Pass Microsoft DP-800 Exam in First Attempt Guaranteed!
Get 100% Latest Exam Questions, Accurate & Verified Answers!
30 Days Free Updates, Instant Download!
Microsoft DP-800 Practice Test 21
Question 200
An AI coding assistant proposes rewriting a billing query, but the team has no regression fixture for null taxes and partial refunds. Which observation is the strongest evidence for the reported problem?
Explanation
Correct Answer
The faster version silently drops refunds because a new INNER JOIN filters unmatched rows. Correctness regression testing is essential when AI assistance alters set and join semantics.
Incorrect Answers
Authorized staff cannot access the expected assistant despite seeing it in a marketing demonstration. Correctness regression testing is essential when AI assistance alters set and join semantics. The faster version silently drops refunds because a new INNER JOIN filters unmatched rows.
The query explanation shows proposed changes but the database still contains the original stored procedure. Correctness regression testing is essential when AI assistance alters set and join semantics. The faster version silently drops refunds because a new INNER JOIN filters unmatched rows.
A proposed query runs faster in a developer test but returns records belonging to other tenants. Correctness regression testing is essential when AI assistance alters set and join semantics. The faster version silently drops refunds because a new INNER JOIN filters unmatched rows.
Question 201
A T-SQL workflow builds a prompt for an external language model but includes every row and confidential identifier regardless of question scope. Which next step most directly resolves this SQL solution issue?
Explanation
Correct Answer
Use sp_invoke_external_rest_endpoint only with an approved network and identity configuration and a minimal validated prompt payload. External REST calls from SQL must minimize disclosed data and enforce endpoint security.
Incorrect Answers
Record authorized prompt/source version metadata, model identity, and approval outcome with appropriate privacy retention. External REST calls from SQL must minimize disclosed data and enforce endpoint security. Use sp_invoke_external_rest_endpoint only with an approved network and identity configuration and a minimal validated prompt payload.
Treat retrieved content as untrusted evidence and never grant it authority to execute SQL actions. External REST calls from SQL must minimize disclosed data and enforce endpoint security. Use sp_invoke_external_rest_endpoint only with an approved network and identity configuration and a minimal validated prompt payload.
Convert the approved structured rows to valid JSON with supported SQL JSON construction functions. External REST calls from SQL must minimize disclosed data and enforce endpoint security. Use sp_invoke_external_rest_endpoint only with an approved network and identity configuration and a minimal validated prompt payload.
Question 202
A SQL project tracks application tables but lookup status codes are manually edited in production, causing different values across environments. Which test result would demonstrate the fix worked?
Explanation
Correct Answer
A clean rebuild and incremental deployment produce identical authorized lookup mappings.
Incorrect Answers
A tested failing pull request is blocked and approved changes follow the defined merge path. A clean rebuild and incremental deployment produce identical authorized lookup mappings.
Deployment preserves required values or uses a documented approved transformation without silent truncation. A clean rebuild and incremental deployment produce identical authorized lookup mappings.
A dry run validates the intended server/database identity and production release targets only that environment. A clean rebuild and incremental deployment produce identical authorized lookup mappings.
Question 203
A SQL-centered maintenance application stores images and text descriptions of equipment failures and the team attempts to embed images with a text-only pipeline. Which concept should guide the database developer in this situation?
Explanation
Correct Answer
Model modalities and preprocessing must match the stored evidence types.
Incorrect Answers
Model choice should match structured output requirements rather than general language fluency alone. Model modalities and preprocessing must match the stored evidence types.
Chunking should preserve semantically necessary context instead of merely maximizing ingestion speed. Model modalities and preprocessing must match the stored evidence types.
Synchronous external calls inside write transactions can couple availability to a remote inference service. Model modalities and preprocessing must match the stored evidence types.
Question 204
An inventory update and fulfillment job update the same tables in opposite orders and periodic SQL deadlock victims trigger retries. Which test result would demonstrate the fix worked?
Explanation
Correct Answer
Repeated parallel tests no longer reproduce the deadlock and functional outcomes remain correct.
Incorrect Answers
The validated remedy shortens lock waits and separate deadlock monitoring remains stable. Repeated parallel tests no longer reproduce the deadlock and functional outcomes remain correct.
The new configuration reduces read blocking while reports avoid uncommitted data and temp/version costs are monitored. Repeated parallel tests no longer reproduce the deadlock and functional outcomes remain correct.
The chosen change lowers total reads without an unacceptable index maintenance penalty. Repeated parallel tests no longer reproduce the deadlock and functional outcomes remain correct.
Question 205
A billing service reports connection timeouts and the team labels every wait a deadlock without inspecting lock ownership. What technical finding would substantiate this root cause?
Explanation
Correct Answer
A blocked request waits behind a long-running transaction but no deadlock victim is recorded. Blocking is waiting for locks; a deadlock involves a cycle that SQL Server must resolve.
Incorrect Answers
The plan includes hundreds of thousands of lookups for columns absent from the nonclustered index. Blocking is waiting for locks; a deadlock involves a cycle that SQL Server must resolve. A blocked request waits behind a long-running transaction but no deadlock victim is recorded.
The captured plan was compiled with a tiny tenant ID and scans badly when reused for a huge tenant. Blocking is waiting for locks; a deadlock involves a cycle that SQL Server must resolve. A blocked request waits behind a long-running transaction but no deadlock victim is recorded.
Query Store shows the same query hash shifted to a more expensive plan after deployment. Blocking is waiting for locks; a deadlock involves a cycle that SQL Server must resolve. A blocked request waits behind a long-running transaction but no deadlock victim is recorded.
Question 206
A team switches between hosted models in a coding assistant and assumes all model endpoints have identical data handling terms. What would be a valid post-database-change verification?
Explanation
Correct Answer
Approved sessions use an allowed endpoint and unapproved configurations are blocked or corrected.
Incorrect Answers
New suggestions follow the documented conventions and still pass independent pull-request review. Approved sessions use an allowed endpoint and unapproved configurations are blocked or corrected.
A selected index improves target queries without breaching transactional and storage budgets. Approved sessions use an allowed endpoint and unapproved configurations are blocked or corrected.
Eligible users can perform an approved task and excluded users cannot access unsupported capabilities. Approved sessions use an allowed endpoint and unapproved configurations are blocked or corrected.
Question 207
A procedure filters customers by region, usually returns ten rows but occasionally one region contains half the table; a cached plan performs poorly for the outlier. Which outcome shows the requirement is met without introducing a new risk?
Explanation
Correct Answer
Representative small and large parameter tests both meet latency with acceptable compilation cost.
Incorrect Answers
A cross-tenant negative test returns no forbidden records while authorized calls still work. Representative small and large parameter tests both meet latency with acceptable compilation cost.
A controlled update follows the intended mapping without corrupting the related tables. Representative small and large parameter tests both meet latency with acceptable compilation cost.
The service can query approved totals and receives an authorization error for direct sensitive table reads. Representative small and large parameter tests both meet latency with acceptable compilation cost.
Question 208
A procedure posts an invoice and then calls a failing validation routine; the customer balance remains changed because the error is swallowed. What is the MOST appropriate initial remediation?
Explanation
Correct Answer
Implement TRY CATCH with transaction state handling, rollback, and error propagation appropriate to the caller. T-SQL error handling must preserve transaction atomicity rather than merely log exceptions.
Incorrect Answers
Extract the required JSON property and convert it to an appropriate numeric type with invalid-data handling. T-SQL error handling must preserve transaction atomicity rather than merely log exceptions. Implement TRY CATCH with transaction state handling, rollback, and error propagation appropriate to the caller.
Evaluate JSON_CONTAINS or another supported JSON-aware predicate for the target engine and index design. T-SQL error handling must preserve transaction atomicity rather than merely log exceptions. Implement TRY CATCH with transaction state handling, rollback, and error propagation appropriate to the caller.
Use OPENJSON with an explicit schema to expand validated line items into relational rows. T-SQL error handling must preserve transaction atomicity rather than merely log exceptions. Implement TRY CATCH with transaction state handling, rollback, and error propagation appropriate to the caller.
Question 209
A SQL service calls an external model and assumes its output is always a plain text answer, but the service contract requires structured fields and safe errors. How should the SQL engineering team confirm the issue is truly resolved?
Explanation
Correct Answer
Valid responses populate the correct schema while failures are logged and retried safely.
Incorrect Answers
The workflow meets token and latency budgets while citing the correct current clauses. Valid responses populate the correct schema while failures are logged and retried safely.
The model summarizes the comment without executing any unauthorized database change. Valid responses populate the correct schema while failures are logged and retried safely.
The request sends only approved fields and the model receives sufficient task context. Valid responses populate the correct schema while failures are logged and retried safely.
Question 210
Readers wait behind frequent updates in an order system and staff propose dirty reads as a quick fix for business dashboards. Which outcome shows the requirement is met without introducing a new risk?
Explanation
Correct Answer
The new configuration reduces read blocking while reports avoid uncommitted data and temp/version costs are monitored. The operating rule is: Versioned reads can improve concurrency without exposing dirty values when correctly configured.
Incorrect Answers
The analysis identifies specific statements and a tested change reduces peak resource use. The operating rule is: Versioned reads can improve concurrency without exposing dirty values when correctly configured. The new configuration reduces read blocking while reports avoid uncommitted data and temp/version costs are monitored.
The validated remedy shortens lock waits and separate deadlock monitoring remains stable. The operating rule is: Versioned reads can improve concurrency without exposing dirty values when correctly configured. The new configuration reduces read blocking while reports avoid uncommitted data and temp/version costs are monitored.
Under concurrent replay at most one valid reservation commits and a clear retry result reaches the other caller. The operating rule is: Versioned reads can improve concurrency without exposing dirty values when correctly configured. The new configuration reduces read blocking while reports avoid uncommitted data and temp/version costs are monitored.
Developing AI-Enabled Database Solutions (DP-800) Practice Tests
- Microsoft DP-800 Practice Test 19 - 10 Questions (Last Reviewed 29-Oct-2026)
- Microsoft DP-800 Practice Test 20 - 10 Questions (Last Reviewed 28-Oct-2026)
- Microsoft DP-800 Practice Test 21 - 10 Questions (Last Reviewed 27-Oct-2026)
- Microsoft DP-800 Practice Test 22 - 10 Questions (Last Reviewed 26-Oct-2026)
- Microsoft DP-800 Practice Test 23 - 10 Questions (Last Reviewed 25-Oct-2026)