Archives of the TeradataForum

Message Posted: Tue, 01 Jun 2010 @ 18:32:01 GMT

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

There are a number of ways to handle this problem.

The characters U+FFFD (x'EFBFBD' in UTF-8) and U+001A (x'1A' in UTF-8) are "substitution" characters. If these are the only characters that are giving you trouble, you can set DBS Control internal field 104 (AcceptReplacementCharacters) to TRUE. Then they will load.

If the problem is more extensive, you should probably call Tak Takahashi. He has worked extensively with Unicode and has some UDF's and other routines that can make this relatively easy for you.

If that doesn't work, you might consider an INMOD or a Perl routine to massage the record before going into FastLoad. By the same token, you could load the characters into a BYTE field and then write a UDF to walk the string and perform whatever conversion you want. Of course these require coding, but c'est la vie.

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