Add vectors-DashVector(DashVector)-阿里云帮助中心

Prerequisites

Before you begin, make sure that you have:

An activated DashVector service. See Activate DashVector
A collection. See Create a collection

Quick start (Python)

import dashvector
import os

client = dashvector.Client(
    api_key=os.environ.get('DASHVECTOR_API_KEY'),
    endpoint=os.environ.get('DASHVECTOR_ENDPOINT')
)

collection = client.get(name='example_collection')

# Add a single vector
collection.insert(
    (
        'vec-001',                        # Primary key ID
        [0.1, 0.2, 0.3, 0.4]             # Vector data (must match collection dimensions)
    )
)

# Add a vector with metadata
collection.insert(
    (
        'vec-002',
        [0.5, 0.6, 0.7, 0.8],
        {'price': 100, 'type': 'dress'}   # Optional metadata (fields)
    )
)

For the full Python SDK reference, see Insert documents (Python). For the Java SDK reference, see Insert documents (Java).

Add a vector in the console

Log on to the DashVector console.
In the left-side navigation pane, click Clusters. Find the target collection and click Details in the Operation column.
In the left-side bar, click Vector Add.
Configure the following parameters and click Confirm.

Parameters

Parameter	API parameter	Required	Description
Vector	`vector`	Yes	The vector data. The number of dimensions and data type must match the collection. Example: `[1.00,2.00,3.00,4.00]`
Primary Key ID	`id`	Yes	A unique identifier for the vector. Supports letters, digits, and the following special characters: `_ - ! @ # $ % + = .` Maximum length: 64 characters.
Partition	`partition`	Yes	The target partition. Default value: `default`. To use a non-default partition, create one first. See Create a partition.
Attribute	`fields`	No	Metadata in JSON format. Example: `{"price":100,"type":"dress"}`

Add vectors with SDKs

DashVector provides Python and Java SDKs for adding vectors programmatically.

Python

import dashvector
import os

client = dashvector.Client(
    api_key=os.environ.get('DASHVECTOR_API_KEY'),
    endpoint=os.environ.get('DASHVECTOR_ENDPOINT')
)

collection = client.get(name='example_collection')

# Add a single vector
ret = collection.insert(
    (
        'id-001',
        [0.1, 0.2, 0.3, 0.4]
    )
)
print(ret)
# Output example:
# {"request_id": "...", "code": 0, "message": "success"}

# Add a vector with metadata and a target partition
ret = collection.insert(
    (
        'id-002',
        [0.5, 0.6, 0.7, 0.8],
        {'price': 100, 'type': 'dress'}
    ),
    partition='my_partition'
)

# Batch add vectors
ret = collection.insert(
    [
        ('id-003', [0.1, 0.2, 0.3, 0.4], {'category': 'shoes'}),
        ('id-004', [0.9, 0.8, 0.7, 0.6], {'category': 'bags'}),
        ('id-005', [0.3, 0.4, 0.5, 0.6], {'category': 'hats'}),
    ]
)

For the full Python SDK reference, see Insert documents.

Java

import com.aliyun.dashvector.DashVectorClient;
import com.aliyun.dashvector.DashVectorCollection;
import com.aliyun.dashvector.models.Doc;
import com.aliyun.dashvector.models.DashVectorResult;

import java.util.*;

public class InsertExample {
    public static void main(String[] args) {
        DashVectorClient client = new DashVectorClient(
            System.getenv("DASHVECTOR_ENDPOINT"),
            System.getenv("DASHVECTOR_API_KEY")
        );

        DashVectorCollection collection = client.get("example_collection");

        // Add a single vector
        Doc doc = Doc.builder()
            .id("id-001")
            .vector(Arrays.asList(0.1f, 0.2f, 0.3f, 0.4f))
            .fields(Map.of("price", 100, "type", "dress"))
            .build();

        DashVectorResult result = collection.insert(doc);
    }
}

For the full Java SDK reference, see Insert documents.

Add a vector with the HTTP API

For the full HTTP API reference, see Insert documents.

Limits

Item	Limit
Primary key ID length	64 characters maximum
Primary key ID characters	Letters, digits, and `_ - ! @ # $ % + = .`
Vector dimensions	Must match the collection configuration
Vector data type	Must match the collection configuration
Metadata format	JSON