We did just that but the datasets we introduced are comercially licensed and they're also from 2010. We keep the osm_id for OSM data so we can distinguish between the two datasets when we update. It took a lot of work to merge those two datasets.