Update payload metadata by re-upserting points with new data. This replaces the entire payload — partial field updates are not supported.
Before you begin, make sure the target points already exist in your collection. If a point ID does not exist, then the upsert operation creates a new point instead of updating.
Update a single payload
The upsert method updates existing points or creates new ones if they do not exist. When updating metadata, you must provide the vector, but payload is optional. It is common to retrieve and preserve the existing vector. The following example updates one point’s payload.
import random
from actian_vectorai import VectorAIClient, PointStruct
DIMENSION = 128
COLLECTION = "products"
# Connect to VectorAI DB server
with VectorAIClient("localhost:50051") as client:
# Get existing point
existing = client.points.get(COLLECTION, ids=[1])
# Preserve existing vector or generate new one
vector = existing[0].vectors if existing[0].vectors else [random.gauss(0, 1) for _ in range(DIMENSION)]
# Update with new payload
point = PointStruct(
id=1, # Point ID to update
vector=vector, # Keep existing vector
payload={ # New metadata (optional)
"name": "Gaming Laptop",
"category": "electronics",
"price": 899.99,
"in_stock": True,
"updated": True
}
)
# Upsert point with updated payload
client.points.upsert(COLLECTION, [point])
print("Payload updated successfully")
Update multiple payloads
Upsert operations perform full replacement of payload data when provided. You must provide the complete new payload — partial updates are not supported. If you need to preserve some fields, then retrieve the existing payload first.
The following example updates multiple payloads at once.
import random
from actian_vectorai import VectorAIClient, PointStruct
DIMENSION = 128
COLLECTION = "products"
# Connect to VectorAI DB server
with VectorAIClient("localhost:50051") as client:
# Retrieve existing points
results = client.points.get(COLLECTION, ids=[1, 2])
# Update multiple points with new payloads
points = [
PointStruct(
id=1, # Point ID
vector=results[0].vectors if results[0].vectors else [random.gauss(0, 1) for _ in range(DIMENSION)], # Preserve vector
payload={"name": "Gaming Laptop", "category": "electronics", "price": 1299.99} # New metadata (optional)
),
PointStruct(
id=2, # Point ID
vector=results[1].vectors if results[1].vectors else [random.gauss(0, 1) for _ in range(DIMENSION)], # Preserve vector
payload={"name": "Smartphone", "category": "electronics", "price": 799.99} # New metadata (optional)
)
]
# Batch upsert updated points
client.points.upsert(COLLECTION, points)
print("Batch updated 2 payloads")
