From the course: LLM Foundations: Vector Databases for Caching and Retrieval Augmented Generation (RAG)

Unlock the full course today

Join today to access over 23,200 courses taught by industry experts.

Indexes in Milvus

Indexes in Milvus

From the course: LLM Foundations: Vector Databases for Caching and Retrieval Augmented Generation (RAG)

Start my 1-month free trial Buy for my team

Indexes in Milvus

“

In this video, we will review indexing capabilities in Milvus. What are Milvus indexes? Indexes play a key role in any database product to help speed up data retrieval. Milvus also has indexes to help with the same. In Milvus, we can create indexes on either scalar or vector fields. Scalar indexes work similar to RDBMS indexes. There can only be one index per field. Also, there is no support for composite indexes in Milvus. Indexes help organize vectors using an ANN metric type like L2 or IP. The index is set up in such a way that will help search using these metrics efficiently. Vector indexes are a prerequisite to perform ANN searches on vector fields. These indexes must be created before any such search. Milvus supports several index types which can be chosen based on the use case. Here is a list of popular index types in Milvus. This list is not exhaustive. Each index type has a specific internal organization. FLAT indexes are suitable for small dataset and have a 100% recall…

Contents