 |
 |
Archives of the TeradataForum
Message Posted: Fri, 30 Jan 2004 @ 16:47:45 GMT
| Subj: | | Re: mload with utf8 charset.. |
| |
| From: | | Victor Sokovin |
Vivek,
| | Since the characters are of variable bytes in a specific field Mload does not know when to take single byte or double byte or triple byte.
So the solution could be to push all the chars in one temp table & then chop them from there using the SUBSTR function. | |
I think you are an easy customer. ML accepts UTF8, and UTF8 is specifically designed to handle multibyte encodings. So, ML should be able
to support them as well. OK, kanji might be extra difficult but the variable length of characters in UTF8 is a common feature. It should work when
you mix together the simplest two-byte characters and usual Latin-based characters. The latter use the one-byte encoding.
Just for the reference, what is the version(s) of ML that you tried out?
I agree that for ML loads the database should not be kanji enabled. You need to enable kanji if you plan to use kanji characters in metadata
(table and column names, compression values etc.)
Regards,
Victor
| |
