Well, the file does not necessarily hold all the meta-data. So I guess that would be a means to “push” the meta-data into the DB along with the file-tag.
Letting the meta-data aside, yes, it would boil down to a single file-scan request.
A single-file metadata-update request would come in handy too, I guess, if somewhen the webinterface would be capable of managing tags and such.