Trouble Storing PDF Document Embeddings in Pinecone Index

Support
1

Captura de pantalla 2023-04-19 a las 18.42.10
Captura de pantalla 2023-04-19 a las 18.42.101241×741 36.5 KB

Hi,
I am encountering difficulties in storing PDF document embeddings into my Pinecone index. I used Langchain to split the documents into chunks and then converted them into OpenAI embeddings. When I check the embeddings locally, they appear to have been generated correctly. However, upon sending them to Pinecone, the vectors appear empty in the index dashboard. The expected quantity of vectors arrives and they are sorted into the correct namespaces, but when querying the vector store, they show up empty. I have attempted to delete my index and create a new one, but to no avail. Additionally, I would appreciate any advice on how to change the region of my index in Pinecone.

Thank you in advance for any assistance you can provide.

2

Hey Vin,

Apologies for the late reply. Without more context, I am unable to determine whether or not the issue was with your index/upsertions, or the use of the legacy Index Browser. Since this post, we have revamped and released a new Index Browser and I’d be curious if you still face this problem.

Additionally, I would appreciate any advice on how to change the region of my index in Pinecone.

Currently the only way to change the region of your index, would be to create a new index in a new project with the desired region. The region of your index is determined by your project environment:

Closing this one out for now, please feel free to repost if this is unresolved for you. Thanks for your patience.