On Mon, Nov 11, 2019 at 5:56 PM Alvaro Herrera <alvhe...@2ndquadrant.com> wrote:
> First, I noticed that there's a significant unanswered issue from Andrew > Gierth about this using a completely different mechanism, namely an > implicit window function. I see that Robert was concerned about the > performance of Andrew's proposed approach, but I don't see any reply > from Surafel on the whole issue. > > i don't know much about window function implementation but am sure teaching window function about limit and implementing limit *variant on top of it will not be much simpler * *and performant than the current implementation. i also think more appropriate place to * *include limit variant is limitNode not other place which can case redundancy and * *maintainability issue * *regards * *Surafel *