Alban Hertroys <alban@magproductions.nl> writes:
> I'd say "unique constraint violation" right here (provided there's a
> unique constraint on this column, of course). The order in which these
> are inserted/updated doesn't seem to matter either. I'm afraid I'm still
> missing the point...
The point is that btree depends on a number of assumptions about the
behavior of the operators in an operator class, and one of those
assumptions is that "equality" is transitive.
regards, tom lane