John Sidney-Woollett <johnsw@wardbrook.com> writes:
> Do upper() and lower() only work correctly for postgres v8 UTF-8 encoded
> databases? (They don't seem to work on chars > standard ascii on my
> 7.4.6 db). Is this locale or encoding specific issue?
Before 8.0, they don't work on multibyte characters, period. In 8.0
they work according to your locale setting.
> Is there likely to be a significant difference in speed between a
> database using a UTF-8 locale and the C locale (if you don't care about
> the small issues you detailed below)?
I'd expect the C locale to be materially faster for text sorting.
Don't have a number offhand.
regards, tom lane