On Fri, Feb 26, 2016 at 1:05 AM, Amit Langote
<Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp> wrote:
> I think setting up N ResultRelInfos in advance where the tuple would only
> ever require one might be superfluous. But that may point to some flaw in
> my original design or thinking about the case.
You have a point. But maybe we should get it working first and
optimize it later.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company