Develop AI Solutions by Using Azure Cosmos DB for NoSQL AI-200 Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which SQL function in Azure Cosmos DB for NoSQL is used to perform a vector similarity search?

   • AVectorDistance
   • BST_DISTANCE
   • CCONTAINS
   • DARRAY_CONTAINS
   Reveal Solution Discussion Select an answer first
2. 22expert · hard

   A legal document management system stores documents in Azure Cosmos DB for NoSQL. Each document has a `content` field and an `embedding` field that is a vector of 1536 dimensions. The team wants to perform vector similarity searches to find related documents. They are concerned about the RU consumption of vector searches. Which approach should they take to optimize RU consumption for vector searches?

   • AUse a larger vector dimension to improve accuracy.
   • BDisable the vector index and use a full scan.
   • CUse a smaller vector dimension by applying dimensionality reduction to the embeddings.
   • DUse a different similarity metric, such as Euclidean distance.
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A logistics company stores shipment documents in Azure Cosmos DB for NoSQL. Each document has a `shipmentId`, `status`, `region`, and `lastUpdated` field. The team frequently runs a query to find all shipments in a specific region with a status of 'Delivered', ordered by `lastUpdated` descending. The query currently performs a cross-partition scan and consumes high RUs. The collection is partitioned by `/shipmentId`. What should the team do to reduce RU consumption for this query?

   • AAdd a single-field index on status and keep the existing query.
   • BChange the partition key to /region and keep the existing composite index.
   • CDisable indexing on the collection to reduce write RU and rely on the query engine's scan.
   • DAdd a composite index on (region, status, lastUpdated DESC) and change the query to filter by region and status.
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   Which clause is typically used with VectorDistance to return the most similar items first?

   • AGROUP BY
   • BDISTINCT
   • CORDER BY
   • DJOIN
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   An e-commerce platform uses Azure Cosmos DB for NoSQL to store product inventory. When a product's price changes, the system needs to update a search index in real time. The team decides to use the change feed processor to detect price changes. Which component must be implemented to ensure the change feed processor can resume processing after a failure without missing changes?

   • AA trigger on the container to invoke the processor.
   • BA lease container to store the processing state.
   • CA dead-letter queue to store failed changes.
   • DA second container to store the processed changes.
   Reveal Solution Discussion Select an answer first