Points overview

Points are the fundamental data units in Actian VectorAI DB. Each point contains the following three components:

A unique ID
A vector embedding
An optional metadata payload

\text{Point} = \text{ID} + \text{Vector} + \text{Payload}

Points represent any data type, including text, images, audio, video, and structured data. Unlike traditional database rows with fixed schemas, VectorAI DB points store high-dimensional embeddings with rich metadata through flexible JSON payloads.

Point structure

Each point contains three key components:

{
  "id": 1,                            // Required: Unique integer identifier
  "vector": [0.12, -0.34, 0.56, ...], // Required: Fixed-size embedding
  "payload": {                        // Optional: JSON metadata
    "name": "Laptop",
    "category": "electronics",
    "price": 999.99,
    "in_stock": true
  }
}

The following table describes each component:

Component	Required	Description
ID	Yes	Unique integer identifier within the collection.
Vector	Yes	Fixed-size embedding that matches your collection dimension configuration.
Payload	No	Nested JSON for rich filtering and metadata storage.

Point operations overview

VectorAI DB supports the following operations for managing points:

Operation	Description
Insert	Add individual points or batches of points to your collection.
Retrieve	Fetch specific points from your collection by their IDs.
Update	Modify existing points by providing new vector and payload data.
Delete	Remove individual points or batches of points from your collection.
Compact	Reclaim storage space and optimize the index after deletions. See Delete points.

Task guides

See step-by-step guides for point operations:

Next steps

Payload management — Attach and filter metadata on points.
Vector embeddings — Understand the embeddings stored in points.
Search operations — Run similarity queries across points.
Filtering — Narrow search results using point payload fields.

Collection workflow Insert points

⌘I

Collections

Points

Vectors

Payload

Search

Filtering

Semantic search

Hybrid search

Distance metrics

Indexing

Point structure

Point operations overview

Task guides

Next steps

​Point structure

​Point operations overview

​Task guides

​Next steps

Point structure

Point operations overview

Task guides

Next steps