Re: [GENERAL] ascii() for utf8

From: Alvaro Herrera
> Decibel! wrote:
> > Moving to -hackers.
> >
> > On Jul 27, 2007, at 1:22 PM, Stuart wrote:
> >> Does Postgresql have a function like ascii() that will
> >> return the unicode codepoint value for a utf8 character?
> >> (And symmetrically same for question chr() of course).
>
> > I suspect that this is just a matter of no one scratching the itch. I
> > suspect a patch would be accepted, or you could possibly put something on
> > pgFoundry.
>
> Nay; there were some discussions about this not long ago, and I think
> one conclusion you could draw from them is that many people want these
> functions in the backend.

That would certainly be my preference. I will be distributing an
application, the database part of which may (not sure yet) require
this function, to multiple platforms including Windows and (though
I have never done it) am anticipating it will be significantly harder
if I have to worry about the recipient compiling an external function
or making sure a dll goes in the right place, gets updated, etc.

> > I'd set it up so that ascii() and chr() act according to the
> > appropriate locale setting (I'm not sure which one would be appropriate).
>
> I don't see why any of them would react to the locale, but they surely
> must honor client encoding.

Wouldn't this be the database encoding? (I have been using
strictly utf-8 and admit I am pretty fuzzy on encoding issues.)

If one had written an external function, how much more effort
would it be to make it acceptable for inclusion in the backend?