Skip to main content


AnalyticDB for PostgreSQL is a massively parallel processing (MPP) data warehousing service that is designed to analyze large volumes of data online.

AnalyticDB for PostgreSQL is developed based on the open source Greenplum Database project and is enhanced with in-depth extensions by Alibaba Cloud. AnalyticDB for PostgreSQL is compatible with the ANSI SQL 2003 syntax and the PostgreSQL and Oracle database ecosystems. AnalyticDB for PostgreSQL also supports row store and column store. AnalyticDB for PostgreSQL processes petabytes of data offline at a high performance level and supports highly concurrent online queries.

This notebook shows how to use functionality related to the AnalyticDB vector database.

To run, you should have an AnalyticDB instance up and running:


Only available on Node.js.


LangChain.js accepts node-postgres as the connections pool for AnalyticDB vectorstore.

npm install -S pg

And we need pg-copy-streams to add batch vectors quickly.

npm install -S pg-copy-streams
npm install @langchain/openai @langchain/community



User-generated data such as usernames should not be used as input for the collection name.
This may lead to SQL Injection!

import { AnalyticDBVectorStore } from "@langchain/community/vectorstores/analyticdb";
import { OpenAIEmbeddings } from "@langchain/openai";

const connectionOptions = {
host: process.env.ANALYTICDB_HOST || "localhost",
port: Number(process.env.ANALYTICDB_PORT) || 5432,
database: process.env.ANALYTICDB_DATABASE || "your_database",
user: process.env.ANALYTICDB_USERNAME || "username",
password: process.env.ANALYTICDB_PASSWORD || "password",

const vectorStore = await AnalyticDBVectorStore.fromTexts(
["foo", "bar", "baz"],
[{ page: 1 }, { page: 2 }, { page: 3 }],
new OpenAIEmbeddings(),
{ connectionOptions }
const result = await vectorStore.similaritySearch("foo", 1);
// [{"pageContent":"foo","metadata":{"page":1}}]

await vectorStore.addDocuments([{ pageContent: "foo", metadata: { page: 4 } }]);

const filterResult = await vectorStore.similaritySearch("foo", 1, {
page: 4,
// [{"pageContent":"foo","metadata":{"page":4}}]

const filterWithScoreResult = await vectorStore.similaritySearchWithScore(
{ page: 3 }
// [[{"pageContent":"baz","metadata":{"page":3}},0.26075905561447144]]

const filterNoMatchResult = await vectorStore.similaritySearchWithScore(
{ page: 5 }
// []

// need to manually close the Connection pool
await vectorStore.end();

API Reference:

Was this page helpful?

You can also leave detailed feedback on GitHub.