Re: [PATCH] Incremental sort (was: PoC: Partial sort) - Mailing list pgsql-hackers

From James Coleman
Subject Re: [PATCH] Incremental sort (was: PoC: Partial sort)
Date
Msg-id CAAaqYe9np1d9V+mL8Q7CGc=DdjwKz3L5nvDaQo5UHq0kYDHDTA@mail.gmail.com
Whole thread Raw
In response to Re: [PATCH] Incremental sort (was: PoC: Partial sort)  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Mon, Mar 23, 2020 at 1:05 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> > ... all plan types that use only one child use the outer one.  They
> > could use either, as long as it does that consistently, I think.
>
> Yeah, exactly.  The outer/inner terminology is really only sensible
> for join nodes, but there isn't a third child-plan pointer reserved
> for single-child node types, so you gotta use one of those.  And
> conventionally we use the "outer" one.
>
>                         regards, tom lane

Great, thanks for the explanation Alvaro and Tom; I'll fix that up in
my next patch series.

I idly wonder if a macro childPlanState() defined exactly the same as
outerPlanState() might _kinda_ make sense here, but I'm also content
to follow convention.

I might submit a small patch to the comment on those macros though to
expand on the explanation.

James



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: [PATCH] Incremental sort (was: PoC: Partial sort)
Next
From: Andres Freund
Date:
Subject: Re: Assert() failures during RI checks