Find answers from the community

Updated 2 months ago

Uploading large files to a vector store with Pinecone

At a glance

The community member is facing issues uploading large amounts of data to a Pinecone vector store. They are encountering timeout errors when using the ingestion pipeline, and the parallel processing approach is resulting in a "Resource already exists" error (409). The community members have suggested trying to lower the batch_size parameter in the PineconeVectorStore constructor to address the timeout issue, but the community member is still concerned about how to upload large amounts of data in less time. There is no explicitly marked answer in the comments.

Hii dudes, when i upload embeddings to pinecone using ingestion pipeline(by pipeline.run) it interrupts in the middle due to the timeout error. and when i try with parrallel processing using " num_workers "option it interrupts with the error409- error":{"code":"ALREADY_EXISTS","message":"Resource already exists"},"status":409. what is the sollution to upload large amount of file to the vectorstore? @Logan M
L
l
4 comments
PineconeVectorStore has a batch_size param in the constructor, try lowering that if it times out. Default is 100
why not parallel processing worked? why the error 409 occurs?
this method worked for uploading the data. my concern is about how to upload large amount of data in less time?
@Logan M please suggest a solution
Add a reply
Sign up and join the conversation on Discord