Re: Very puzzling sort behavior - Mailing list pgsql-general

From Andrew Sullivan
Subject Re: Very puzzling sort behavior
Date
Msg-id 20150910220317.GT28424@crankycanuck.ca
Whole thread Raw
In response to Re: Very puzzling sort behavior  (Ken Tanzer <ken.tanzer@gmail.com>)
Responses Re: Very puzzling sort behavior
Re: Very puzzling sort behavior
List pgsql-general
On Thu, Sep 10, 2015 at 02:54:31PM -0700, Ken Tanzer wrote:
> Thanks, but I guess I should have been clearer.  Thanks to y'all wonderful
> mailing list folks, I get it now as to why the two sorts are not the same.
> I'm hoping for practical suggestions or advice about how to get C locale
> sorting without having to rewrite all my existing queries.

Why not change the collation for the column?
http://www.postgresql.org/docs/9.4/interactive/sql-altertable.html

> be clinging to futile hope, but is there really no way to specify a
> collation for the return value of a function?

I don't believe so.  I think you need to specify the collation for the
data itself.

It strikes me that you might be able to similate this with a
materialized view or something like that, which has a different
collation than the source table.  That seems like it'd be pretty
awkward, but if there's some reason you can't use C collation on the
source table itself that might work.  I guess you could do the same
thing with a temporary table inside the function, thereby getting a
different collation on the data than in the source table.  These both
seem like hacks, but if you need a bugfix prior to a real solution
they'd give you a path.

A

--
Andrew Sullivan
ajs@crankycanuck.ca


pgsql-general by date:

Previous
From: Ken Tanzer
Date:
Subject: Re: Very puzzling sort behavior
Next
From: Joe Conway
Date:
Subject: Re: Very puzzling sort behavior