Re: MERGE SQL statement for PG12

Hi,

On 2019-01-15 14:05:25 -0500, Robert Haas wrote:
> On Mon, Jan 14, 2019 at 1:37 AM Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com> wrote:
> > Can you please help me understand what's fundamentally wrong with
> > the approach and more importantly, can you please explain what would
> > the the architecturally sound way to do this? The same also applies
> > to the executor side where the current approach is deemed wrong, but
> > very little is said on what's the correct way.
>
> [ Long and good explanation by Robert ]

> I want to point out that it is not as if nobody has reviewed this
> patch previously. Here is Andres making basically the same point
> about parse analysis that I'm making here -- FWIW, I didn't find his
> reply until after I'd written the above:
>
> https://www.postgresql.org/message-id/20180403021800.b5nsgiclzanobiup%40alap3.anarazel.de
>
> Here he is explaining these points some more:
>
> https://www.postgresql.org/message-id/20180405200003.gar3j26gsk32gqpe%40alap3.anarazel.de
>
> And here's Peter Geoghegan not only explaining the same problem but
> having a go at fixing it:
>
> https://www.postgresql.org/message-id/CAH2-Wz%3DZwNQvp11XjeHo-dBLHr9GDRi1vao8w1j7LQ8mOsUkzw%40mail.gmail.com
>
> Actually, I see that Peter Geoghegan not just the emails above but a
> blizzard of other emails explaining the structural problems with the
> patch, which I now see include not only the parse analysis concerns
> but also the business of multiple RTEs which I mentioned above.

+ many.

Pavan, I think the reason you're not getting much further feedback is
that you and Simon have gotten a lot and only incorporated feedback only
very grudgingly, if at all. You've not even attempted to sketch out a
move of the main merge handling from parse-analysis to planner, as far
as I can tell, despite this being one of the main criticisms for about a
year. Given that not much besides rebasing has happened since v11, I
don't find it surprising that people don't want to invest more energy in
this patch.

Greetings,

Andres Freund