Re: Open issues for collations - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Open issues for collations
Date
Msg-id 1303156367.14718.8.camel@vanquo.pezone.net
Whole thread Raw
In response to Re: Open issues for collations  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Mon, 2011-04-18 at 10:41 -0400, Tom Lane wrote:
> However, I've come across a new issue that maybe requires discussion:
> what collation should be associated with a multi-row VALUES in FROM?
> For instance, in
> 
>         SELECT ... FROM
>           (VALUES (1, 'foo'), (2, 'bar' COLLATE "C")) v(a,b),
>           ...
> 
> what collation should be imputed to references to v.b?
> 
> The way the code currently works is that the first row of the VALUES
> list is inspected to determine what collation to report

Hmm, I do see this on my list of things to address, so yes, it should be
fixed.  I likely got stuck because it's pretty complicated.



pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: Typed table DDL loose ends
Next
From: Peter Eisentraut
Date:
Subject: ORDER BY 1 COLLATE