Re: Change of database characterset
From: joel garry <joel-garry_at_home.com>
Date: Fri, 9 Jan 2009 10:00:14 -0800 (PST)
Message-ID: <043315d9-e074-470d-bb63-16f0382f365b_at_g3g2000pre.googlegroups.com>
On Jan 9, 9:46 am, trub3101 <trub3..._at_sky.com> wrote:
> Hi all,
>
> I am about to change our dev database characterset from WE8ISO8859P1
> to UTF8 (new database,export/import) to bring dev in line with
> production.
>
> However, before I do this I would be grateful if someone could explain
> the fundamental differences between the two charactersets i.e. is
> there any possibility of data lost? Indeed is this conversion
> necessary?
>
> Many thanks in advance,
>
> tb3101
Date: Fri, 9 Jan 2009 10:00:14 -0800 (PST)
Message-ID: <043315d9-e074-470d-bb63-16f0382f365b_at_g3g2000pre.googlegroups.com>
On Jan 9, 9:46 am, trub3101 <trub3..._at_sky.com> wrote:
> Hi all,
>
> I am about to change our dev database characterset from WE8ISO8859P1
> to UTF8 (new database,export/import) to bring dev in line with
> production.
>
> However, before I do this I would be grateful if someone could explain
> the fundamental differences between the two charactersets i.e. is
> there any possibility of data lost? Indeed is this conversion
> necessary?
>
> Many thanks in advance,
>
> tb3101
Yes. See metalink notes 469020.1 and 124721.1 (among many others, metalink and google are useful!), and the docs on csscan. The basic issue is, some things that are single characters in P1 become multiple characters in utf, so blow column length.
It is not necessary, for some definitions of necessary. It _is_ necessary to keep your dev and prod the same, and it is much easier to avoid data loss if you don't have conversions.
jg
-- _at_home.com is bogus. 'Fantastic! We have a comic-book geek in the White House.' - Joe QuesadaReceived on Fri Jan 09 2009 - 12:00:14 CST