Vector indexes

Vector indexes are a specialized type of index designed to efficiently query and manage vector data stored within a collection of documents. Amazon DocumentDB supports Hierarchical Navigable Small World (HNSW) and Inverted File with Flat Compression (IVFFlat) indexes.

For more information, see Vector search for Amazon DocumentDB.

Vector indexes are beneficial for machine learning and generative AI use cases, such as:

• semantic search

• product recommendation

• personalization

• chatbots

• fraud detection

• anomaly detection

Supported index properties

Option	3.6	4.0	5.0	8.0	Elastic Cluster
name	No	No	Yes	Yes	No

Creating a vector index

Use the createIndex command with the runCommand() method to create a vector index. The syntax is:

db.runCommand({
  "createIndexes": "<collection>", 
  "indexes": [{
    "key": {
      "<field>": "vector"
    },
    "name": "<name>",
    "vectorOptions": {
      "type": "<hnsw> | <ivfflat>",
      "dimensions": <number of dimensions>,
      "similarity": "<euclidean>|<cosine>|<dotProduct>",
      "lists": <number_of_lists> [applicable for IVFFlat],
      "m": <max number of connections> [applicable for HNSW],
      "efConstruction": <size of the dynamic list for index build> [applicable for HNSW]
    }
  }] 
})

The key parameter is a JSON document that specifies the field and vector index type:

{
  "<field>": "vector"
}

See Index Properties for examples of creating vector indexes.