--- In firebird-support@yahoogroups.com, Helen Borrie <helebor@...> wrote:

> There is no such thing as a "collation
> set". There is a CHARACTER SET - in your case
> ISO8859_1 - that determines which set of
> character images responds to the bytes that you
> store in your character fields. Then, there is
> COLLATION, which determines which code page your
> stored data refers to. Collations generally
> apply the rules that determine things like
> lower-case/upper-case correspondences and the
> sorting order of the characters in local dictionaries.
>
> So - the attributes of the DATA will be no
> different, since you will not be changing the CHARACTER SET.
>
> The pieces you need to attend to are those that
> have COLLATE clauses set on objects that affect
> output and searching. They will typically be
> distributed between your metadata (domain, field
> and index definitions) and your applications
> (ORDER BY and GROUP BY statements). It is likely
> also that you have applied collation settings to
> variables and statements in stored procedures and triggers.
>
> By the way, you don't absolutely need a
> "waterfall" approach to this change. It is quite
> normal to apply different collations to different
> tasks in a database, i.e. moving from the
> existing case-sensitive collation to the new
> case-insensitive one can be as selective as you need it to be.
>
> ./heLen
>

Hello Helen.

I know I can use a different collate in a particular query or command,
but I prefer to not worry about that, define it once in domains that
uses all my fields, and lets the engine to ever use the same collate.

I think I will create an empty database, that uses the news collates
in Firebird 2, and pump data to them.

Thanks for your advises.