Re: tuple radix sort

From: John Naylor <johncnaylorls(at)gmail(dot)com>
To: cca5507 <cca5507(at)qq(dot)com>
Cc: zengman <zengman(at)halodbtech(dot)com>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: tuple radix sort
Date: 2026-04-08 07:17:22
Message-ID: CANWCAZZzXHoDpCx394OAhu0ppW_OanbJzYifVwuALeNMhmF8TA@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Apr 1, 2026 at 4:27 PM cca5507 <cca5507(at)qq(dot)com> wrote:
>
> Hi John,
>
> How about adding an assertion here:

> + Assert(next_level > level);
> +

Good idea.

I also thought we should change this cast:

if (ssup->comparator == ssup_datum_signed_cmp)
{
- norm_datum1 = orig + ((uint64) PG_INT64_MAX) + 1;
+ norm_datum1 = orig + (Int64GetDatum(PG_INT64_MAX)) + 1;
}

Upthread you mention something else about treating Datum as an
integer, but I'm not sure if this is what you meant since you didn't
say. If you have concrete suggestions, feel free to share them.

--
John Naylor
Amazon Web Services

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message John Naylor 2026-04-08 07:22:27 Re: vectorized CRC on ARM64
Previous Message Lakshmi N 2026-04-08 07:09:23 DOCS: pg_plan_advice minor doc fixes