On Tue, Aug 16, 2016 at 2:43 AM, Alexander Korotkov
<a.korotkov@postgrespro.ru> wrote:g of pg_operator.oprassociative...
> Another problem is computational complexity of such transformations. AFAIR,
> few patches for making optimizer smarter with expressions were already
> rejected because of this reason.
s/few/many/
> Also, even if we would have such transformation of expressions, floating
> points types would be still problematic, because (a + b) + c = a + (b + c)
> is not exactly true for them because of computational error.
Right. I think matching expressions exactly is plenty good enough.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company