vignesh C <vignes...@gmail.com> writes:
> On Wed, 19 Mar 2025 at 18:12, David Rowley <dgrowle...@gmail.com> wrote:
>> While VACUUM ONLY on a partitioned table has no effect, the same isn't
>> true for inheritance parents. 62ddf7ee9 did change the behaviour of
>> VACUUM for these so that vacuuming the inheritance parent now vacuums
>> all of its children, unless ONLY is used. So, I'd say if the tab
>> completion is only added to ANALYZE, then it'd be incomplete still.
>> (I've not looked at the v3 patch to see which of those it handles.)
> I also felt it is necessary for both ANALYZE and VACUUM, and the v3
> patch includes changes for both.
Agreed.
Pushed with some cosmetic adjustments. I made the order of the
options match the syntax diagrams in the comments, which is just
neatnik-ism (it changes no behavior) but seemed to read better.
I also realized that we could simplify the match patterns for the
various VACUUM cases by using MatchAnyN --- the existing code was
getting rather contorted there. That wasn't something your patch
introduced, but might as well fix it while we're here.
regards, tom lane
