"dev" <dev(at)strategma(dot)bg> writes:
> my database is in UNICODE
> I try this (with Bulgarian chars):
> SELECT ' ' ~* '' as result
> returns FALSE
> above query is same as:
> SELECT 'Test exam' ~* 'test' as result
> retuns TRUE
The regexp stuff is not currently capable of dealing with case
conversions in multi-byte character sets (such as Unicode).
Is there a Latin-N encoding you could use?
regards, tom lane