Archives of the TeradataForum

Message Posted: Thu, 21 Sep 2000 @ 21:53:51 GMT

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

Glen, thank you for your response.

In the example that you use, the client identifies the host data type changes for Teradata. The issue is that the data will be coming to us in Unicode, and not in MBCS or DBCS. The data will have already been rectified to Unicode Japanese, Korean, Chinese, Lao, Tamil, English, Spanish, and any other language that is on the 'net. The data is already mixed, and in many columns.

I actually tried to move the data, by creating a table on SQL server, with three columns: an integer, a latin char, and a unicode varchar. I inserted data mixing latin and unicode in the unicode field by linking a table in msaccess. I verifed that the Unicode data was stored in SQLServer by creating a new link and selecting from the table.

I then created a package in DTS that copied the data from SQLServer to the V2R4 system, and found the resulting data is not the same.... I tried it both with and without Unicode as the translation. With the translation, the connect fails because the package makes a call to the dictionary that does not return the data in - Unicode...

     CREATE SET TABLE NCNTBASET.unicode_test ,NO FALLBACK ,
          NO BEFORE JOURNAL,
          NO AFTER JOURNAL
          (
           i1 INTEGER,
           SBCS CHAR(10) CHARACTER SET LATIN NOT CASESPECIFIC,
           Unicode VARCHAR(30) CHARACTER SET UNICODE NOT CASESPECIFIC)
     PRIMARY INDEX ( i1 );

     select i1, char2hexint(unicode) from unicode_test;

      *** Query completed. 8 rows found. 2 columns returned.
      *** Total elapsed time was 1 second.

              i1 Char2HexInt(Unicode)
          ------ --------------------------------------------
               2 003F
               5 003F003F003F003F003F003F003F003F
               1 0061
              10 003F003F003F003F003F003F003F003F00200020

Note that 00F3 is "?"...

I would appreciate any help with engineering that I can get. I presume that I am doing something wrong, but the manual clearly states that Unicode is not supported by the driver. See page 4-22 of the ODBC manual for 2.08.

Thanks -

Michael McIntire

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