Re: case insensitive match in unicode

In article <20060327114037.GD30791@svana.org>, kleptog@svana.org says...
> On Mon, Mar 27, 2006 at 12:45:05PM +0200, SunWuKung wrote:
> > This sounds like a very interesting concept.
> > It wouldn't be 'case insensitive' just insensitive.
> >
> > The way I imagine it now is a special case of the ~ function.
> > I create matchgroups in a table and check each character if it is in the
> > group. If it is I will replace the character with the group in [éÉE],
> > [oóOÓ??] and do a regexp with that.
>
> No need to reinvent the wheel. ICU provides a range of services to deal
> with this. For example the following filter in ICU:
>
>  NFD; [:Nonspacing Mark:] Remove; NFC.
>
> Will remove all accents from characters. And it works for all Unicode
> characters. With a bit more thinking you can work with case variations
> also.
>
> There is also a locale-independant case-mapping module there plus
> various locale specific ones also.
>
> http://icu.sourceforge.net/userguide/Transform.html
> http://icu.sourceforge.net/userguide/caseMappings.html
> http://icu.sourceforge.net/userguide/normalization.html
>
> Have a nice day,
>
Thanks, I looked at this and it looks like something that would indeed
solve the problem.
However I was so far unable to figure out how could I use this from
within Postgres. If you have experience with it could you give me an
example?

Thanks
Balázs