Design and Implement Intelligent Search DP-800 Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A team is evaluating vector index types for a 50-million-row dataset. They need to balance query latency, memory usage, and recall. The ANN index they tested has high recall but uses too much memory. They want to reduce memory consumption while keeping recall acceptable. What should they do?

   • AIncrease the number of probes in the ANN index to improve recall.
   • BReduce the number of dimensions in the stored vectors by using a dimensionality reduction technique.
   • CSwitch to an exact nearest neighbor (ENN) index because it uses less memory.
   • DUse a full-text index instead of a vector index to reduce memory usage.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   What is a prerequisite for using VECTOR_SEARCH on a column?

   • AThe column must have a vector index
   • BThe column must be of type NVARCHAR
   • CThe column must be part of a primary key
   • DThe table must have a full-text catalog
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which function in SQL Server 2025 is used to calculate the distance between two vectors?

   • AVECTORPROPERTY
   • BVECTOR_NORMALIZE
   • CVECTOR_DISTANCE
   • DVECTOR_SEARCH
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A library catalog system stores book titles and descriptions. Users often search for exact phrases like 'data science' and also for topics like 'machine learning basics'. The system currently uses full-text search only. Users complain that conceptual searches return poor results. What should the team do?

   • AReplace full-text search with vector search only.
   • BAdd semantic vector search and combine results with full-text using RRF.
   • CUse a thesaurus in full-text search to expand synonyms.
   • DAdd more stop words to the full-text index.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A customer support portal stores ticket descriptions in Azure SQL Database. The team wants users to search for tickets by describing their issue in natural language, even when the exact words do not appear in the ticket. Which approach should they use?

   • AUse the SOUNDEX function to match phonetically similar words in ticket descriptions.
   • BCreate a full-text index and use the CONTAINS predicate with multiple OR conditions for likely keywords.
   • CGenerate embeddings for ticket descriptions using a text-embedding model and store them in a VECTOR column, then use VECTOR_SEARCH to find similar tickets.
   • DEnable full-text search with thesaurus support to expand synonyms for common support terms.
   Reveal Solution Discussion Select an answer first