Heikki Linnakangas <heikki.linnakangas@enterprisedb.com> writes:
> OFF is a reserved keyword. It's not a reserved keyword in the SQL spec,
> and it's not hard to see people using off as a variable or column name,
> so it would be nice to relax that.
While I can see the value of doing something about that, this seems
awfully fragile:
> + /*
> + * OFF is also accepted as a boolean value, but is not listed
> + * here to avoid making it a reserved keyword. All uses of
> + * opt_boolean rule also accept a ColId with the same action -
> + * OFF is handled via that route.
> + */
The production's correctness now depends on how it's used, and there's
no way to prevent somebody from misusing it.
I think it'd be better if you were to refactor the grammar so that ColId
was actually one of the alternatives in this very production (call it
opt_boolean_or_name, or something like that). Then at least there'd be
less of a flavor of action-at-a-distance about the assumption that OFF
was handled in a compatible fashion.
regards, tom lane