- Deletion is permanent and cannot be undone
- Underlying vector data remains intact
- Queries will fall back to sequential scan
- Running queries during deletion may be slower
- Use run_with_orchestration=True for large indices to prevent timeouts
- Consider index dependencies before deletion
Authorizations
The access token received from the authorization server in the OAuth 2.0 flow.
Path Parameters
The table of vector embeddings to delete (e.g. vectors, entity, document_collections)
Available options:
chunks
, documents_entities
, graphs_entities
, graphs_communities
The name of the index to delete
Response
200