What causes "document not found" in infiniflow/ragflow?

Treat the returned error text as per-file: remove the stale id from the request and re-submit the remainder.

infiniflow/ragflow · warning · RuntimeError

document not found

Error message

document not found

What it means

Raised in FileService.delete_docs when DocumentService.get_by_id(doc_id) returns not-found for one of the submitted document ids. The per-doc try block catches it and appends it to the errors string returned to the caller, so one bad id does not abort the whole batch.

Source

Thrown at api/db/services/file_service.py:715

    @staticmethod
    def put_blob(user_id, location, blob):
        bname = f"{user_id}-downloads"
        return settings.STORAGE_IMPL.put(bname, location, blob)

    @classmethod
    @DB.connection_context()
    def delete_docs(cls, doc_ids, tenant_id):
        root_folder = FileService.get_root_folder(tenant_id)
        pf_id = root_folder["id"]
        FileService.init_knowledgebase_docs(pf_id, tenant_id)
        errors = ""
        kb_table_num_map = {}
        for doc_id in doc_ids:
            try:
                e, doc = DocumentService.get_by_id(doc_id)
                if not e:
                    raise RuntimeError("document not found")
                tenant_id = DocumentService.get_tenant_id(doc_id)
                if not tenant_id:
                    raise RuntimeError("Tenant not found!")

                b, n = File2DocumentService.get_storage_address(doc_id=doc_id)

                TaskService.filter_delete([Task.doc_id == doc_id])
                if not DocumentService.remove_document(doc, tenant_id):
                    raise RuntimeError("Database error (Document removal)!")

                f2d = File2DocumentService.get_by_document_id(doc_id)
                deleted_file_count = 0
                if f2d:
                    deleted_file_count = FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
                File2DocumentService.delete_by_document_id(doc_id)
                if deleted_file_count > 0:
                    settings.STORAGE_IMPL.rm(b, n)

View on GitHub (pinned to 554fb1133a)

Solutions

Treat the returned error text as per-file: remove the stale id from the request and re-submit the remainder.
Before deleting, verify each id with DocumentService.get_by_id or the document-list API.
Make deletion idempotent in your client: treat 'document not found' as success.
Check for duplicate ids in the payload.

Example fix

# before
FileService.delete_docs(['doc1', 'doc1'], tenant_id)  # second lookup fails

# after
FileService.delete_docs(list(dict.fromkeys(doc_ids)), tenant_id)  # dedupe

Defensive patterns

Strategy: try-catch

Validate before calling

doc_ids = list(dict.fromkeys(doc_ids))  # dedupe
valid_ids = [d for d in doc_ids if DocumentService.get_by_id(d)[0]]

Try / catch

errors, _ = FileService.delete_docs(doc_ids, tenant_id)
if 'not found' in errors:
    # treat as already-deleted: idempotent success
    pass

Prevention

Make client-side deletion idempotent: 'not found' means already deleted.
Deduplicate id lists before submitting batch deletes.
Refresh document lists after deletions instead of reusing stale ids.

When it happens

Trigger: Calling the document deletion API with a doc id that was already deleted, belongs to another tenant, or is malformed; duplicate ids in the list where the first delete removes the doc and the second lookup fails.

Common situations: Stale document id in the UI after deletion in another tab; retrying a partially completed delete request; scripts re-running against an already-cleaned dataset.

Related errors

AI-assisted analysis of infiniflow/ragflow@554fb1133a (2026-08-15). Data as JSON: /api/errors/98b4e435a9955ea5. Report an issue: GitHub.