Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> It's probably also worth noting that the reason I used DISTINCT
> originally is because it's already a keyword.
True.
It occurs to me that the pg_stats view already exposes "n_distinct"
as a column name. I wouldn't object to using "n_distinct" and
"n_distinct_inherited" or some such.
regards, tom lane