Unable to upsert vectors to new serverless vector

Support
,
1

Unable to upsert vectors (via python) to new serverless index. However the complex process of upserting works very well for pod index. Currently i am using pip install pinecone-client==2.2.2, pip install pinecone-client[grpc] in requirements.txt, should i upgrade to => 3.0.0 pinecone? I have 300,000 vectors and for pod index i ran batch process, should i run similar or will it be completly different, many thanks

2

Hi @sameerwpat and welcome to the Pinecone community forums!

Thank you for your question.

Yes, please upgrade to Pinecone client version 3.0.0 or above. This version of the client and newer has support for serverless included.

The main thing that will change in your code after the upgrade is the way that you create serverless indexes via code:

from pinecone.grpc import PineconeGRPC as Pinecone
from pinecone import ServerlessSpec

pc = Pinecone(api_key="YOUR_API_KEY")

pc.create_index(
  name="serverless-index",
  dimension=1536,
  metric="cosine",
  spec=ServerlessSpec(
    cloud="aws",
    region="us-east-1"
  )
)

See more in our official docs here.

Hope that helps!

Best,
Zack