Skip to main content


DashVector is a fully-managed vectorDB service that supports high-dimension dense and sparse vectors, real-time insertion and filtered search. It is built to scale automatically and can adapt to different application requirements.

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

To use DashVector, you must have an API key. Here are the installation instructions.


%pip install --upgrade --quiet  dashvector dashscope

We want to use DashScopeEmbeddings so we also have to get the Dashscope API Key.

import getpass
import os

os.environ["DASHVECTOR_API_KEY"] = getpass.getpass("DashVector API Key:")
os.environ["DASHSCOPE_API_KEY"] = getpass.getpass("DashScope API Key:")


from langchain.text_splitter import CharacterTextSplitter
from langchain_community.embeddings.dashscope import DashScopeEmbeddings
from langchain_community.vectorstores import DashVector
from langchain_community.document_loaders import TextLoader

loader = TextLoader("../../modules/state_of_the_union.txt")
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
docs = text_splitter.split_documents(documents)

embeddings = DashScopeEmbeddings()

We can create DashVector from documents.

dashvector = DashVector.from_documents(docs, embeddings)

query = "What did the president say about Ketanji Brown Jackson"
docs = dashvector.similarity_search(query)

We can add texts with meta datas and ids, and search with meta filter.

texts = ["foo", "bar", "baz"]
metadatas = [{"key": i} for i in range(len(texts))]
ids = ["0", "1", "2"]

dashvector.add_texts(texts, metadatas=metadatas, ids=ids)

docs = dashvector.similarity_search("foo", filter="key = 2")
[Document(page_content='baz', metadata={'key': 2})]