Re: Out of date comment in cached_plan_cost

On Mon, Dec 11, 2017 at 5:00 AM, David Rowley
<david(dot)rowley(at)2ndquadrant(dot)com> wrote:
> On 11 December 2017 at 21:39, Ashutosh Bapat
> <ashutosh(dot)bapat(at)enterprisedb(dot)com> wrote:
>> I don't see much difference in the old and new wording. The word
>> "generally" confuses more than clarifying the cases when the planning
>> cost curves do not change with the number of relations i.e.
>> partitions.
>
> I added that to remove the false claim that inheritance children don't
> make the join problem more complex. This was only true before we had
> partition-wise joins.
>
> I've re-read my original patch and I don't really see the problem with
> it. The comment is talking about inheritance child relations, which
> you could either interpret to mean INHERITS (sometable), or some table
> listed in pg_inherits. The statement that I added forces me into
> thinking of the former rather than the latter, so I don't really see
> any issue.
>
> I'm going to leave it here. I don't want to spend too much effort
> rewording a comment. My vote is for the original patch I sent. I only
> changed it because Robert complained that technically an inheritance
> child could actually be a partition.

I basically feel like we're not really solving any problem by changing
this. I mean, partition-wise join makes this statement less true, but
adding the word "generally" doesn't really help anybody understand the
situation better. If we're going to add anything here, I think it
should be something like:

(This might need to be rethought in light of partition-wise join.)

If that's more specific than we want to get, then let's just leave it
alone. Partition-wise join isn't even enabled by default at this
point.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company