You will need to hack your own version of strxfrm which takes an aragument for locale. If you mess with locale,you will corrupt the indexes I've been told.
Greg Stark wrote:
Dennis Gearon <[EMAIL PROTECTED]> writes:
I was thinking of INGNORING locale, since it is basically fixed for a DB for long periods of time.
If a table/column HAD it's own locale, that could be used,
but I was more interested in a function taht would allow the explicit
declaration of the encoding(s) to look for.
Indeed for my purposes that's what I'll have to do.
but the strxfrm function uses the current application locale, so I'll have to call setlocale to set it, call strxfrm, then call setlocale to set it back.
I fear that some implementations might do a lot of work when setlocale is called loading large data files and might leak memory expecting it to only be called once at program initialization. That would suck
BTW, what is l10n
l10n = localization i18n = internationalization
arguably i should have said i18n actually.
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
