Re: Frames vs partitions: is SQL2008 completely insane?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: pgsql-hackers(at)postgreSQL(dot)org, Hitoshi Harada <umi(dot)tanuki(at)gmail(dot)com>
Subject: Re: Frames vs partitions: is SQL2008 completely insane?
Date: 2008-12-26 21:41:01
Message-ID: 15681.1230327661@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I wrote:
> Lastly, for a simple aggregate used with an OVER clause, the current
> patch seems to define the aggregate as being taken over the frame
> rather than the partition, but I cannot find anything in SQL2008 that
> lends any support to *either* definition.

Never mind that --- I found it in 10.9 syntax rule 4.b.

But what this seems to boil down to is that LEAD() and LAST_VALUE()
are completely useless unless you're allowed to specify a nondefault
framing clause ... and don't mind ignoring the clearly-insane
restriction of 6.10 syntax rule 6.b.

The minimum extra functionality needed to make these functions useful
would seem to be to allow UNBOUNDED FOLLOWING ...

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2008-12-26 23:28:28 Tuplestore trimming in window-functions patch
Previous Message Tom Lane 2008-12-26 21:30:20 Re: Window-functions patch handling of aggregates