Martijn van Oosterhout escribió:
> The most surprising this is that to_ascii won't accept a bytea. TBH the
> whole to_ascii function seems somewhat half-baked. If what you're
> trying to do is remove accents, there are perl functions around that do
> that. Basically, the switch to a different normal form then drop all
> the accent characters.
Yeah, it's been a common suggestion to use convert() in combination with
to_ascii on UTF-8 databases, and I didn't notice that the convert()
shuffling would take that ability away :-( I don't think requiring
plperl is nice however. Perhaps we could get around the problem by
using byteaout/textin.
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.