Update points

Modify existing points with new vector and payload data.

Update existing points by providing a new vector and optionally new payload data. The upsert operation replaces the point components you provide.

Update a single point

The update operation replaces the vector and optionally the payload. When updating a point, you must provide the vector, but payload is optional.

To update only specific payload fields, first retrieve the existing point, modify the payload, then upsert it back.

Python

import random
from actian_vectorai import VectorAIClient, PointStruct

DIMENSION = 128
COLLECTION = "products"

# Connect to VectorAI DB server
with VectorAIClient("localhost:6574") as client:
    # Generate new vector
    updated_vector = [random.gauss(0, 1) for _ in range(DIMENSION)]
    
    # Update point with new data
    point = PointStruct(
        id=1,  # Point ID to update
        vector=updated_vector,  # New vector embedding
        payload={  # New metadata (optional)
            "name": "Gaming Laptop",
            "category": "electronics",
            "price": 1499.99,
            "updated": True
        }
    )
    
    # Upsert updated point
    client.points.upsert(COLLECTION, [point])
    
    print("Point updated successfully")

Update multiple points

The same replacement rules from single-point updates apply here. You must provide the vector, and payload is optional. Partial payload updates are not supported.

Python

import random
from actian_vectorai import VectorAIClient, PointStruct

DIMENSION = 128
COLLECTION = "products"

# Connect to VectorAI DB server
with VectorAIClient("localhost:6574") as client:
    # Prepare update data
    updates = [
        {"id": 2, "name": "5G Smartphone", "category": "electronics", "price": 899.99, "updated": True},
        {"id": 3, "name": "Pro Tablet", "category": "electronics", "price": 799.99, "updated": True},
        {"id": 4, "name": "Premium T-Shirt", "category": "clothing", "price": 49.99, "updated": True}
    ]
    
    # Create points with new data
    points = [
        PointStruct(
            id=item["id"],  # Point ID to update
            vector=[random.gauss(0, 1) for _ in range(DIMENSION)],  # New vector
            payload={k: v for k, v in item.items() if k != "id"}  # Extract payload fields (optional)
        )
        for item in updates
    ]
    
    # Batch update all points
    client.points.upsert(COLLECTION, points)
    
    print(f"Updated {len(points)} points")