100% Latest & Updated Microsoft DP-800 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 960
An organization wants RAG that respects row-level security. A test instead finds: User sees facts from another tenant. Which underlying issue best explains this finding?
- Context was flooded with low-relevance passages
- No evaluated ground-truth and citation evidence exists
- Logging pipeline persists unredacted secrets
- RAG pipeline bypasses SQL authorization
Explanation
Correct Answer
D:RAG pipeline bypasses SQL authorization. It accounts for the specific observed failure: user sees facts from another tenant. The intended solution is run retrieval under requesting user's authorized context with the required setup.
Incorrect Answers
A:Context was flooded with low-relevance passages. This concerns token cost spikes without improved grounding. The relevant SQL or system operation is RAG pipeline bypasses SQL authorization.
B:No evaluated ground-truth and citation evidence exists. This concerns team declares success from fluent demos only. The relevant SQL or system operation is RAG pipeline bypasses SQL authorization.
C:Logging pipeline persists unredacted secrets. This concerns prompt history stores connection strings verbatim. The relevant SQL or system operation is RAG pipeline bypasses SQL authorization.
Question 961
A policy bot should cite exact source version and timestamp. Which TWO listed actions form the most appropriate database design and its concrete implementation? (Choose TWO.)
- Use groundedness, relevance and citation-based evaluation
- Include source identifiers, version and timestamps in prompt context
- Define prompt template with instructions and bounded evidence section
- Prioritize authoritative current source and reconcile conflicts
- Retrieve chunk metadata alongside text and expose citations
Explanation
Correct Answers
B:Include source identifiers, version and timestamps in prompt context. The implementation is retrieve chunk metadata alongside text and expose citations; success can be checked by open cited source and confirm effective date. The method addresses the stated data and workload constraint.
E:This is the concrete configuration of the specified design: retrieve chunk metadata alongside text and expose citations. The choice makes the implementation testable in the given workload.
Incorrect Answers
A:Use groundedness, relevance and citation-based evaluation. This concerns checking whether rag helped over unguided prompts. Here, include source identifiers, version and timestamps in prompt context.
C:Define prompt template with instructions and bounded evidence section. This concerns a reusable prompt with SQL database retrieval. Here, include source identifiers, version and timestamps in prompt context.
D:Prioritize authoritative current source and reconcile conflicts. This concerns a rag answer includes contradictory facts from old and new records. Here, include source identifiers, version and timestamps in prompt context.
Question 962
A policy bot should cite exact source version and timestamp. The team selects: Include source identifiers, version and timestamps in prompt context. Which implementation action puts this design into practice?
- Retrieve authorized current passages and pass them with question
- Set candidate ranking and token budget for retrieved chunks
- Create SQL projection excluding secrets before external request
- Retrieve chunk metadata alongside text and expose citations
Explanation
Correct Answer
D:Retrieve chunk metadata alongside text and expose citations. This is the concrete setup for include source identifiers, version and timestamps in prompt context, and open cited source and confirm effective date supplies a direct confirmation.
Incorrect Answers
A:Retrieve authorized current passages and pass them with question. This concerns an ai assistant answers questions from frequently changing internal procedures. Instead, retrieve chunk metadata alongside text and expose citations.
B:Set candidate ranking and token budget for retrieved chunks. This concerns a SQL-backed assistant runs retrieval but sends too many irrelevant rows. Instead, retrieve chunk metadata alongside text and expose citations.
C:Create SQL projection excluding secrets before external request. This concerns a support bot routes answers to an external model via rest. Instead, retrieve chunk metadata alongside text and expose citations.
Question 963
A policy bot should cite exact source version and timestamp. After implementing the design, which check most directly validates the expected behavior?
- Test out-of-scope question returns grounded uncertainty
- Test prompt injection phrase embedded in source document
- Open cited source and confirm effective date
- Compare retrieved source IDs for different users
Explanation
Correct Answer
C:Open cited source and confirm effective date. This verifies the requirement in the case rather than merely confirming that a job ran; an opposing result can reveal that retrieval filter ignored effective-date metadata.
Incorrect Answers
A:Test out-of-scope question returns grounded uncertainty. This concerns not invent answer when SQL retrieval returns no supporting rows. Instead, open cited source and confirm effective date.
B:Test prompt injection phrase embedded in source document. This concerns a reusable prompt with SQL database retrieval. Instead, open cited source and confirm effective date.
D:Compare retrieved source IDs for different users. This concerns rag that respects row-level security. Instead, open cited source and confirm effective date.
Question 964
A policy bot should cite exact source version and timestamp. A test instead finds: Citation points to superseded policy revision. Which underlying issue best explains this finding?
- RAG context construction over-shared database columns
- Prompt lacks no-evidence behavior or model ignored it
- Retrieval included stale records or skipped freshness filter
- Retrieval filter ignored effective-date metadata
Explanation
Correct Answer
D:Retrieval filter ignored effective-date metadata. It accounts for the specific observed failure: citation points to superseded policy revision. The intended solution is include source identifiers, version and timestamps in prompt context with the required setup.
Incorrect Answers
A:RAG context construction over-shared database columns. This concerns external model receives unrelated customer details. Here, retrieval filter ignored effective-date metadata.
B:Prompt lacks no-evidence behavior or model ignored it. This concerns assistant confidently makes up non-existent product policy. Here, retrieval filter ignored effective-date metadata.
C:Retrieval included stale records or skipped freshness filter. This concerns assistant answers with last year's retirement policy. Here, retrieval filter ignored effective-date metadata.
Question 965
A RAG answer includes contradictory facts from old and new records. Which database design or engineering approach most directly meets this requirement?
- Use RAG to ground generation in current approved records
- Use sp_invoke_external_rest_endpoint where supported and configured
- Prioritize authoritative current source and reconcile conflicts
- Include source identifiers, version and timestamps in prompt context
Explanation
Correct Answer
C:Prioritize authoritative current source and reconcile conflicts. The implementation is deduplicate by document id and select newest approved version; success can be checked by test contradictory version pair yields current rule. The method addresses the stated data and workload constraint.
Incorrect Answers
A:Use RAG to ground generation in current approved records. This concerns an ai assistant answers questions from frequently changing internal procedures. Here, prioritize authoritative current source and reconcile conflicts.
B:Use sp_invoke_external_rest_endpoint where supported and configured. This concerns a stored procedure sends an allowed request to an external text model. Here, prioritize authoritative current source and reconcile conflicts.
D:Include source identifiers, version and timestamps in prompt context. This concerns cite exact source version and timestamp. Here, prioritize authoritative current source and reconcile conflicts.
Question 966
A RAG answer includes contradictory facts from old and new records. The team selects: Prioritize authoritative current source and reconcile conflicts. Which implementation action puts this design into practice?
- Apply RLS predicate and per-document ACL before prompt construction
- Explicitly instruct refusal or escalation when context lacks answer
- Retrieve chunk metadata alongside text and expose citations
- Deduplicate by document ID and select newest approved version
Explanation
Correct Answer
D:Deduplicate by document ID and select newest approved version. This is the concrete setup for prioritize authoritative current source and reconcile conflicts, and test contradictory version pair yields current rule supplies a direct confirmation.
Incorrect Answers
A:Apply RLS predicate and per-document ACL before prompt construction. This concerns rag that respects row-level security. Here, deduplicate by document ID and select newest approved version.
B:Explicitly instruct refusal or escalation when context lacks answer. This concerns not invent answer when SQL retrieval returns no supporting rows. Here, deduplicate by document ID and select newest approved version.
C:Retrieve chunk metadata alongside text and expose citations. This concerns cite exact source version and timestamp. Here, deduplicate by document ID and select newest approved version.
Question 967
A RAG answer includes contradictory facts from old and new records. After implementing the design, which check most directly validates the expected behavior?
- Compare retrieved source IDs for different users
- Compare hallucination rate and answer citation accuracy
- Test contradictory version pair yields current rule
- Verify generated answer cites valid current passage
Explanation
Correct Answer
C:Test contradictory version pair yields current rule. This verifies the requirement in the case rather than merely confirming that a job ran; an opposing result can reveal that context included duplicate conflicting revisions.
Incorrect Answers
A:Compare retrieved source IDs for different users. This concerns rag that respects row-level security. Instead, test contradictory version pair yields current rule.
B:Compare hallucination rate and answer citation accuracy. This concerns checking whether rag helped over unguided prompts. Instead, test contradictory version pair yields current rule.
D:Verify generated answer cites valid current passage. This concerns an ai assistant answers questions from frequently changing internal procedures. Instead, test contradictory version pair yields current rule.
Question 968
A RAG answer includes contradictory facts from old and new records. A test instead finds: Model quotes obsolete value despite new approved record. Which underlying issue best explains this finding?
- Retrieval included stale records or skipped freshness filter
- No evaluated ground-truth and citation evidence exists
- Context included duplicate conflicting revisions
- SQL rows were flattened without structured context
Explanation
Correct Answer
C:Context included duplicate conflicting revisions. It accounts for the specific observed failure: model quotes obsolete value despite new approved record. The intended solution is prioritize authoritative current source and reconcile conflicts with the required setup.
Incorrect Answers
A:Retrieval included stale records or skipped freshness filter. This concerns assistant answers with last year's retirement policy. Here, context included duplicate conflicting revisions.
B:No evaluated ground-truth and citation evidence exists. This concerns team declares success from fluent demos only. Here, context included duplicate conflicting revisions.
D:SQL rows were flattened without structured context. This concerns llm receives ambiguous comma-separated values. Here, context included duplicate conflicting revisions.
Question 969
A SQL-backed assistant runs retrieval but sends too many irrelevant rows. Which database design or engineering approach most directly meets this requirement?
- Use evidence-aware abstention in RAG prompting
- Limit top-K and trim passages by relevance
- Use approved Foundry endpoint and managed authentication
- Serialize controlled result fields to JSON
Explanation
Correct Answer
B:Limit top-K and trim passages by relevance. The implementation is set candidate ranking and token budget for retrieved chunks; success can be checked by measure answer quality and input token count. The method addresses the stated data and workload constraint.
Incorrect Answers
A:Use evidence-aware abstention in RAG prompting. This concerns not invent answer when SQL retrieval returns no supporting rows. Here, limit top-K and trim passages by relevance.
C:Use approved Foundry endpoint and managed authentication. This concerns connecting SQL retrieval to microsoft foundry model. Here, limit top-K and trim passages by relevance.
D:Serialize controlled result fields to JSON. This is relevant when a chatbot uses SQL result rows as context for a model. Here, limit top-K and trim passages by relevance.
Question 970
A SQL-backed assistant runs retrieval but sends too many irrelevant rows. The team selects: Limit top-K and trim passages by relevance. Which implementation action puts this design into practice?
- Set candidate ranking and token budget for retrieved chunks
- Explicitly instruct refusal or escalation when context lacks answer
- Generate hybrid top-K with RRF and bounded source context
- Use FOR JSON PATH or JSON_OBJECT projection of allowed fields
Explanation
Correct Answer
A:Set candidate ranking and token budget for retrieved chunks. This is the concrete setup for limit top-k and trim passages by relevance, and measure answer quality and input token count supplies a direct confirmation.
Incorrect Answers
B:Explicitly instruct refusal or escalation when context lacks answer. This concerns not invent answer when SQL retrieval returns no supporting rows. Instead, set candidate ranking and token budget for retrieved chunks.
C:Generate hybrid top-K with RRF and bounded source context. This concerns hybrid SQL/vector candidates before generating answer. Instead, set candidate ranking and token budget for retrieved chunks.
D:Use FOR JSON PATH or JSON_OBJECT projection of allowed fields. This is relevant when a chatbot uses SQL result rows as context for a model. Instead, set candidate ranking and token budget for retrieved chunks.
Developing AI-Enabled Database Solutions (DP-800) Practice Tests