I think my only other thought is this makes using a plain iterator hard? With a larger pages / object size, could this hit the document size limit?

I suppose running two queries could work, but that's still two queries. Plus potential concurrency issues doing so