 65

Which command is most appropriate for applying multiple inserts, updates, and deletes to a Delta table based on changes from a source dataset?

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

Correct Answer: 1

Explanation

The MERGE operation is designed to synchronize a Delta table with a source dataset when different actions may be required depending on whether matching records exist. A MERGE statement can perform updates for matched records and inserts for unmatched records, and it can also support delete conditions. This makes MERGE particularly useful for change data capture and incremental data processing scenarios. DESCRIBE provides metadata information, OPTIMIZE improves storage layout and query performance, and VACUUM removes obsolete files according to retention settings. For applying multiple conditional data changes in one operation, MERGE is the appropriate Delta Lake capability.

Question 66

Which feature can help prevent unauthorized users from seeing particular columns in query results?

1. Data skipping
2. Partition pruning
3. Column masking
4. File compaction

Correct Answer: 4

Explanation

Column masking can be used to restrict how sensitive column values are exposed to users. Depending on the implemented policy, authorized users may see the original value while other users receive a masked or transformed representation. This provides a governance mechanism for protecting sensitive information such as personal identifiers or confidential business values. Data skipping and partition pruning are performance-related techniques that reduce unnecessary data processing. File compaction improves storage efficiency and query performance by reducing inefficient file layouts. Column masking addresses data visibility directly, making it the appropriate feature when access to sensitive column values needs to be controlled.

Question 67

A data engineer wants to prevent duplicate records from being introduced when processing the same input more than once. Which approach is most appropriate?

1. Increase cluster size
2. Use idempotent processing logic
3. Disable checkpointing
4. Remove schema validation

Correct Answer: 2

Explanation

Idempotent processing means that processing the same input repeatedly produces the same intended final result without creating unwanted duplicate effects. This is especially important in data pipelines where retries, failures, or repeated file delivery can cause the same data to be processed more than once. Techniques such as MERGE operations using stable business keys, deduplication, and properly managed checkpoints can support idempotent designs. Increasing cluster size addresses performance rather than correctness. Disabling checkpointing can make streaming reliability worse, while removing schema validation does not prevent duplicates. Therefore, designing the pipeline to be idempotent is an effective way to handle repeated input safely.

Question 68

Which Structured Streaming feature is used to maintain progress information and state so processing can recover after a failure?

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

Correct Answer: 3

Explanation

A checkpoint stores important progress and state information for a Structured Streaming query. It enables the streaming application to recover from failures and continue processing from an appropriate point rather than starting over from the beginning. Checkpoints are particularly important for reliable streaming pipelines that may contain stateful operations or need consistent processing across restarts. A catalog is part of the Unity Catalog namespace, a schema organizes objects, and a volume provides governed file storage access. These objects do not perform the recovery role provided by streaming checkpoints. Proper checkpoint configuration is therefore a fundamental part of building resilient Structured Streaming workloads.

Question 69

Which Unity Catalog object is designed to provide governed file-based access within a catalog and schema?

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

Correct Answer: 4

Explanation

A volume is a Unity Catalog object that provides governed access to files within a catalog and schema. Volumes are useful when data engineers need to work with non-tabular files such as JSON, CSV, images, documents, or other files while still applying Unity Catalog governance. A storage credential represents authentication information for accessing cloud storage. An external location represents a governed storage path, while a warehouse provides SQL compute rather than file governance. Volumes therefore provide an important bridge between file-based data and Unity Catalog’s permission and governance model, allowing organizations to control access to files through familiar Unity Catalog privileges.

Question 70

Which operation is commonly used to reduce the number of small files in a Delta table and improve query performance?

1. OPTIMIZE
2. MERGE
3. GRANT
4. SHOW

Correct Answer: 1

Explanation

The OPTIMIZE operation is used to improve the physical organization of Delta table data. One important use is compacting many small files into fewer larger files, which can reduce file-management overhead and improve query performance. Small files are often produced by frequent incremental writes or streaming workloads. OPTIMIZE can therefore be useful as part of ongoing table maintenance. MERGE is used for conditional inserts, updates, and deletes. GRANT manages permissions, while SHOW commands display metadata or configuration information. OPTIMIZE should be used when the primary requirement is improving the physical layout and reducing small-file overhead rather than changing the logical contents of the table.

Question 71

Which architecture pattern separates raw, cleaned, and business-ready data into progressively refined layers?

1. Star schema
2. Lambda architecture
3. Medallion architecture
4. Snowflake schema

Correct Answer: 3

Explanation

The medallion architecture organizes data into progressively refined layers, commonly called Bronze, Silver, and Gold. The Bronze layer generally contains raw or minimally processed data. The Silver layer applies cleaning, validation, deduplication, and transformation. The Gold layer contains curated, business-oriented datasets designed for analytics and reporting. This structure helps data engineering teams separate ingestion from transformation and consumption while improving traceability and maintainability. Star and snowflake schemas are dimensional modeling approaches used primarily for analytical data models, while Lambda architecture combines batch and streaming processing patterns. The medallion approach is particularly common in modern lakehouse implementations using Azure Databricks.

Question 72

Which SQL statement is used to grant a user or group permission to select data from a table?

1. DENY SELECT
2. GRANT SELECT
3. CREATE SELECT
4. ALLOW TABLE

Correct Answer: 2

Explanation

The GRANT SELECT statement is used to provide permission to read data from a table or another supported securable object. In Unity Catalog, administrators and object owners can grant privileges to users, groups, or service principals according to the organization’s governance requirements. Access generally follows the Unity Catalog privilege model, where users must have the necessary permissions on relevant parent objects as well as the object itself. DENY SELECT is not the standard Unity Catalog syntax for granting access. CREATE SELECT and ALLOW TABLE are also not valid commands for this purpose. Therefore, GRANT SELECT is the appropriate statement for providing read access.

Question 73

Which feature can be used to detect and manage changes in the schema of files processed by Auto Loader?

1. Schema evolution
2. Data masking
3. Query caching
4. File compaction

Correct Answer: 4

Explanation

Schema evolution allows ingestion pipelines to respond to changes in incoming data structures. In Auto Loader workloads, schema evolution capabilities can help accommodate newly added columns or other supported schema changes without requiring the pipeline to be completely redesigned. This is particularly valuable when upstream systems evolve over time. Depending on the configuration and type of schema change, the pipeline may update its schema or route problematic records according to the selected behavior. Data masking protects sensitive values, query caching improves repeated query performance, and file compaction addresses storage layout. Therefore, schema evolution is the relevant capability for handling changing input schemas.

Question 74

A streaming pipeline must ignore records that arrive significantly later than the expected event time. Which Structured Streaming feature should be considered?

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

Correct Answer: 1

Explanation

A watermark allows Structured Streaming to track event-time progress and determine how long the system should wait for late-arriving data. When a watermark is combined with appropriate stateful operations, the streaming engine can eventually remove old state and limit how much historical information must be maintained. This helps control memory usage and allows the pipeline to handle delayed events within a defined tolerance. A catalog, volume, or view does not provide event-time management. Watermarking is therefore particularly useful for streaming workloads involving out-of-order or late-arriving events, especially when aggregations or other stateful operations depend on event-time information.

Question 75

Which Azure Databricks capability allows a data engineering team to define tasks, dependencies, schedules, and retries for a pipeline workload?

1. Unity Catalog
2. Databricks Jobs
3. Delta Sharing
4. DBFS

Correct Answer: 2

Explanation

Databricks Jobs, also referred to as Lakeflow Jobs in current Databricks terminology, provides orchestration capabilities for data engineering workloads. Teams can define multiple tasks, establish dependencies between those tasks, configure schedules or triggers, and specify retry behavior for failed executions. This allows complex workflows to be automated rather than manually executed. Unity Catalog focuses on governance and data access. Delta Sharing enables secure data sharing with other users or organizations, while DBFS is associated with file and storage access concepts. A job workflow can connect notebooks, SQL tasks, Python scripts, pipelines, and other supported workload types into an automated execution sequence.

Question 76

Which technique can improve query performance by avoiding unnecessary scanning of data based on filter conditions?

1. Predicate pushdown
2. Data duplication
3. Schema deletion
4. Checkpoint removal

Correct Answer: 3

Explanation

Predicate pushdown improves query efficiency by applying filter conditions as close to the data source as possible. When supported, this allows the execution engine to avoid reading data that cannot satisfy the query predicate. Reducing unnecessary data scanning can lower I/O, processing requirements, and query latency. Other optimization mechanisms, such as partition pruning and data skipping, can provide additional benefits depending on the table layout and query conditions. Data duplication generally increases storage and processing requirements, while deleting schemas or removing checkpoints does not provide a valid query optimization strategy. Predicate pushdown is therefore a useful technique for minimizing unnecessary data access during query execution.

Question 77

Which privilege is generally required to create a table inside a Unity Catalog schema?

1. USE SCHEMA and CREATE TABLE
2. SELECT only
3. USE CATALOG only
4. READ VOLUME only

Correct Answer: 1

Explanation

Creating a table in a Unity Catalog schema requires the appropriate privileges on the parent catalog and schema as well as the privilege to create the table object. In a typical scenario, the user needs USE CATALOG on the catalog, USE SCHEMA on the schema, and CREATE TABLE on the schema. SELECT alone provides read access and does not allow object creation. USE CATALOG by itself also does not grant the ability to create tables inside a schema. READ VOLUME is intended for accessing files within volumes. Unity Catalog’s hierarchical permission model ensures that users have both namespace access and the specific privilege required for the operation they want to perform.

Question 78

Which Delta Lake capability allows a data engineer to query a previous version of a table after changes have been committed?

1. Auto Loader
2. Liquid clustering
3. Time travel
4. Data masking

Correct Answer: 4

Explanation

Delta Lake time travel allows users to access previous versions of a table using historical versions or timestamps, provided the required transaction-log and data files are still available. This capability is useful for auditing, debugging, reproducing previous analytical results, and recovering from certain accidental changes. Delta Lake maintains transaction history that records changes to the table over time. Auto Loader is an ingestion technology, liquid clustering is a data-layout optimization technique, and data masking controls how sensitive values are displayed. Time travel therefore directly addresses the requirement to query a table as it existed at an earlier point in its history.

Question 79

Which operation can be used to redistribute data across a larger number of partitions during a Spark transformation?

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

Correct Answer: 3

Explanation

The repartition operation redistributes data across a specified number of Spark partitions. It can be useful when the current partitioning does not provide suitable parallelism or when a workload requires data to be redistributed before a transformation or join. Because repartitioning can involve a shuffle, it should be used carefully and based on workload requirements. CACHE stores computed data for potential reuse, GRANT manages permissions, and VACUUM removes obsolete files from Delta table storage according to retention rules. Repartitioning is therefore the appropriate Spark technique when the goal is to change the distribution and number of partitions for subsequent processing.

Question 80

Which Azure Databricks feature provides centralized governance for tables, views, volumes, catalogs, and other data assets?

1. Unity Catalog
2. Spark UI
3. Databricks Repos
4. Cluster logs

Correct Answer: 2

Explanation

Unity Catalog provides centralized governance for data and AI assets across Azure Databricks environments. It supports a unified permission model for catalogs, schemas, tables, views, volumes, external locations, and other securable objects. It also helps organizations implement consistent access controls, auditing, discovery, and data governance practices. The Spark UI is primarily used for monitoring Spark applications and understanding execution performance. Databricks Repos supports source-control integration and development workflows, while cluster logs provide operational information about compute resources. Unity Catalog is therefore the component intended to provide centralized governance and access management across governed data assets.