Microsoft DP-900 Practice Test 106
Question 1050
A team mistakenly filters SUM totals using a row-level WHERE condition. What is the MOST appropriate next validation or correction?
- Test a group with many small charges whose total exceeds threshold.
- Test connectivity from allowed and denied network locations.
- Test parameter validation and execution permissions.
- Add a WHERE Amount greater than 1000 and assume it finds all customers spending over 1000 in total.
Explanation
Correct Answer
A:Test a group with many small charges whose total exceeds threshold. This tests the intended result: only groups whose computed total crosses the threshold remain.
Incorrect Answers
B:Test connectivity from allowed and denied network locations. It tests Azure SQL firewall and private endpoint rather than the critical SQL HAVING after grouping behavior. HAVING filters aggregated groups while WHERE filters source rows.
C:Test parameter validation and execution permissions. It tests Stored procedure reusable server operation rather than the critical SQL HAVING after grouping behavior. HAVING filters aggregated groups while WHERE filters source rows.
D:Add a WHERE Amount greater than 1000 and assume it finds all customers spending over 1000 in total. This unsafe shortcut leaves the SQL HAVING after grouping problem unresolved. Filtering individual rows cannot replace a condition on group totals.
Question 1051
A team needs to build reports locally before publishing to the service. What is the MOST appropriate next validation or correction?
- Check selected locations against known coordinates and category labels.
- Rename a spreadsheet .pbix and assume it becomes a model with working reports.
- Simulate upstream failures and inspect pipeline run dependencies.
- Validate source connection types and loaded sample values.
Explanation
Correct Answer
D:Validate source connection types and loaded sample values. This tests the intended result: imported tables and approved relationships support intended reports.
Incorrect Answers
A:Check selected locations against known coordinates and category labels. It tests Map visual geographic measures rather than the critical Power BI Desktop data import behavior. Power BI Desktop is an authoring tool for data modeling and reports.
B:Rename a spreadsheet .pbix and assume it becomes a model with working reports. This unsafe shortcut leaves the Power BI Desktop data import problem unresolved. Assuming a file import automatically fixes wrong data types can misstate totals.
C:Simulate upstream failures and inspect pipeline run dependencies. It tests Fabric Data Factory pipeline orchestration rather than the critical Power BI Desktop data import behavior. Power BI Desktop is an authoring tool for data modeling and reports.
Question 1052
Architects must choose between file-first and SQL-first Fabric analytics. What is the MOST appropriate next validation or correction?
- Inspect preview transformations and loaded destination records.
- Kill a test consumer and reconcile events after restart.
- Assume Lakehouse and Warehouse are exactly identical because both use OneLake.
- Prototype one Spark pipeline and one reporting query.
Explanation
Correct Answer
D:Prototype one Spark pipeline and one reporting query. This tests the intended result: the chosen experience supports required processing and consumption without undue copying.
Incorrect Answers
A:Inspect preview transformations and loaded destination records. It tests Fabric Dataflow Gen2 transformations rather than the critical Lakehouse versus Warehouse selection behavior. Fabric workloads have complementary analytics designs rather than one universal model.
B:Kill a test consumer and reconcile events after restart. It tests Streaming checkpoints and replay rather than the critical Lakehouse versus Warehouse selection behavior. Fabric workloads have complementary analytics designs rather than one universal model.
C:Assume Lakehouse and Warehouse are exactly identical because both use OneLake. This unsafe shortcut leaves the Lakehouse versus Warehouse selection problem unresolved. Picking a SQL-first store for raw notebook-only work may cause extra conversions.
Question 1053
Logistics reports delivery volume by state and region. After selecting the proper Map visual geographic measures approach, which observation gives the clearest evidence of success?
- Slices sum to the intended whole and readers can compare shares.
- Output windows contain correct event aggregates under steady ingestion.
- Source records arrive with tracked counts and errors for each system.
- Locations resolve to intended regions and values match actual places.
Explanation
Correct Answer
D:Locations resolve to intended regions and values match actual places. This is the directly observable result for Map visual geographic measures; geographic visuals require accurate location encoding and appropriate map granularity.
Incorrect Answers
A:Slices sum to the intended whole and readers can compare shares. This supports Pie donut chart proportion caution, not confirmation of Map visual geographic measures. The relevant evidence is that locations resolve to intended regions and values match actual places.
B:Output windows contain correct event aggregates under steady ingestion. This supports Azure Stream Analytics SQL-like streaming, not confirmation of Map visual geographic measures. The relevant evidence is that locations resolve to intended regions and values match actual places.
C:Source records arrive with tracked counts and errors for each system. This supports Analytical ingestion from operational systems, not confirmation of Map visual geographic measures. The relevant evidence is that locations resolve to intended regions and values match actual places.
Question 1054
A legacy workload updates fixed byte ranges rather than replacing an entire file. What is the MOST appropriate next validation or correction?
- Assume a page blob is an ordinary sequential append log.
- Check method scope expiry and recipient handling with test requests.
- Verify supported alignment and read-write range operations.
- Test multi-edge routes and direction filters on sample graph.
Explanation
Correct Answer
C:Verify supported alignment and read-write range operations. This tests the intended result: expected ranges update while other pages remain unchanged.
Incorrect Answers
A:Assume a page blob is an ordinary sequential append log. This unsafe shortcut leaves the Page blob random access problem unresolved. Using append-only semantics for virtual disk page updates is inappropriate.
B:Check method scope expiry and recipient handling with test requests. It tests Blob SAS delegated access rather than the critical Page blob random access behavior. Page blobs are suited to random I/O scenarios and VHD backing.
D:Test multi-edge routes and direction filters on sample graph. It tests Cosmos DB API for Apache Gremlin graph rather than the critical Page blob random access behavior. Page blobs are suited to random I/O scenarios and VHD backing.
Question 1055
Large analytical scans are slow because files are very small. After selecting the proper Data engineer partition and file layout approach, which observation gives the clearest evidence of success?
- Analytical reads avoid excessive file enumeration and irrelevant partitions.
- Users can answer agreed questions from accurate accessible visuals.
- Reviewers trace a reported measure back to approved inputs and rules.
- The model supports intended joins or nested attribute reads without excessive conversions.
Explanation
Correct Answer
A:Analytical reads avoid excessive file enumeration and irrelevant partitions. This is the directly observable result for Data engineer partition and file layout; data engineering includes practical physical data layout decisions.
Incorrect Answers
B:Users can answer agreed questions from accurate accessible visuals. This supports Data analyst report design, not confirmation of Data engineer partition and file layout. The relevant evidence is that analytical reads avoid excessive file enumeration and irrelevant partitions.
C:Reviewers trace a reported measure back to approved inputs and rules. This supports Data lineage from source to report, not confirmation of Data engineer partition and file layout. The relevant evidence is that analytical reads avoid excessive file enumeration and irrelevant partitions.
D:The model supports intended joins or nested attribute reads without excessive conversions. This supports Relational rows versus document attributes, not confirmation of Data engineer partition and file layout. The relevant evidence is that analytical reads avoid excessive file enumeration and irrelevant partitions.
Question 1056
A catalog API adds nested attributes that differ between product categories. Which principle BEST explains the correct Semi structured JSON documents design decision?
- ADLS Gen2 adds data-lake semantics on Azure Blob storage.
- Transactional workloads emphasize consistent low-latency create update delete operations.
- JSON arrays hold ordered elements whereas objects group named properties.
- Semi structured data has organizational markers without requiring identical columns in every record.
Explanation
Correct Answer
D:Semi structured data has organizational markers without requiring identical columns in every record. In this case the applicable decision is to represent variable properties in JSON documents with parseable key value structures.
Incorrect Answers
A:ADLS Gen2 adds data-lake semantics on Azure Blob storage. That concept explains Azure Data Lake Storage Gen2 analytics landing; the situation instead concerns Semi structured JSON documents. Assuming every JSON object has a fixed relational schema causes null errors and brittle ingestion.
B:Transactional workloads emphasize consistent low-latency create update delete operations. That concept explains OLTP atomic purchase transaction; the situation instead concerns Semi structured JSON documents. Assuming every JSON object has a fixed relational schema causes null errors and brittle ingestion.
C:JSON arrays hold ordered elements whereas objects group named properties. That concept explains JSON arrays and nested objects; the situation instead concerns Semi structured JSON documents. Assuming every JSON object has a fixed relational schema causes null errors and brittle ingestion.
Question 1057
A production app runs steady high-volume reads requiring predictable reserved capacity. What is the MOST appropriate next validation or correction?
- Assume serverless removes partition-key requirements and query resource costs.
- Simulate item creation/update and inspect processor checkpoints.
- Check method scope expiry and recipient handling with test requests.
- Benchmark bursts and steady demand for each candidate option.
Explanation
Correct Answer
D:Benchmark bursts and steady demand for each candidate option. This tests the intended result: billing and throttling behavior match measured usage and service limits.
Incorrect Answers
A:Assume serverless removes partition-key requirements and query resource costs. This unsafe shortcut leaves the Cosmos DB serverless versus provisioned throughput problem unresolved. Choosing serverless without checking workload scale constraints may be costly.
B:Simulate item creation/update and inspect processor checkpoints. It tests Cosmos DB change feed updates rather than the critical Cosmos DB serverless versus provisioned throughput behavior. Cosmos DB consumption and reserved throughput models suit different demand patterns.
C:Check method scope expiry and recipient handling with test requests. It tests Blob SAS delegated access rather than the critical Cosmos DB serverless versus provisioned throughput behavior. Cosmos DB consumption and reserved throughput models suit different demand patterns.
Question 1058
A team publishes an interactive multi-page analytical report to coworkers. What is the MOST appropriate next validation or correction?
- Pin a screenshot to a wall and call it a live Power BI dashboard with governed sharing.
- Test hierarchy navigation and chronological sorting.
- Inject interleaved customer events across partitions and inspect offsets.
- Test app audience membership and semantic model access.
Explanation
Correct Answer
D:Test app audience membership and semantic model access. This tests the intended result: intended users can view the right report pages or dashboard tiles with governed permissions.
Incorrect Answers
A:Pin a screenshot to a wall and call it a live Power BI dashboard with governed sharing. This unsafe shortcut leaves the Power BI reports dashboards and sharing problem unresolved. Sharing one report URL does not automatically authorize every underlying data source.
B:Test hierarchy navigation and chronological sorting. It tests Drill-down hierarchy for time rather than the critical Power BI reports dashboards and sharing behavior. Reports and dashboards have different presentation and interaction roles.
C:Inject interleaved customer events across partitions and inspect offsets. It tests Event streams partitioning and key order rather than the critical Power BI reports dashboards and sharing behavior. Reports and dashboards have different presentation and interaction roles.
Question 1059
A system receives daily data exports that need validation by morning. Which principle BEST explains the correct Batch scheduled processing design decision?
- Descriptive analytics summarizes history; diagnostic analytics explores causes.
- Batch work processes bounded collections at selected times.
- Data engineering includes practical physical data layout decisions.
- Table Storage is a key-attribute NoSQL store not a relational SQL engine.
Explanation
Correct Answer
B:Batch work processes bounded collections at selected times. In this case the applicable decision is to run a scheduled batch ingestion with clear checkpoints and retries.
Incorrect Answers
A:Descriptive analytics summarizes history; diagnostic analytics explores causes. That concept explains Descriptive versus diagnostic analytics; the situation instead concerns Batch scheduled processing. Lack of idempotency can duplicate rows on rerun.
C:Data engineering includes practical physical data layout decisions. That concept explains Data engineer partition and file layout; the situation instead concerns Batch scheduled processing. Lack of idempotency can duplicate rows on rerun.
D:Table Storage is a key-attribute NoSQL store not a relational SQL engine. That concept explains Azure Table Storage key-value entities; the situation instead concerns Batch scheduled processing. Lack of idempotency can duplicate rows on rerun.
Question 1060
A team wants each event counted in exactly one fixed five-minute bucket. What is the MOST appropriate next validation or correction?
- Create a row for each second and call it a five-minute tumbling window.
- Test events at minute boundaries and compare bucket counts.
- Validate file paths lineage and SQL measure reconciliation.
- Verify discoverability and access for distinct Fabric users.
Explanation
Correct Answer
B:Test events at minute boundaries and compare bucket counts. This tests the intended result: events are assigned to nonoverlapping fixed-duration windows.
Incorrect Answers
A:Create a row for each second and call it a five-minute tumbling window. This unsafe shortcut leaves the Tumbling time windows problem unresolved. Using overlapping hopping windows can count an event in several summaries.
C:Validate file paths lineage and SQL measure reconciliation. It tests Structured warehouse and unstructured lake coexistence rather than the critical Tumbling time windows behavior. Tumbling windows have contiguous nonoverlapping time intervals.
D:Verify discoverability and access for distinct Fabric users. It tests Fabric Real-Time hub stream discovery rather than the critical Tumbling time windows behavior. Tumbling windows have contiguous nonoverlapping time intervals.
