On Fri, Oct 13, 2006 at 03:40:17PM +0200, Tomi NA wrote:
> This is a reoccurring topic on the list: sure, it's possible to
> misconfigure pg so that uppercase/lowercase/ilike/tsearch2/order don't
> work with a single letter outside of the English alphabet, but the
> problem Martins seems to be facing is one we've seen here before
> (myself being one of those affected). There's no way Martins can set
> up pg - UTF or no UTF - so that collation and case insensitivity-based
> functions work in both Latvian an Russian.
While sorting for multiple languages simultaneously is an issue, that's
not the problem here. Linux/GLibc *does* support correct sorting for
all language/charset combinations, and that's what he's using. Just for
the hell of it I setup lv_LV.utf8 on my laptop and verifed that it
sorts just fine:
$ LC_ALL=lv_LV.utf8 sort < /tmp/test2 | recode utf8..latin1
a
á
B
d
F
ü
Z
$
(hope it looks ok on your screen, the source file is utf8 but the
recode is for display).
Similarly, upper/lower are also supported, although postgresql doesn't
take advantage of the system support in that case.
So yes, PostgreSQL does have some issues in this area, but for the
platform he is using sorting *should* work.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.