bytedance/deer-flow · error · ValueError

Fact {fact_id!r} cannot be upserted and deleted together

Error message

Fact {fact_id!r} cannot be upserted and deleted together

What it means

Error "Fact {fact_id!r} cannot be upserted and deleted together" thrown in bytedance/deer-flow.

Source

Thrown at backend/packages/harness/deerflow/agents/memory/backends/deermem/deermem/core/storage.py:763

            if upsert_revisions is not None and fact_id in upsert_revisions:
                expected_fact_revision = upsert_revisions[fact_id]
                if expected_fact_revision is None and existing is not None:
                    raise MemoryFactRevisionConflict(f"Fact {fact_id!r} must not already exist")
                if expected_fact_revision is not None:
                    actual_fact_revision = None if existing is None else existing.get("revision")
                    if actual_fact_revision != expected_fact_revision:
                        raise MemoryFactRevisionConflict(f"Expected fact {fact_id!r} revision {expected_fact_revision}, found {actual_fact_revision}")
            normalized = _normalize_fact(candidate, scope=scope, existing=existing)
            if existing != normalized:
                prepared[fact_id] = (normalized, existing, fact_path)

        delete_ids = [str(fact_id) for fact_id in deletes]
        if len(delete_ids) != len(set(delete_ids)):
            raise ValueError("Duplicate fact ids are not allowed in deletes")
        removals: dict[str, tuple[dict[str, Any], Path]] = {}
        for fact_id in delete_ids:
            if fact_id in prepared:
                raise ValueError(f"Fact {fact_id!r} cannot be upserted and deleted together")
            if agent_name is None:  # guarded above
                raise ValueError("agent_name is required for fact repository changes")
            existing, fact_path = self._read_fact(path, fact_id, user_id=user_id, agent_name=agent_name)
            if existing is None:
                continue
            if delete_revisions and fact_id in delete_revisions and delete_revisions[fact_id] != existing.get("revision"):
                raise MemoryFactRevisionConflict(f"Expected fact {fact_id!r} revision {delete_revisions[fact_id]}, found {existing.get('revision')}")
            removals[fact_id] = (existing, fact_path)

        base = current_memory or create_empty_memory()
        user_section = copy.deepcopy(base.get("user", {}))
        history_section = copy.deepcopy(base.get("history", {}))
        if summaries is not None:
            if not isinstance(summaries, dict):
                raise ValueError("change_set.summaries must be an object")
            if "user" in summaries:
                if not isinstance(summaries["user"], dict):
                    raise ValueError("change_set.summaries.user must be an object")

View on GitHub (pinned to 1dd6ba1acb)

Solutions

  1. Split the change set: delete the fact in one operation or upsert it in another, not both.
  2. Decide the intended final state and send only that operation.

When it happens

Trigger: Thrown at backend/packages/harness/deerflow/agents/memory/backends/deermem/deermem/core/storage.py:763 when the library encounters an invalid state.

Common situations: See trigger scenarios.


AI-assisted analysis of bytedance/deer-flow@1dd6ba1acb (2026-08-14). Data as JSON: /api/errors/08f28311b2460bee. Report an issue: GitHub.