From: | David Rowley <dgrowleyml(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
Subject: | Re: Speeding up ruleutils' name de-duplication code, redux |
Date: | 2024-09-10 20:33:59 |
Message-ID: | CAApHDvqMRD10-HEK5eXZDgo1NqjizdSg8Taqbakno1kAxaNuyA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Wed, 11 Sept 2024 at 03:06, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> We could accomplish what you suggest by re-ordering the calls so that
> we build the hash table before enlarging the array. 0001 attached
> is the same as before (modulo line number changes from being rebased
> up to HEAD) and then 0002 implements this idea on top. On the whole
> though I find 0002 fairly ugly and would prefer to stick to 0001.
> I really doubt that scanning any newly-created column positions is
> going to take long enough to justify intertwining things like this.
I'm fine with that. I did test the performance with and without
v2-0002 and the performance is just a little too noisy to tell. Both
runs I did with v2-0002, it was slower, so I agree it's not worth
making the code uglier for.
I've no more comments. Looks good.
David
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2024-09-10 20:36:00 | Re: Speeding up ruleutils' name de-duplication code, redux |
Previous Message | Tom Lane | 2024-09-10 20:24:43 | Re: [BUG?] XMLSERIALIZE( ... INDENT) won't work with blank nodes |