Re: [PATCH] pg_permissions - Mailing list pgsql-hackers

From Tom Lane
Subject Re: [PATCH] pg_permissions
Date
Msg-id 1577736.1616764567@sss.pgh.pa.us
Whole thread Raw
Responses Re: [PATCH] pg_permissions  ("Joel Jacobson" <joel@compiler.org>)
List pgsql-hackers
Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> On 2021-Mar-26, Joel Jacobson wrote:
>> On Thu, Mar 25, 2021, at 17:51, Tom Lane wrote:
>> I wonder what performance will be like with lots o' objects.

> I guess he is concerned about the number of catalog accesses.

My concern is basically that you're forcing the join between
pg_shdepend and $everything_else to be done as a nested loop.
It will work well, up to where you have so many objects that
it doesn't ... but the planner will have no way to improve it.

Having said that, I don't really see a better way either.
Materializing $everything_else via a UNION ALL seems like
no fun from a maintenance perspective, plus we're not that
great on optimizing such constructs either.

            regards, tom lane



pgsql-hackers by date:

Previous
From: David Steele
Date:
Subject: Re: Stronger safeguard for archive recovery not to miss data
Next
From: Andrew Dunstan
Date:
Subject: Re: Allow matching whole DN from a client certificate