For multiset tables, I'd like to be able to track as a metric the number of distinct rows in a dataset. Currently, this is possible for individual column values, but not for entire rows. This would allow me to keep track of whether I've removed any important columns or need to perform aggregations upstream.
