At the same time, if a URL wasn't scraped at all on this run, we want to remove it from the table entirely. I know this must involve some kind of a merge, but I'm having trouble with the details.