Is this a new bug?
Current Behavior
I am trying to test the serverless option as we are using a bunch of indexes in our organization. For us is mandatory to delete or at least locate vectorIds by metadata. Are these limitations permanent?
https://docs.pinecone.io/docs/limits#serverless-index-limitations
During ingestion we associate document information that then need to locate all chunks from there during a delete operation. Now we are using a filter like
...
const filter = {
id: identifier
};
...
await namespace.deleteMany(filter);
Expected Behavior
Followed the comments from here but still cannot locate the vectors I need to delete.
https://docs.pinecone.io/docs/manage-rag-documents#fetch-all-records-for-a-parent-document
Steps To Reproduce
create a serverless index, try to delete some vectors by some metadata
...
const filter = {
id: identifier
};
...
await namespace.deleteMany(filter);
Relevant log output
No response
Environment
- **OS**: windows 10
- **Language version**: nodejs 18
- **Pinecone client version**: 2.0.0
Additional Context
Would like to know the timeframe for this support as without this option is not possible to evaluate the new serverless option.
Besides would be nice to have a method on the namespace to be able to list documents, otherwise extrametadata is needed to do a POST as detailed here, we do not have the metadata to fully locate the index location
https://docs.pinecone.io/docs/manage-rag-documents#fetch-all-records-for-a-parent-document
In short for the library it would be grat to
- support the same options as non serverless for delete operations
- otherwise provide an option to get ids from metadata to then do the deletion manually. That again, could be wrapped in the library.
standing by, thanks in advance
Is this a new bug?
Current Behavior
I am trying to test the serverless option as we are using a bunch of indexes in our organization. For us is mandatory to delete or at least locate vectorIds by metadata. Are these limitations permanent?
https://docs.pinecone.io/docs/limits#serverless-index-limitations
During ingestion we associate document information that then need to locate all chunks from there during a delete operation. Now we are using a filter like
Expected Behavior
Followed the comments from here but still cannot locate the vectors I need to delete.
https://docs.pinecone.io/docs/manage-rag-documents#fetch-all-records-for-a-parent-document
Steps To Reproduce
create a serverless index, try to delete some vectors by some metadata
Relevant log output
No response
Environment
Additional Context
Would like to know the timeframe for this support as without this option is not possible to evaluate the new serverless option.
Besides would be nice to have a method on the namespace to be able to list documents, otherwise extrametadata is needed to do a POST as detailed here, we do not have the metadata to fully locate the index location
https://docs.pinecone.io/docs/manage-rag-documents#fetch-all-records-for-a-parent-document
In short for the library it would be grat to
standing by, thanks in advance