Thank you again for the clarification! > I'm not sure how to explain it any more clearly: the backslash in this > example is not a backslash. It's a byte within a multibyte character, > which *entirely coincidentally* happens to have the same numeric value > as an ASCII backslash. But it isn't a backslash. And it won't be > processed as a backslash by any Big5-aware code. If I understand this explanation correctly, then you have given me the answer I needed! Please correct me if my understanding that follows is wrong again: PostgreSQL is Big5-aware code. It does not blindly unescape every backlash it encounters in SQL literals. Instead, it sees backslash as part of some Big5 characters as they are supposed to be when client encoding is set to Big5 (SET CLIENT_ENCODING TO BIG5). In other words, PostgreSQL sees the backslash in "C:\134" differently from that being part of multi-byte characters depending on which client encoding is used. Regards, CN -- http://www.fastmail.fm - And now for something completely different?
