Horton

Re: MergeAppend could consider sorting cheapest child path

Andrei Lepikhov <lepihov@gmail.com>

View thread
From: Andrei Lepikhov <lepihov@gmail.com>
To: Alexander Pyhalov <a.pyhalov@postgrespro.ru>
Cc: Andy Fan <zhihuifan1213@163.com>, Bruce Momjian <bruce@momjian.us>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Alexander Korotkov <aekorotkov@gmail.com>, Nikita Malakhov <HukuToc@gmail.com>
Date: 2025-05-07T09:03:04Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Mostly-cosmetic adjustments to estimate_multivariate_bucketsize().

  2. Consider fractional paths in generate_orderedappend_paths

On 7/5/2025 08:57, Alexander Pyhalov wrote:
> Andrei Lepikhov писал(а) 2025-05-07 08:02:
>> On 5/5/2025 15:56, Alexander Pyhalov wrote:
>>> Andrei Lepikhov писал(а) 2025-05-05 14:38:
>>> Also logic a bit differs if path is NULL. In 
>>> get_cheapest_path_for_pathkeys_ext() we explicitly check for path 
>>> being NULL, in get_cheapest_fractional_path_for_pathkeys_ext() only 
>>> after calculating sort cost.
>>>
>>> I've tried to fix comments a bit and unified functions definitions.
>> Generally seems ok, I'm not a native speaker to judge the comments. But:
>> if (base_path && path != base_path)
>>
>> What is the case in your mind where the base_path pointer still may be 
>> null at that point?
> 
> Hi.
> 
> It seems if some childrel doesn't have valid pathlist, subpaths_valid 
> would be false in add_paths_to_append_rel()
> and generate_orderedappend_paths() will not  be called. So when we 
> iterate over live_childrels,
> all of them will have cheapest_total path. This is why we can assert 
> that base_path is not NULL.
I'm not sure I understand you correctly. Under which conditions will 
rel->cheapest_total_path be set to NULL for a childrel? Could you 
provide an example?

-- 
regards, Andrei Lepikhov