> Shouldn't this be implemented in a more generic manner? An ordinary
Unfortunately I don't see a way for that. GIN is generalized - and it doesn't
know semantic. Actually, we could fix first five strategy numbers for BTree's
strategies, and then we could teach GIN core to use BTRee semantic, but what
about with already existed operator classes?
>
> The documentation describes btree-gin as providing "GIN operator
> classes that implement B-tree equivalent behavior", but now the
> behavior diverges.
Anyway GIN couldn't be used for ORDER BY clause.
--
Teodor Sigaev E-mail: teodor@sigaev.ru
WWW: http://www.sigaev.ru/