Vector Index
← Back to DP-420 Terminology Reference
Category: Optimization
Summary: Enables approximate nearest-neighbor similarity search over embeddings for AI/RAG scenarios.
Overview
A vector index enables efficient approximate nearest-neighbor (ANN) similarity search over high-dimensional vector embeddings stored directly in Azure Cosmos DB items — powering AI scenarios such as retrieval-augmented generation (RAG) without needing a separate vector database. Azure Cosmos DB for NoSQL supports vector index types such as flat, quantizedFlat, and DiskANN, configured via a vector embedding policy alongside the container's indexing policy.
More Info & Links
DP-420 Style Practice Questions
Q1. What is the primary purpose of a vector index in Azure Cosmos DB?
- A. Efficient approximate nearest-neighbor similarity search over vector embeddings for AI/RAG scenarios
- B. Enforcing uniqueness of a property across a logical partition
- C. Sorting items by their partition key
- D. Accelerating geospatial distance calculations
Q2. Which of the following is a supported vector index type in Azure Cosmos DB for NoSQL?
- A. DiskANN
- B. Composite
- C. Wildcard
- D. Hash
Answer Key
Answers are listed at the very end so you can attempt the questions first.
- Q1 (Vector Index): A. Efficient approximate nearest-neighbor similarity search over vector embeddings for AI/RAG scenarios — Vector indexes enable fast similarity search over embeddings, supporting AI workloads like retrieval-augmented generation.
- Q2 (Vector Index): A. DiskANN — Azure Cosmos DB for NoSQL supports vector index types including flat, quantizedFlat, and DiskANN.
DP-420: Designing and Implementing Cloud-Native Applications Using Microsoft Azure Cosmos DB — term detail page
