MIME-Version: 1.0
References: <98fd5198e1fe4e162f1166fd0575844de8bb7714.camel@loonybin.net> <CAKAnmmLBxrQ0n4+BPb_xBQKrDRoiq4eSKdQYDbnmo7KTQVFuRg@mail.gmail.com>
In-Reply-To: <CAKAnmmLBxrQ0n4+BPb_xBQKrDRoiq4eSKdQYDbnmo7KTQVFuRg@mail.gmail.com>
From: Isaac Morland <isaac.morland@gmail.com>
Date: Wed, 22 May 2024 11:13:12 -0400
Message-ID: <CAMsGm5e5=YcjaUGQogkxMqrj9_beuSmVOrXTAuQPsBdmkbZPSA@mail.gmail.com>
Subject: Re: Finding "most recent" using daterange
To: Greg Sabino Mullane <htamfids@gmail.com>
Cc: Rob Foehl <rwf@loonybin.net>, pgsql-general@lists.postgresql.org
Content-Type: multipart/alternative; boundary="0000000000003861d206190c6061"
Archived-At: <https://www.postgresql.org/message-id/CAMsGm5e5%3DYcjaUGQogkxMqrj9_beuSmVOrXTAuQPsBdmkbZPSA%40mail.gmail.com>
Precedence: bulk

--0000000000003861d206190c6061
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, 22 May 2024 at 10:15, Greg Sabino Mullane <htamfids@gmail.com>
wrote:

> This is a good candidate for a window function. Also note that nulls
> already get sorted correctly by the DESC so no need to get 'infinity'
> involved, although you could write 'DESC NULLS FIRST' to be explicit abou=
t
> it.
>
> with x as (select *,  row_number() over (partition by id order by
> upper(dates) desc, lower(dates) desc) from example)
>   select id,value,dates from x where row_number =3D 1;
>

Don=E2=80=99t you need NULLS LAST for the lower bounds? There NULL means so=
mething
closer to -infinity and should appear after the non-NULL values in a
descending sort.

Actually it strikes me that this sorting issue could be a reason to avoid
NULL bounds on ranges and prefer the use of +/-infinity if the underlying
data type supports it.

--0000000000003861d206190c6061
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr">On Wed, 22 May 2024 at 10:15, Greg Sabino=
 Mullane &lt;<a href=3D"mailto:htamfids@gmail.com">htamfids@gmail.com</a>&g=
t; wrote:<br></div><div class=3D"gmail_quote"><blockquote class=3D"gmail_qu=
ote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-st=
yle:solid;border-left-color:rgb(204,204,204);padding-left:1ex"><div dir=3D"=
ltr">This is a good candidate for a window function. Also note that nulls a=
lready get sorted correctly by the DESC so no need to get &#39;infinity&#39=
; involved, although you could write &#39;DESC NULLS FIRST&#39; to be expli=
cit about it.<div><br></div><div>with x as (select *, =C2=A0row_number() ov=
er (partition by id order by upper(dates) desc, lower(dates) desc) from exa=
mple)</div><div>=C2=A0 select id,value,dates from x where row_number =3D 1;=
</div></div></blockquote><div><br></div><div>Don=E2=80=99t you need NULLS L=
AST for the lower bounds? There NULL means something closer to -infinity an=
d should appear after the=C2=A0non-NULL values in a descending sort.</div><=
div><br></div><div>Actually it strikes me that this sorting issue could be =
a reason to avoid NULL bounds on ranges and prefer the use of=C2=A0+/-infin=
ity if the underlying data type supports it.</div></div></div>

--0000000000003861d206190c6061--