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 76
Question 750
A device catalog adds hundreds of mostly empty optional fields as standard columns and read performance is deteriorating. Which TWO options together specify a corrective step and a meaningful test?
Explanation
Correct Answers
The revised schema supports required searches while reducing unnecessary storage and migration burden. Evaluate normalized, JSON, or sparse-column designs against actual query needs and engine support.
Evaluate normalized, JSON, or sparse-column designs against actual query needs and engine support.
Incorrect Answers
Checkout transaction latency remains stable while reporting queries use purpose-built structures. Evaluate normalized, JSON, or sparse-column designs against actual query needs and engine support. The revised schema supports required searches while reducing unnecessary storage and migration burden.
Design JSON access and indexing according to the engine's supported JSON features and test the predicate. Evaluate normalized, JSON, or sparse-column designs against actual query needs and engine support.
Add supported CHECK and DEFAULT constraints consistent with the application contract. Evaluate normalized, JSON, or sparse-column designs against actual query needs and engine support. The revised schema supports required searches while reducing unnecessary storage and migration burden.
Question 751
An AI-enabled SQL workflow calls a model service with an API key stored in a database object accessible to developers. Choose TWO statements: a suitable intervention and evidence of its success.
Explanation
Correct Answers
Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers.
Authorized model calls succeed with least privilege and no plaintext key is exposed to unrelated readers. Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers.
Incorrect Answers
Add appropriate query/activity monitoring and tested audit correlation without logging unnecessary sensitive payloads. Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers. Authorized model calls succeed with least privilege and no plaintext key is exposed to unrelated readers.
Compromised low-privilege identities cannot read unrelated protected data in a controlled test. Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers. Authorized model calls succeed with least privilege and no plaintext key is exposed to unrelated readers.
Enforce authorized tenant scoping through backend policies and service identity rather than trust the caller's filter. Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers.
Question 752
A SQL MCP tool returns a table comment that instructs an AI coding assistant to read a secret table and send data to another endpoint. How should the SQL engineering team confirm the issue is truly resolved?
Explanation
Correct Answer
The assistant ignores the embedded command and logs no attempt to access the secret table. The operating rule is: Untrusted tool output must not acquire higher instruction priority than authorized developer intent.
Incorrect Answers
Approved sessions use an allowed endpoint and unapproved configurations are blocked or corrected. The operating rule is: Untrusted tool output must not acquire higher instruction priority than authorized developer intent. The assistant ignores the embedded command and logs no attempt to access the secret table.
Tests confirm exact expected totals for refunds, null values, and normal billing records. The operating rule is: Untrusted tool output must not acquire higher instruction priority than authorized developer intent. The assistant ignores the embedded command and logs no attempt to access the secret table.
New suggestions follow the documented conventions and still pass independent pull-request review. The operating rule is: Untrusted tool output must not acquire higher instruction priority than authorized developer intent. The assistant ignores the embedded command and logs no attempt to access the secret table.
Question 753
A security team wants to confirm that cross-tenant access attempts are detected, not only blocked by its new row-level security predicate. What should the database developer do FIRST within approved procedures?
Explanation
Correct Answer
Add appropriate query/activity monitoring and tested audit correlation without logging unnecessary sensitive payloads. Access prevention and accountable security monitoring are complementary controls.
Incorrect Answers
Configure appropriate SQL auditing to a protected destination with tested retention and access reviews. Access prevention and accountable security monitoring are complementary controls. Add appropriate query/activity monitoring and tested audit correlation without logging unnecessary sensitive payloads.
Plan coordinated key rotation and recovery testing before replacing or retiring key material. Access prevention and accountable security monitoring are complementary controls. Add appropriate query/activity monitoring and tested audit correlation without logging unnecessary sensitive payloads.
Choose an approved managed identity access route for the model endpoint where supported and restrict configuration readers. Access prevention and accountable security monitoring are complementary controls. Add appropriate query/activity monitoring and tested audit correlation without logging unnecessary sensitive payloads.
Question 754
A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path. Select TWO: the best action and the outcome that confirms it worked.
Explanation
Correct Answers
Capture API query shapes and SQL dependency telemetry before scaling hardware. A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path.
A bounded query or optimized mapping meets latency targets under representative load. A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path. Capture API query shapes and SQL dependency telemetry before scaling hardware.
Incorrect Answers
Use a suitable Logic Apps or event-triggered integration with narrow filters and approved recipients. A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path. Capture API query shapes and SQL dependency telemetry before scaling hardware.
Configure the supported DAB GraphQL relationship mapping and test nested authorization across entities. A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path. Capture API query shapes and SQL dependency telemetry before scaling hardware.
An API test enforces page size and returns deterministic next pages without missing or duplicating items. A GraphQL response is slow for deeply nested selections and the team considers adding larger SQL capacity without identifying the expensive path. Capture API query shapes and SQL dependency telemetry before scaling hardware.
Question 755
A pricing service caches DAB responses but customers continue receiving superseded prices after a time-sensitive promotion update. What technical finding would substantiate this root cause?
Explanation
Correct Answer
SQL shows the approved new price while the API still returns the old cached payload. Caching must respect correctness and freshness requirements for mutable business records.
Incorrect Answers
A test can call the DAB backend URL directly without the gateway's identity checks. Caching must respect correctness and freshness requirements for mutable business records. SQL shows the approved new price while the API still returns the old cached payload.
The existing solution grants multiple consumers direct read access to core transactional tables. Caching must respect correctness and freshness requirements for mutable business records. SQL shows the approved new price while the API still returns the old cached payload.
Users report timeouts while infrastructure charts show normal CPU and no request trace. Caching must respect correctness and freshness requirements for mutable business records. SQL shows the approved new price while the API still returns the old cached payload.
Question 756
Readers wait behind frequent updates in an order system and staff propose dirty reads as a quick fix for business dashboards. Which concept should guide the database developer in this situation?
Explanation
Correct Answer
Versioned reads can improve concurrency without exposing dirty values when correctly configured.
Incorrect Answers
Business data consistency requires protecting read-modify-write sequences. Versioned reads can improve concurrency without exposing dirty values when correctly configured.
Query Store provides historical query-plan evidence and supports evidence-based regression diagnosis. Versioned reads can improve concurrency without exposing dirty values when correctly configured.
Index maintenance should be justified by observed behavior and storage characteristics. Versioned reads can improve concurrency without exposing dirty values when correctly configured.
Question 757
An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction. Choose TWO statements: a suitable intervention and evidence of its success.
Explanation
Correct Answers
The FAQ write commits promptly and a background worker creates a matching authorized vector. An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction.
Use an event-driven Azure Function SQL trigger pattern with durable retries and source version checks. An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction.
Incorrect Answers
A changed row's embedding is refreshed within the accepted time window and old results no longer dominate. An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction. The FAQ write commits promptly and a background worker creates a matching authorized vector.
Evaluate a CDC-based incremental vector update process with checkpoints and delete propagation. An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction. Use an event-driven Azure Function SQL trigger pattern with durable retries and source version checks.
Move inference to an appropriate asynchronous change processing flow or tightly bound the operation after evaluating failure risk. An application wants new FAQ entries to receive embeddings soon after insertion without embedding generation blocking the user-facing SQL transaction. The FAQ write commits promptly and a background worker creates a matching authorized vector.
Question 758
An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits. Which TWO options together specify a corrective step and a meaningful test?
Explanation
Correct Answers
The workflow meets token and latency budgets while citing the correct current clauses. An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits.
Bound and rerank retrieved context, preserve meaningful chunk metadata, and benchmark answer coverage. An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits.
Incorrect Answers
Every material policy claim cites a current accessible source or is marked unsupported. An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits. The workflow meets token and latency budgets while citing the correct current clauses.
Record authorized prompt/source version metadata, model identity, and approval outcome with appropriate privacy retention. An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits. The workflow meets token and latency budgets while citing the correct current clauses.
Include controlled document identifiers, version metadata, and explicit citation requirements in the RAG prompt. An SQL-based RAG procedure sends hundreds of long retrieved chunks to a model and becomes expensive while often omitting the most relevant clauses due to context limits. The workflow meets token and latency budgets while citing the correct current clauses.
Question 759
An engineering help-desk needs exact error-code matching plus conceptually relevant troubleshooting articles and neither keyword nor vectors alone performs well. How should the SQL engineering team confirm the issue is truly resolved?
Explanation
Correct Answer
A combined benchmark retrieves both precise code matches and relevant related explanations.
Incorrect Answers
The combined ranking balances relevant candidates from both lists according to measured relevance. A combined benchmark retrieves both precise code matches and relevant related explanations.
A pre-insert check identifies incompatible vectors and routes them to repair before indexing. A combined benchmark retrieves both precise code matches and relevant related explanations.
A tested query returns the expected nearest vector rows with reproducible distance values. A combined benchmark retrieves both precise code matches and relevant related explanations.
Question 760
A customer analytics query calls an expensive scalar UDF for every row and slows dramatically during monthly aggregation. Select TWO: the best action and the outcome that confirms it worked.
Explanation
Correct Answers
Assess supported scalar UDF inlining or rewrite the computation into a set-based expression. A revised query returns identical outputs with materially less CPU under production-like data.
A revised query returns identical outputs with materially less CPU under production-like data. The corresponding check is whether assess supported scalar UDF inlining or rewrite the computation into a set-based expression.
Incorrect Answers
Either the complete authorized order is committed or the entire unit of work is rolled back. A revised query returns identical outputs with materially less CPU under production-like data. The corresponding check is whether assess supported scalar UDF inlining or rewrite the computation into a set-based expression.
Make reference-data seeding idempotent using stable keys and transactional checks. A revised query returns identical outputs with materially less CPU under production-like data.
Create an appropriately permissioned view that projects only authorized aggregated fields. A revised query returns identical outputs with materially less CPU under production-like data. Assess supported scalar UDF inlining or rewrite the computation into a set-based expression.
Developing AI-Enabled Database Solutions (DP-800) Practice Tests
- Microsoft DP-800 Practice Test 74 - 10 Questions (Last Reviewed 31-Oct-2026)
- Microsoft DP-800 Practice Test 75 - 10 Questions (Last Reviewed 30-Oct-2026)
- Microsoft DP-800 Practice Test 76 - 10 Questions (Last Reviewed 29-Oct-2026)
- Microsoft DP-800 Practice Test 77 - 10 Questions (Last Reviewed 28-Oct-2026)
- Microsoft DP-800 Practice Test 78 - 10 Questions (Last Reviewed 27-Oct-2026)