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

 

Question 321

A data engineer needs to ensure that a pipeline does not create duplicate records when the same input batch is accidentally processed more than once. Which design principle should be applied?

1. Partition pruning
2. Data masking
3. Idempotent processing
4. Cluster autoscaling

Correct Answer: 3

Explanation

Idempotent processing ensures that executing the same processing operation multiple times produces the same intended result rather than creating additional duplicate data. This is particularly important in production data pipelines because retries, job reruns, network failures, and operational mistakes can cause the same input to be processed more than once. A pipeline can achieve idempotency through techniques such as deterministic keys, merge operations, checkpointing, or carefully designed write logic. Partition pruning improves query performance, data masking protects sensitive information, and cluster autoscaling changes compute capacity. Therefore, when duplicate results must be avoided during repeated processing, idempotent pipeline design is the appropriate principle.

Question 322

A team wants to automatically increase or decrease the number of workers based on the workload of a Databricks compute resource. Which capability should they configure?

1. Autoscaling
2. Column masking
3. Delta constraints
4. Unity Catalog views

Correct Answer: 1

Explanation

Autoscaling allows compute resources to adjust the number of workers according to workload requirements within configured limits. When demand increases, additional workers can be added to provide more processing capacity. When demand decreases, workers can be removed to reduce unnecessary resource consumption. Autoscaling is especially useful for workloads with variable processing requirements. Column masking is a data-governance feature for protecting sensitive values, Delta constraints help enforce data-quality conditions, and Unity Catalog views provide logical data representations. These features do not dynamically adjust worker capacity. Therefore, when compute capacity needs to adapt to changing workload demand, autoscaling should be configured.

Question 323

A Delta table contains many small files after frequent incremental writes. Which operation can help consolidate files and improve data layout?

1. SHOW GRANTS
2. OPTIMIZE
3. CREATE VIEW
4. DROP SCHEMA

Correct Answer: 2

Explanation

OPTIMIZE can improve Delta table file organization by compacting smaller files into larger files. Frequent incremental writes can produce many small files, which may increase metadata overhead and reduce read efficiency. Optimizing the table can reduce the number of files that query engines need to manage and can improve subsequent query performance. Depending on the table design and workload, additional optimization techniques such as clustering may also be considered. SHOW GRANTS concerns permissions, CREATE VIEW creates a logical SQL object, and DROP SCHEMA removes a schema and its objects according to the applicable options. Therefore, OPTIMIZE is the operation designed to address small-file problems in Delta tables.

Question 324

A data engineer needs to apply the same data transformation logic to each micro-batch of a Structured Streaming query. Which feature is appropriate?

1. foreachBatch
2. VACUUM
3. DESCRIBE DETAIL
4. Cluster policy

Correct Answer: 1

Explanation

foreachBatch allows custom processing logic to be applied to each micro-batch generated by a Structured Streaming query. This is useful when a streaming pipeline needs to reuse batch-oriented operations or write each micro-batch to a destination using custom logic. It can also support patterns such as applying MERGE operations against Delta tables when designed correctly. VACUUM is used for removing obsolete Delta files, DESCRIBE DETAIL provides table metadata, and cluster policies govern compute configuration. These features do not provide custom processing for individual streaming micro-batches. Therefore, foreachBatch is the appropriate capability when each micro-batch requires customized processing logic.

Question 325

An administrator wants to prevent users from creating compute resources with configurations that violate organizational standards. Which feature should be used?

1. Materialized view
2. Watermark
3. Cluster policy
4. Delta history

Correct Answer: 3

Explanation

Cluster policies allow administrators to define approved limits and configuration rules for Databricks compute resources. They can restrict settings such as runtime versions, node types, worker counts, autoscaling ranges, and other supported parameters. This helps organizations standardize compute environments and reduce the risk of uncontrolled resource usage. A materialized view stores query-derived results, a watermark manages late-arriving data in streaming workloads, and Delta history records table transactions. None of these features controls how users configure compute resources. Therefore, when administrators need to enforce organizational standards for compute configuration, a cluster policy is the appropriate mechanism.

Question 326

A streaming aggregation should stop retaining state for events that are older than the configured lateness threshold. Which feature helps control this state?

1. Git integration
2. Watermarking
3. Storage credentials
4. External locations

Correct Answer: 2

Explanation

Watermarking helps manage state in stateful Structured Streaming operations by defining how long the system should wait for late-arriving events based on event-time processing. Once the watermark advances beyond a particular point, records older than the configured threshold can be excluded from certain stateful operations, allowing the engine to eventually clean up related state. This is useful for controlling memory and state-store growth in long-running streaming applications. Git integration manages source-code versions, storage credentials provide authentication to cloud storage, and external locations govern access to defined storage paths. Therefore, watermarking is the appropriate feature for managing state associated with late-arriving events.

Question 327

A company wants to use a storage credential to access cloud storage and then define a governed path that references that storage. Which Unity Catalog object uses the credential to represent the path?

1. External location
2. Temporary view
3. SQL warehouse
4. Materialized view

Correct Answer: 1

Explanation

An external location combines a cloud storage path with a Unity Catalog storage credential to provide governed access to external data. The storage credential defines how Databricks authenticates to the underlying cloud storage, while the external location identifies the specific path that is being governed. Administrators can then apply permissions to control access to that location. Temporary views are session-level logical objects, SQL warehouses provide SQL compute, and materialized views contain persisted query results. These objects do not represent governed external storage paths. Therefore, when a storage credential is used to govern access to a specific cloud storage location, an external location is the appropriate Unity Catalog object.

Question 328

A pipeline receives files with evolving JSON structures, and new fields may appear over time. Which Auto Loader capability can help manage this situation?

1. Schema evolution
2. Query caching
3. Cluster termination
4. Table ownership

Correct Answer: 1

Explanation

Auto Loader supports schema management capabilities that can help pipelines handle changes in incoming file structures. Schema evolution can allow newly introduced fields to be incorporated into the evolving input schema according to the configured behavior. This is useful when source systems add attributes over time and manually updating the ingestion process for every change would be inefficient. Query caching is a performance mechanism, cluster termination concerns compute lifecycle, and table ownership controls administrative privileges. None of these features directly manages changing input schemas. Therefore, when an Auto Loader pipeline needs to accommodate newly appearing fields, schema evolution is an appropriate capability to consider.

Question 329

A data engineer wants to prevent a streaming query from reprocessing previously handled source data after a restart. Which mechanism is most important?

1. Cluster policy
2. Checkpointing
3. Column mask
4. SQL warehouse

Correct Answer: 2

Explanation

Checkpointing stores the progress and state information required for a Structured Streaming query to recover after a restart. When a query is restarted using the same appropriate checkpoint location, the streaming engine can use the recorded progress to continue from the previously processed position rather than treating all prior input as new. Checkpointing is therefore an important part of reliable streaming pipeline design. A cluster policy controls compute configuration, a column mask protects sensitive data, and a SQL warehouse provides SQL compute resources. None of these mechanisms records streaming progress. Therefore, checkpointing is the appropriate mechanism for recovering a streaming query without unnecessarily reprocessing previously handled input.

Question 330

A data engineer wants queries to skip unnecessary data files by using information about the values stored in those files. Which performance technique is relevant?

1. Data skipping
2. User provisioning
3. Secret rotation
4. Job scheduling

Correct Answer: 1

Explanation

Data skipping allows the query engine to avoid reading files that cannot contain records matching the query conditions. Metadata about file contents can help determine which files are relevant, reducing unnecessary I/O and improving query performance. This is particularly useful for large Delta datasets where a query may otherwise need to inspect many files. User provisioning controls access to accounts, secret rotation manages credentials, and job scheduling determines when workflows execute. None of these features directly reduces the amount of data read during a query. Therefore, when the goal is to avoid scanning irrelevant files based on available data information, data skipping is the relevant performance technique.

Question 331

A production job requires a specific Python package that is not included in the default Databricks runtime. What should the data engineer configure?

1. A library dependency
2. A watermark
3. A row filter
4. A Delta history query

Correct Answer: 1

Explanation

A library dependency allows a Databricks workload to use packages that are not included in the standard runtime environment. Dependencies can be configured at an appropriate scope depending on the workload and deployment model. Managing dependencies explicitly helps ensure that required Python packages are available when a job executes and supports more reproducible environments. A watermark is used by streaming workloads to manage late-arriving data, a row filter controls data visibility, and Delta history provides information about table transactions. These features do not install or provide Python packages. Therefore, when a production job requires an additional Python package, the appropriate library dependency should be configured.

Question 332

A data engineer wants to restrict rows returned from a table based on the identity of the user querying the table. Which governance capability is designed for this requirement?

1. Row filtering
2. Autoscaling
3. OPTIMIZE
4. Git branching

Correct Answer: 1

Explanation

Row filtering can restrict which records are visible to users based on defined access rules. This is useful for implementing row-level security scenarios where different users or groups should see different subsets of the same table. For example, access rules can be designed so that regional users see only records associated with their authorized region. Autoscaling controls compute capacity, OPTIMIZE improves Delta file organization, and Git branching manages source-code versions. These capabilities do not determine which table rows a user is allowed to see. Therefore, when access must be restricted at the row level according to user identity or other conditions, row filtering is the appropriate governance capability.

Question 333

A pipeline needs to identify and process only files that have not previously been discovered by the ingestion process. Which Auto Loader behavior supports this requirement?

1. Incremental file discovery
2. Table ownership
3. SQL result caching
4. Cluster policy enforcement

Correct Answer: 1

Explanation

Auto Loader is designed for incremental file ingestion and maintains information about files that have already been discovered and processed. This enables a pipeline to focus on newly arriving files rather than repeatedly scanning and processing the entire historical directory as new input. Incremental discovery is particularly useful for large cloud-storage locations where files continuously arrive over time. Table ownership controls administrative permissions, SQL result caching concerns query performance, and cluster policy enforcement controls compute configuration. None of these features provides incremental file discovery. Therefore, when the ingestion process needs to identify newly arriving files efficiently, Auto Loader’s incremental discovery capabilities are appropriate.

Question 334

A team wants to maintain separate development and production versions of its Databricks project while allowing changes to be reviewed before deployment. Which practice is most appropriate?

1. Disable version control
2. Use Git branches and pull requests
3. Increase table retention
4. Create additional storage credentials

Correct Answer: 2

Explanation

Git branches and pull requests provide a structured workflow for developing, reviewing, and promoting changes. Developers can work on isolated branches without immediately changing the production version. Once changes are ready, a pull request can be used for review, testing, and approval before merging them into the branch used for deployment. This supports traceability and reduces the risk of deploying unreviewed changes. Increasing table retention affects data history, while additional storage credentials concern access to external storage. Disabling version control removes important change-management capabilities. Therefore, Git branches combined with pull requests are appropriate when development and production versions need controlled promotion.

Question 335

A data engineer needs to ensure that a numeric column in a Delta table never contains negative values. Which feature can enforce this requirement at the table level?

1. Delta constraint
2. SQL warehouse
3. External location
4. Git folder

Correct Answer: 1

Explanation

Delta constraints can enforce data-quality conditions on table writes. A check constraint can specify a rule such as requiring a numeric column to be greater than or equal to zero. When an incoming record violates the defined constraint, the write can fail instead of allowing invalid data into the table. This provides an enforcement mechanism directly associated with the table rather than relying entirely on downstream validation. SQL warehouses provide compute for SQL workloads, external locations govern cloud storage paths, and Git folders support source-code organization. Therefore, when a table must enforce a condition such as preventing negative values, a Delta constraint is an appropriate solution.

Question 336

A workflow contains multiple independent tasks that can safely execute at the same time. What approach can reduce unnecessary workflow waiting?

1. Serialize every task
2. Configure independent tasks without unnecessary dependencies
3. Disable task monitoring
4. Increase table retention

Correct Answer: 2

Explanation

Independent tasks that do not depend on each other’s outputs can often execute in parallel. Avoiding unnecessary dependencies allows the workflow scheduler to start eligible tasks without waiting for unrelated tasks to finish. This can reduce total workflow duration and improve utilization of available compute resources. Dependencies should still be defined whenever a task genuinely requires the result of another task. Serializing every task can unnecessarily increase execution time, while disabling monitoring removes useful operational visibility. Table retention concerns historical data rather than workflow execution. Therefore, when tasks are genuinely independent, configuring them without unnecessary dependencies can enable parallel execution and improve workflow efficiency.

Question 337

A data engineer needs to inspect metadata such as the location, format, and other properties of a Delta table. Which command is appropriate?

1. DESCRIBE DETAIL
2. SHOW GRANTS
3. VACUUM
4. MERGE

Correct Answer: 1

Explanation

DESCRIBE DETAIL provides detailed metadata about a Delta table. The returned information can include properties such as table location, format, size-related information, and other useful metadata depending on the environment and table configuration. This command is helpful during troubleshooting, performance investigations, and operational analysis when an engineer needs more information than the table schema alone provides. SHOW GRANTS focuses on permissions, VACUUM removes obsolete files, and MERGE synchronizes source and target records. None of these operations is intended primarily for examining table metadata. Therefore, when detailed Delta table properties need to be inspected, DESCRIBE DETAIL is the appropriate command.

Question 338

A streaming pipeline should execute a processing trigger at a fixed interval rather than continuously processing as soon as data is available. Which configuration concept is relevant?

1. Trigger interval
2. Table ownership
3. Cluster policy
4. Column masking

Correct Answer: 1

Explanation

A trigger interval controls when Structured Streaming processing is initiated for available data according to the selected trigger configuration. Using a fixed processing interval can be useful when workloads should run periodically instead of processing continuously whenever new records arrive. The choice of trigger depends on workload requirements, latency expectations, and resource considerations. Table ownership determines administrative control, cluster policies govern compute configuration, and column masking protects sensitive values. These features do not determine the execution timing of streaming micro-batches. Therefore, when a streaming workload needs processing to occur at a defined recurring interval, the trigger interval configuration is relevant.

Question 339

A data engineer wants to compare the current contents of a Delta table with an earlier version without permanently restoring the table. Which capability can be used?

1. Delta time travel
2. Cluster autoscaling
3. External location
4. Secret scope

Correct Answer: 1

Explanation

Delta time travel allows users to access previous versions of a Delta table using historical version or timestamp information, subject to the available retention period. This is useful for auditing, troubleshooting, investigating unexpected changes, and comparing historical data with the current table state. Accessing an earlier version does not necessarily mean replacing the current table contents. Cluster autoscaling changes compute capacity, an external location governs access to cloud storage, and secret scopes provide mechanisms for managing secrets. None of these features provides historical table querying. Therefore, when a previous Delta table state needs to be inspected or compared without permanently changing the current state, time travel is appropriate.

Question 340

A data engineer needs to pass a value such as an environment name from a workflow configuration into a job task. Which capability should be considered?

1. Job parameters
2. Delta constraints
3. File compaction
4. Data masking

Correct Answer: 1

Explanation

Job parameters allow values to be supplied to workflow tasks so that the same job logic can operate with different runtime inputs. For example, an environment parameter can identify development, testing, or production behavior without requiring separate copies of the underlying code. Parameters can improve reusability and make workflows easier to configure and maintain. Delta constraints enforce data-quality rules, file compaction improves table file organization, and data masking protects sensitive values. These features do not provide a mechanism for passing runtime configuration values into job tasks. Therefore, when a workflow needs to supply values such as an environment name to a task, job parameters are appropriate.