Archives of the TeradataForum

Message Posted: Sat, 07 Jul 2007 @ 11:22:46 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Thanks guys (Dave Wellman,Dieter Noeth,Victor Sokovin, and others if they arrive after this note).

Yes, I define all my character columns up to 30 bytes as CHAR and anything over that as VARCHAR (I have my reasons).

I didn't realize the server side was UTF-16. Actually, I'm a little annoyed with myself, because I've been meaning to check the space difference between a LATIN and UNICODE table for months, but just didn't get around to it.

Your combined comments - apart from alarming me - are going to make me stop and think what I'm trying to achieve.

One immediate thought is that I could change my approach to the following :

Only define columns as UNICODE where it is known to contain double byte characters, rather than just set all character based columns to Unicode. I can't see this as being easy to guarantee but I can no longer ignore the approach.

Set all Unicode character columns to VARCHAR, which I assume (I'll check it this time!) will only take up the space it needs. crawl back under a stone and hope it all goes away !

I think I'll also have a word with my NCR consultants and ask how they plan to move forward with Unicode on the Server side.

As usual, have a virtual beer on me (and if I run in to you at a conference, I'll materialize them !)

Dave Clough
Database Designer
Express ICS

www.tnt.com

	<Prev	Next>		<<First	<Prev	Next>	Last>>