McAfee-Secured Website

Exam Bundle

Exam Code: DP-900

Exam Name Microsoft Azure Data Fundamentals

Certification Provider: Microsoft

Corresponding Certification: Microsoft Certified: Azure Data Fundamentals

Microsoft DP-900 Bundle $25.00

100% Microsoft DP-900 Guaranteed Success With Testking Training Materials

Get Real DP-900 Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

  • Questions & Answers

    DP-900 Questions & Answers

    332 Questions & Answers

    The ultimate exam preparation tool, Questions & Answers mirrors the actual DP-900 exam while allowing you to hone and measure every aspect of your test-taking ability.

  • DP-900 Video Course

    DP-900 Video Course

    32 Video Lectures

    DP-900 Video Course is developed by Microsoft Professionals to help you pass the DP-900 exam.

    Description

    This course will improve your knowledge and skills required to pass Microsoft Azure Data Fundamentals exam.

Microsoft Azure Data Fundamentals (DP-900) Practice Tests

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Total Cost: $164.98
Bundle Price: $139.98

Purchase Individually

  • Questions & Answers

    Questions & Answers

    332 Questions

    $124.99
  • DP-900 Video Course

    Video Course

    32 Video Lectures

    $39.99

Microsoft DP-900 Practice Test 63

Question 620

A data architect recommends this approach: Use an ETL stage that transforms or masks the data before loading. Which situation most clearly calls for that approach?

  1. A source system exports sensitive information that must be masked before records enter the reporting database.
  2. A team building analytics natively in Microsoft Fabric needs managed data movement and transformation orchestration.
  3. A reporting department needs a managed relational analytical warehouse with familiar SQL-oriented access in Microsoft Fabric.
  4. A data science team can wait until 2 AM to refresh a large feature dataset from a day's events.

Explanation

Correct Answer

A:

This situation calls for ETL before landing. The reporting target receives already-prepared information.

Incorrect Answers

B:

The other situation is better matched to Use Data Factory capabilities within Microsoft Fabric. Integrated Fabric pipelines and dataflows can prepare data for downstream workloads.

C:

The other situation is better matched to Use a Fabric Data Warehouse. Analysts can query curated dimensional and fact tables with SQL semantics.

D:

The other situation is better matched to Run the transformation as a scheduled batch analytics job. A bounded set of accumulated events is processed at the planned interval.

Question 621

An hourly pipeline currently re-reads a complete ten-year sales dataset even when only a few rows change. Which data concept, Azure service, or design decision most directly satisfies the requirement?

  1. Use Azure Databricks with Apache Spark for the transformation workload.
  2. Store the raw extracts in a bronze or raw lake zone.
  3. Partition analytical data using an appropriate time-based partition strategy.
  4. Load new or changed records incrementally using a reliable change marker.

Explanation

Correct Answer

D:

The requirement concerns Incremental loading. Repeated ingestion can process a smaller changed subset rather than the full history.

Incorrect Answers

A:

That action instead addresses Databricks Spark: Parallel distributed processing handles larger-than-single-server transformations. It does not directly satisfy the described Incremental loading need.

B:

That action instead addresses Landing raw data: Original records remain available for replay and audit of downstream transformations. It does not directly satisfy the described Incremental loading need.

C:

That action instead addresses Analytical partitioning: Queries can scan fewer partitions when predicates match the partition layout. It does not directly satisfy the described Incremental loading need.

Question 622

A data professional makes this decision: Load new or changed records incrementally using a reliable change marker. Which outcome is the most direct expected result?

  1. Repeated ingestion can process a smaller changed subset rather than the full history.
  2. Queries can scan fewer partitions when predicates match the partition layout.
  3. Parallel distributed processing handles larger-than-single-server transformations.
  4. Original records remain available for replay and audit of downstream transformations.

Explanation

Correct Answer

A:

This is the expected outcome for Incremental loading: Repeated ingestion can process a smaller changed subset rather than the full history.

Incorrect Answers

B:

This result would follow the different Analytical partitioning decision (Partition analytical data using an appropriate time-based partition strategy.) rather than the specified design.

C:

This result would follow the different Databricks Spark decision (Use Azure Databricks with Apache Spark for the transformation workload.) rather than the specified design.

D:

This result would follow the different Landing raw data decision (Store the raw extracts in a bronze or raw lake zone.) rather than the specified design.

Question 623

An hourly pipeline currently re-reads a complete ten-year sales dataset even when only a few rows change. Someone instead proposes: Partition analytical data using an appropriate time-based partition strategy. Which alternative is the best correction for the original requirement?

  1. Use Azure Databricks with Apache Spark for the transformation workload.
  2. Load new or changed records incrementally using a reliable change marker.
  3. Partition analytical data using an appropriate time-based partition strategy.
  4. Store the raw extracts in a bronze or raw lake zone.

Explanation

Correct Answer

B:

This correction addresses Incremental loading. Repeated ingestion can process a smaller changed subset rather than the full history.

Incorrect Answers

A:

This choice is for Databricks Spark, not the specific requirement. Parallel distributed processing handles larger-than-single-server transformations.

C:

This choice is for Analytical partitioning, not the specific requirement. Queries can scan fewer partitions when predicates match the partition layout.

D:

This choice is for Landing raw data, not the specific requirement. Original records remain available for replay and audit of downstream transformations.

Question 624

A data architect recommends this approach: Load new or changed records incrementally using a reliable change marker. Which situation most clearly calls for that approach?

  1. A 20-terabyte event table is typically filtered to only the last two days.
  2. A pipeline must transform terabytes of raw JSON in distributed compute jobs.
  3. An hourly pipeline currently re-reads a complete ten-year sales dataset even when only a few rows change.
  4. A data platform must preserve unmodified supplier extracts before any cleansing or enrichment.

Explanation

Correct Answer

C:

This situation calls for Incremental loading. Repeated ingestion can process a smaller changed subset rather than the full history.

Incorrect Answers

A:

The other situation is better matched to Partition analytical data using an appropriate time-based partition strategy. Queries can scan fewer partitions when predicates match the partition layout.

B:

The other situation is better matched to Use Azure Databricks with Apache Spark for the transformation workload. Parallel distributed processing handles larger-than-single-server transformations.

D:

The other situation is better matched to Store the raw extracts in a bronze or raw lake zone. Original records remain available for replay and audit of downstream transformations.

Question 625

A failed nightly load went unnoticed, leaving dashboards with yesterday's stale totals. Which data concept, Azure service, or design decision most directly satisfies the requirement?

  1. Configure pipeline run monitoring and alerts for failed activities.
  2. Use distributed or warehouse-optimized analytical aggregation compute.
  3. Apply quality and schema normalization in a curated silver layer.
  4. Use distributed Spark DataFrame transformations.

Explanation

Correct Answer

A:

The requirement concerns Data pipeline monitoring. Operators can detect and investigate ingestion failures before reporting deadlines.

Incorrect Answers

B:

That action instead addresses Large-scale aggregation: Wide historical scans and group calculations can run on scalable resources. It does not directly satisfy the described Data pipeline monitoring need.

C:

That action instead addresses Standardized silver: Cleaned and standardized records can be reused by multiple analytical consumers. It does not directly satisfy the described Data pipeline monitoring need.

D:

That action instead addresses Spark dataframe: The join can be executed across the cluster rather than on one desktop machine. It does not directly satisfy the described Data pipeline monitoring need.

Question 626

A data professional makes this decision: Configure pipeline run monitoring and alerts for failed activities. Which outcome is the most direct expected result?

  1. Wide historical scans and group calculations can run on scalable resources.
  2. Operators can detect and investigate ingestion failures before reporting deadlines.
  3. Cleaned and standardized records can be reused by multiple analytical consumers.
  4. The join can be executed across the cluster rather than on one desktop machine.

Explanation

Correct Answer

B:

This is the expected outcome for Data pipeline monitoring: Operators can detect and investigate ingestion failures before reporting deadlines.

Incorrect Answers

A:

This result would follow the different Large-scale aggregation decision (Use distributed or warehouse-optimized analytical aggregation compute.) rather than the specified design.

C:

This result would follow the different Standardized silver decision (Apply quality and schema normalization in a curated silver layer.) rather than the specified design.

D:

This result would follow the different Spark dataframe decision (Use distributed Spark DataFrame transformations.) rather than the specified design.

Question 627

A failed nightly load went unnoticed, leaving dashboards with yesterday's stale totals. Someone instead proposes: Use distributed or warehouse-optimized analytical aggregation compute. Which alternative is the best correction for the original requirement?

  1. Use distributed Spark DataFrame transformations.
  2. Use distributed or warehouse-optimized analytical aggregation compute.
  3. Configure pipeline run monitoring and alerts for failed activities.
  4. Apply quality and schema normalization in a curated silver layer.

Explanation

Correct Answer

C:

This correction addresses Data pipeline monitoring. Operators can detect and investigate ingestion failures before reporting deadlines.

Incorrect Answers

A:

This choice is for Spark dataframe, not the specific requirement. The join can be executed across the cluster rather than on one desktop machine.

B:

This choice is for Large-scale aggregation, not the specific requirement. Wide historical scans and group calculations can run on scalable resources.

D:

This choice is for Standardized silver, not the specific requirement. Cleaned and standardized records can be reused by multiple analytical consumers.

Question 628

A data architect recommends this approach: Configure pipeline run monitoring and alerts for failed activities. Which situation most clearly calls for that approach?

  1. A data engineer needs to join two large datasets across a distributed cluster and project selected attributes.
  2. A retailer needs to group billions of click records by channel and week for recurring analysis.
  3. An analytics pipeline receives inconsistent customer field names and duplicate transaction IDs.
  4. A failed nightly load went unnoticed, leaving dashboards with yesterday's stale totals.

Explanation

Correct Answer

D:

This situation calls for Data pipeline monitoring. Operators can detect and investigate ingestion failures before reporting deadlines.

Incorrect Answers

A:

The other situation is better matched to Use distributed Spark DataFrame transformations. The join can be executed across the cluster rather than on one desktop machine.

B:

The other situation is better matched to Use distributed or warehouse-optimized analytical aggregation compute. Wide historical scans and group calculations can run on scalable resources.

C:

The other situation is better matched to Apply quality and schema normalization in a curated silver layer. Cleaned and standardized records can be reused by multiple analytical consumers.

Question 629

An upstream JSON feed adds an optional attribute and the pipeline must tolerate legitimate source changes. Which data concept, Azure service, or design decision most directly satisfies the requirement?

  1. Publish a governed shared semantic model for the business measures.
  2. Publish approved business aggregates in a gold serving layer.
  3. Use a Delta table format where supported.
  4. Design ingestion transformations with explicit handling of supported schema drift.

Explanation

Correct Answer

D:

The requirement concerns Analytics schema drift. Allowed new fields are processed without silently corrupting the curated dataset.

Incorrect Answers

A:

That action instead addresses Curated semantic layer: Downstream consumers can reuse consistently defined metrics. It does not directly satisfy the described Analytics schema drift need.

B:

That action instead addresses Gold business data: Reporting users consume curated tables designed around shared business measures. It does not directly satisfy the described Analytics schema drift need.

C:

That action instead addresses Delta Lake table: A transaction log provides managed table state over the underlying analytical files. It does not directly satisfy the described Analytics schema drift need.

Question 630

A data professional makes this decision: Design ingestion transformations with explicit handling of supported schema drift. Which outcome is the most direct expected result?

  1. Allowed new fields are processed without silently corrupting the curated dataset.
  2. Downstream consumers can reuse consistently defined metrics.
  3. A transaction log provides managed table state over the underlying analytical files.
  4. Reporting users consume curated tables designed around shared business measures.

Explanation

Correct Answer

A:

This is the expected outcome for Analytics schema drift: Allowed new fields are processed without silently corrupting the curated dataset.

Incorrect Answers

B:

This result would follow the different Curated semantic layer decision (Publish a governed shared semantic model for the business measures.) rather than the specified design.

C:

This result would follow the different Delta Lake table decision (Use a Delta table format where supported.) rather than the specified design.

D:

This result would follow the different Gold business data decision (Publish approved business aggregates in a gold serving layer.) rather than the specified design.