MIME-Version: 1.0
References: <7f9f9b1d-b612-d0bd-4674-cfb5a8b2d343@gmail.com>
In-Reply-To: <7f9f9b1d-b612-d0bd-4674-cfb5a8b2d343@gmail.com>
From: Samed YILDIRIM <samed@reddoc.net>
Date: Sun, 27 Nov 2022 18:23:38 +0200
Message-ID: 
 <CAAo1mb=+SrSf9h7B55yGOu+Q70Suq7oQ0BQ3QuPJmooOM9MkZA@mail.gmail.com>
Subject: Re: [QUESTION] Window function with partition by and order by
To: Ankit Kumar Pandey <itsankitkp@gmail.com>
Cc: pgsql-sql@lists.postgresql.org
Content-Type: multipart/alternative; boundary="00000000000006ac4105ee762ef3"
Archived-At: 
 <https://www.postgresql.org/message-id/CAAo1mb%3D%2BSrSf9h7B55yGOu%2BQ70Suq7oQ0BQ3QuPJmooOM9MkZA%40mail.gmail.com>
Precedence: bulk

--00000000000006ac4105ee762ef3
Content-Type: text/plain; charset="UTF-8"

Hello Ankit,

It is absolutely expected behaviour of a window function with ORDER BY
clause. The default frame clause of window definition is *RANGE BETWEEN
UNBOUNDED PRECEDING AND CURRENT ROW*. If you add an ORDER BY clause in a
window definition, PostgreSQL takes the current row and all rows before it
within the partition into calculation. If you don't add, it means all rows
within the partition are peers, and PostgreSQL uses all rows for
calculation. I'm putting the related part from the documentation and its
link below.

The default framing option is RANGE UNBOUNDED PRECEDING, which is the same
> as RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW; it sets the frame to
> be all rows from the partition start up through the current row's last peer
> (a row that the window's ORDER BY clause considers equivalent to the
> current row; all rows are peers if there is no ORDER BY).


https://www.postgresql.org/docs/15/sql-select.html#SQL-WINDOW

Best regards.
Samed YILDIRIM


On Sun, 27 Nov 2022 at 18:08, Ankit Kumar Pandey <itsankitkp@gmail.com>
wrote:

> Hello,
>
> While looking at aggregates in window function, I found something
> unusual and would be glad I could get some clarification.
>
> Consider following table (mytable):
>
> id, name
>
> 1, A
>
> 1, A
>
> 2, B
>
> 3, A
>
> 1, A
>
>
> select *, avg(id) over (partition by name, order by id) from mytable;
>
> Output:
>
> id, name, avg
>
> 1, A, 1
>
> 1, A, 1
>
> 1, A, 1
>
> 3, A, 1.5
>
> 2, B, 2
>
>
> Question is: Average of id for partition name (A) should be 6/4 = 1.5
> for all rows in that partition but this result is seen only at the last
> one row in partition (A). Am I missing here something?
>
>
> Thanks
>
>
> --
> Regards,
> Ankit Kumar Pandey
>
>
>
>

--00000000000006ac4105ee762ef3
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Hello Ankit,<br></div><div><br></div><div>It is absol=
utely expected behaviour of a window function with ORDER BY clause. The def=
ault frame clause of window definition is <b><span style=3D"background-colo=
r:rgb(255,255,255)">RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW</span=
></b><span style=3D"background-color:rgb(255,255,255)">. If you add an ORDE=
R BY clause in a window definition, PostgreSQL takes the current row and al=
l rows before it within the partition into calculation. If you don&#39;t ad=
d, it means all rows within the partition are peers, and PostgreSQL uses al=
l rows for calculation. I&#39;m putting the related part from the documenta=
tion and its link below.<br></span></div><div><span style=3D"background-col=
or:rgb(255,255,255)"><br></span></div><blockquote class=3D"gmail_quote" sty=
le=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);paddi=
ng-left:1ex">The default framing option is RANGE UNBOUNDED PRECEDING, which=
 is the same as RANGE BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW; it sets =
the frame to be all rows from the partition start up through the current ro=
w&#39;s last peer (a row that the window&#39;s ORDER BY clause considers eq=
uivalent to the current row; all rows are peers if there is no ORDER BY).</=
blockquote><div>=C2=A0</div><div><a href=3D"https://www.postgresql.org/docs=
/15/sql-select.html#SQL-WINDOW">https://www.postgresql.org/docs/15/sql-sele=
ct.html#SQL-WINDOW</a></div><div><br></div><div><div><div dir=3D"ltr" class=
=3D"gmail_signature" data-smartmail=3D"gmail_signature"><div dir=3D"ltr"><d=
iv>Best regards.</div><div>Samed YILDIRIM<br></div></div></div></div><br></=
div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_at=
tr">On Sun, 27 Nov 2022 at 18:08, Ankit Kumar Pandey &lt;<a href=3D"mailto:=
itsankitkp@gmail.com">itsankitkp@gmail.com</a>&gt; wrote:<br></div><blockqu=
ote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px=
 solid rgb(204,204,204);padding-left:1ex">Hello,<br>
<br>
While looking at aggregates in window function, I found something <br>
unusual and would be glad I could get some clarification.<br>
<br>
Consider following table (mytable):<br>
<br>
id, name<br>
<br>
1, A<br>
<br>
1, A<br>
<br>
2, B<br>
<br>
3, A<br>
<br>
1, A<br>
<br>
<br>
select *, avg(id) over (partition by name, order by id) from mytable;<br>
<br>
Output:<br>
<br>
id, name, avg<br>
<br>
1, A, 1<br>
<br>
1, A, 1<br>
<br>
1, A, 1<br>
<br>
3, A, 1.5<br>
<br>
2, B, 2<br>
<br>
<br>
Question is: Average of id for partition name (A) should be 6/4 =3D 1.5 <br=
>
for all rows in that partition but this result is seen only at the last <br=
>
one row in partition (A). Am I missing here something?<br>
<br>
<br>
Thanks<br>
<br>
<br>
-- <br>
Regards,<br>
Ankit Kumar Pandey<br>
<br>
<br>
<br>
</blockquote></div>

--00000000000006ac4105ee762ef3--