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

 

Question 81

Which Unity Catalog feature allows administrators to manage access to cloud storage without exposing the underlying storage credentials directly to users?

1. Storage credential
2. SQL warehouse
3. Materialized view
4. Notebook

Correct Answer: 1

Explanation

A storage credential is a Unity Catalog securable object that represents an authentication mechanism used to access cloud storage. It allows administrators to centralize and govern access to storage resources without requiring individual users to manage or directly receive the underlying credentials. Storage credentials are commonly associated with external locations, which define specific cloud storage paths that can be accessed through Unity Catalog. SQL warehouses provide compute for SQL workloads, materialized views store query results for faster access, and notebooks provide development environments. Storage credentials therefore play an important role in securely connecting Unity Catalog governance with external cloud storage resources.

Question 82

Which Azure Databricks capability is most appropriate for storing reusable SQL query results that are automatically maintained for analytical workloads?

1. External location
2. Materialized view
3. Storage credential
4. Volume

Correct Answer: 2

Explanation

A materialized view stores the results of a query so that repeated analytical queries can access precomputed data rather than performing the complete underlying computation each time. This can improve performance for workloads involving expensive aggregations or transformations, especially when the underlying data changes according to supported refresh mechanisms. Materialized views are different from standard views because a standard view generally stores the query definition rather than the resulting dataset. External locations manage governed cloud storage paths, storage credentials provide authentication, and volumes provide governed file access. Therefore, a materialized view is appropriate when reusable, maintained query results are required for analytical workloads.

Question 83

A data engineer needs to ensure that a pipeline does not accidentally write records with NULL values into a required Delta column. Which Delta Lake feature can enforce this requirement?

1. Time travel
2. Data skipping
3. NOT NULL constraint
4. Liquid clustering

Correct Answer: 3

Explanation

A NOT NULL constraint can enforce that a specified column must contain a value when records are written to a Delta table. This provides an important data-quality control because invalid records that violate the constraint can be rejected rather than silently becoming part of the dataset. Constraints are useful for protecting data integrity at the table level and complement other validation logic implemented in pipelines. Time travel is used to access historical table versions, data skipping improves query performance, and liquid clustering helps organize data for efficient access. When the requirement is specifically to prevent NULL values in a required column, a NOT NULL constraint directly addresses that requirement.

Question 84

Which operation should be used when an existing Delta table needs to have outdated data files removed after the applicable retention period?

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

Correct Answer: 4

Explanation

VACUUM removes old data files that are no longer required by the Delta table according to the configured retention rules. It is primarily a storage-maintenance operation and can help reduce the amount of unused data occupying cloud storage. OPTIMIZE serves a different purpose by improving the physical organization of data and reducing small-file problems. MERGE changes table records based on matching conditions, while DESCRIBE provides metadata information. VACUUM must be used carefully because removing old files can affect the ability to access historical table versions that depend on those files. Therefore, VACUUM is the appropriate operation for cleaning up obsolete Delta data files.

Question 85

Which approach is most appropriate when a data engineer wants to process only newly arrived files rather than repeatedly processing all files in a cloud storage directory?

1. Auto Loader
2. VACUUM
3. OPTIMIZE
4. GRANT

Correct Answer: 1

Explanation

Auto Loader is designed for efficient incremental file ingestion from cloud storage. It tracks files that have already been discovered and processes newly arriving files without requiring the pipeline to repeatedly scan and process the entire directory contents. This makes it particularly useful for scalable ingestion workloads where files continuously arrive in cloud storage. Auto Loader also integrates with Structured Streaming and supports schema inference and schema evolution capabilities. VACUUM manages obsolete Delta files, OPTIMIZE improves table layout and file organization, and GRANT manages permissions. When the requirement is efficient incremental ingestion of newly arriving files, Auto Loader is the appropriate solution.

Question 86

Which Databricks feature allows developers to synchronize notebooks and source code with a Git repository?

1. Unity Catalog
2. Databricks Git folders
3. Delta Sharing
4. SQL warehouse

Correct Answer: 4

Explanation

Databricks Git folders provide integration between Databricks development environments and Git repositories. They allow data engineers and developers to synchronize notebooks and supported source files with version-control systems. This helps teams track changes, collaborate on development, review modifications, and maintain deployment workflows. Unity Catalog focuses on data governance and access control rather than source-code versioning. Delta Sharing is intended for securely sharing data, while SQL warehouses provide compute for SQL analytics. Git integration is therefore the appropriate capability when the requirement involves connecting Databricks development work with source-control repositories and maintaining versioned engineering assets.

Question 87

Which technique can reduce the amount of data processed when a query filters on a partitioning column?

1. Predicate pruning
2. Partition pruning
3. Schema evolution
4. Checkpointing

Correct Answer: 2

Explanation

Partition pruning allows the query engine to skip partitions that cannot contain records matching the query’s filter conditions. This can significantly reduce the amount of data that must be read and processed. For example, if a table is partitioned by date and a query requests only one specific date, the engine can avoid scanning unrelated date partitions. This improves query efficiency and can reduce I/O and processing costs. Predicate pushdown and data skipping can provide additional optimization benefits, but partition pruning specifically concerns eliminating unnecessary partitions. Schema evolution manages changing schemas, while checkpointing supports streaming recovery rather than query filtering.

Question 88

Which feature can be used to define a reusable query definition without physically storing the query results as a separate dataset?

1. Volume
2. External location
3. Standard view
4. Storage credential

Correct Answer: 3

Explanation

A standard view stores a SQL query definition that can be reused by users and applications. When the view is queried, the underlying query is executed against the referenced data rather than relying on a separately maintained stored result. Views are useful for simplifying complex queries, providing consistent business logic, and controlling which columns or rows are exposed to users. A volume provides governed file access, an external location represents a governed cloud storage path, and a storage credential provides authentication for external storage. Therefore, a standard view is the appropriate object when reusable query logic is required without creating a separately stored result dataset.

Question 89

A streaming pipeline processes events that can arrive several minutes late. Which configuration helps control how long the system retains state for late events?

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

Correct Answer: 1

Explanation

A watermark establishes an event-time threshold that helps Structured Streaming determine how long state should be retained for late-arriving records. This is particularly useful for stateful operations such as streaming aggregations, joins, and deduplication. By defining an appropriate lateness threshold, the pipeline can continue accepting reasonably late events while eventually removing older state that is unlikely to receive additional relevant records. This helps control memory consumption and maintain scalable streaming workloads. Catalogs and schemas organize data assets, while volumes provide governed access to files. A watermark is therefore the appropriate mechanism when a streaming pipeline must manage late events and state retention.

Question 90

Which operation can be used to update existing records and insert new records into a Delta table based on a matching key?

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

Correct Answer: 3

Explanation

MERGE is designed to synchronize records between a source dataset and a target Delta table using specified matching conditions. When a source record matches an existing target record, the MERGE statement can update that record. When no matching record exists, it can insert a new record. Additional conditions can also support delete operations. This makes MERGE particularly useful for change data capture, incremental loading, and maintaining current-state tables from source systems. OPTIMIZE manages physical file organization, VACUUM removes obsolete files, and DESCRIBE returns metadata. Therefore, MERGE is the appropriate operation when both updates and inserts need to be performed according to a matching key.

Question 91

Which Unity Catalog permission allows a principal to access objects within a catalog namespace?

1. SELECT
2. USE CATALOG
3. MODIFY
4. READ VOLUME

Correct Answer: 2

Explanation

USE CATALOG provides permission to access a catalog namespace in Unity Catalog. It is generally required when a user needs to work with objects contained within that catalog, together with the appropriate permissions on the schema and specific object. SELECT provides read access to supported data objects, MODIFY allows changes to data where granted, and READ VOLUME provides read access to files in a volume. Unity Catalog uses hierarchical permissions, so having a privilege on a table alone may not be sufficient if the required parent-level permissions are missing. USE CATALOG therefore plays an important role in enabling access to objects within a governed catalog.

Question 92

Which Delta Lake capability provides transactional consistency when multiple operations modify a table?

1. Delta transaction log
2. SQL warehouse
3. Git folder
4. Cluster policy

Correct Answer: 4

Explanation

The Delta transaction log records changes made to a Delta table and provides the foundation for transactional consistency. Each table modification is represented in the transaction history, allowing Delta Lake to provide reliable reads and writes even when multiple operations occur. The transaction log also supports capabilities such as time travel, schema enforcement, and recovery from certain failures. SQL warehouses provide compute for SQL workloads, Git folders support source-code versioning, and cluster policies govern compute configurations. These features do not provide Delta’s table-level transaction mechanism. Therefore, the Delta transaction log is fundamental to maintaining consistent and reliable transactional behavior for Delta tables.

Question 93

Which technique is useful when a pipeline should isolate records that fail data-quality validation instead of stopping the entire workload?

1. Quarantine invalid records
2. Disable validation
3. Delete the source data
4. Increase the cluster size

Correct Answer: 4

Explanation

Quarantining invalid records allows a data pipeline to separate records that fail validation from records that meet quality requirements. Instead of stopping the entire workload, valid records can continue through the normal processing path while rejected records are stored separately for investigation and possible correction. This approach supports reliable data-quality management and makes failures easier to monitor and troubleshoot. Disabling validation would allow potentially incorrect data to enter downstream datasets. Deleting source data risks permanent data loss, while increasing cluster size addresses compute capacity rather than data correctness. A quarantine path is therefore a practical design for handling rejected records without unnecessarily stopping successful processing.

Question 94

Which feature can be used to restrict access to rows based on the identity of the querying user?

1. File compaction
2. Row filtering
3. Schema evolution
4. Data skipping

Correct Answer: 2

Explanation

Row filtering can restrict which records a user is allowed to see based on defined access rules. This can be useful in scenarios where different users or groups should access different subsets of the same table. For example, a regional user might be allowed to view only records associated with their assigned region. Row-level security can therefore provide fine-grained data governance without requiring separate physical tables for every user group. File compaction improves storage layout, schema evolution handles changes to data structures, and data skipping improves query performance. When the requirement is to control which rows are visible to different users, row filtering is the relevant capability.

Question 95

Which Azure Databricks component is primarily responsible for executing Spark-based data engineering workloads?

1. Compute
2. Unity Catalog
3. External location
4. Git repository

Correct Answer: 1

Explanation

Databricks compute provides the processing resources required to execute Spark-based data engineering workloads. Depending on the workload and configuration, compute can support notebooks, jobs, streaming pipelines, SQL workloads, and other data processing activities. Compute configuration can include factors such as runtime version, worker resources, autoscaling, and access mode. Unity Catalog provides governance, external locations define governed cloud storage paths, and Git repositories provide source-code management. Without suitable compute resources, Spark transformations and other processing tasks cannot execute. Therefore, compute is the Azure Databricks component responsible for providing the processing environment in which data engineering workloads run.

Question 96

Which approach is most useful for ensuring that a pipeline can be safely rerun after a failure without creating duplicate business records?

1. Remove all checkpoints
2. Use deterministic keys and idempotent logic
3. Disable Delta transactions
4. Process every historical file repeatedly

Correct Answer: 3

Explanation

Deterministic keys combined with idempotent processing logic allow a pipeline to be rerun without creating unwanted duplicate business records. A stable business key can identify whether a record already exists, while operations such as MERGE can update an existing record or insert it only when appropriate. This design is especially valuable when jobs retry after failures or when the same input can be delivered more than once. Removing checkpoints can reduce streaming recovery capabilities, disabling Delta transactions would weaken reliability, and repeatedly processing historical files increases duplication risk. Idempotent design therefore provides an important foundation for reliable and recoverable data engineering pipelines.

Question 97

Which operation can improve the physical layout of a Delta table by reorganizing data files for more efficient access?

1. GRANT
2. DESCRIBE
3. OPTIMIZE
4. CREATE VIEW

Correct Answer: 3

Explanation

OPTIMIZE improves the physical organization of Delta table data by reorganizing files and reducing inefficient small-file layouts. It can be useful for tables that receive frequent incremental writes, where many small files may accumulate over time. Better file organization can reduce the number of files that query engines need to manage and may improve query performance. GRANT manages permissions, DESCRIBE provides metadata information, and CREATE VIEW creates a reusable query definition. OPTIMIZE does not fundamentally change the logical meaning of the table’s records; instead, it focuses on improving the underlying physical layout. Therefore, OPTIMIZE is appropriate for Delta table maintenance and performance improvement.

Question 98

Which Structured Streaming operation can apply custom batch-oriented processing logic to each micro-batch?

1. foreachBatch
2. USE CATALOG
3. VACUUM
4. SHOW GRANTS

Correct Answer: 1

Explanation

foreachBatch allows a Structured Streaming query to apply custom processing logic to each micro-batch as it is generated. This can be useful when streaming data needs to be written to systems or processed using logic that is easier to express with batch-oriented APIs. For example, a pipeline can use foreachBatch to perform custom writes, execute MERGE operations, or integrate with supported downstream processing patterns. USE CATALOG changes the active catalog context, VACUUM manages obsolete Delta files, and SHOW GRANTS displays permissions. Therefore, foreachBatch is the appropriate Structured Streaming capability when custom logic needs to run against each micro-batch.

Question 99

Which feature helps administrators standardize and control the configuration of Databricks compute resources?

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

Correct Answer: 4

Explanation

A cluster policy can define and restrict permitted compute configuration settings for Databricks resources. Organizations can use policies to standardize items such as node types, runtime versions, autoscaling limits, and other supported configuration parameters. This helps administrators enforce operational and governance requirements while reducing the risk of users creating unsuitable or unnecessarily expensive compute configurations. Time travel provides access to historical Delta table versions, materialized views support reusable stored query results, and Auto Loader handles incremental file ingestion. Cluster policies therefore address compute governance rather than data storage, querying, or ingestion. They are particularly useful in larger environments where consistent compute configuration is required across teams.

Question 100

A data engineer wants to monitor the execution history of scheduled workloads and investigate failed task runs. Which capability should be used?

1. Unity Catalog
2. Job run history
3. External location
4. Volume

Correct Answer: 2

Explanation

Job run history provides information about executions of scheduled or triggered Databricks workloads. Data engineers can use run details to determine whether tasks succeeded or failed, review execution timing, inspect task dependencies, and investigate failures. This information is valuable for operational monitoring and troubleshooting automated data pipelines. Unity Catalog focuses on data governance and permissions, external locations govern access to cloud storage paths, and volumes provide governed file access. Job monitoring should therefore be performed through the workload’s run information and associated task details. Reviewing job history helps teams identify recurring failures, understand pipeline behavior, and maintain reliable scheduled data engineering workloads.