Originally posted by TimHall Not necessarily. On one of our old 7.3.4 UNIX boxes we have Japanese characters stored in VARCHAR2 datatypes using the US7ASCII characterset.
I wonder how this can be!? AFAIK US7ASCII is 7-byte characterset, so it properly stores only ASCII codes from 0 to 127. You can't even store West European special characters in there, because MSB bit is trimmed off, so I wonder how were you able to stuff Japanese characters in there (or beter to say, how did you manage to retrieve them properly afterwards)?
Jurij Modic ASCII a stupid question, get a stupid ANSI
24 hours in a day .... 24 beer in a case .... coincidence?
To be honest I'm not totally sure. I wasn't here when it got set up.
It works through a third party bit of software. When a Japanese character is typed the software notices, converts it to two individual character codes that represent the multibyte character. These are then stored in the database.
When a select is done you just get gibberish (non-printable characters) unless you are running this software. When the software is running it intercepts certain character code combinations and translates them to a multibyte character set and sends that data to windows. This means that SQL*Plus shows Japanese characters on screen.
It works pretty well. At the time the Oracle unicode implementation was so dodgy that Oracle recommended waiting until Oracle 8 was released, which was not an option at the time.
The software is called NJWIN. Not sure what version they use here as it's a client tool and I'm not involved in that project much.