Archives of the TeradataForum
Message Posted: Thu, 21 Sep 2000 @ 20:11:51 GMT
First the disclaimer. I may be speaking out of turn, because I have no experience with V2R4 MBCS nor ODBC 2.08 MBCS.
However, I have had considerable experience with V2R2, V2R3 and MBCS data being transfered via a variety of ODBC revisions (Most recently around the 2.02 vintage).
Our experience has been with both Korean and Japanese characters. We have also succesfully tested english character data (i.e. SBCS) being stored into tables with character columns defined as UNICODE.
In all cases, the data was being transfered to / from Teradata via Visual Basic (v 5 and v 6) applications. We also used Queryman and MS-Access to view the results of our testing. Further, the MBCS character sets were generated by the VB applications whilst they were running on the language edition of Windows NT i.e. we generated Korean characters by running our app on a Korean Windoes NT 4 SP-3 system, similarly Japanese was generated from a Japanese edition of Windows NT 4 SP3 and so on.
I guess I am answering your question 2 when I say that all of this simply works. In the case of the Japanese platform, we actually used "Shift JIS" as the character set used by the ODBC driver - even though Teradata was storing data in UNICODE and Windows-NT uses Unicode. Again, it simply works. I don't know what character set the English language edition of Windows NT uses - some people tell me ASCII (which was supported by some tests), some tell me UNICODE (which was also supported by some tests). The ODBC driver (at least v2.02 vintage) certainly doesn't allow you to specify the character set (eg Shift-JIS, UNICODE etc) when running on an English language edition of Windows NT and from all the traces we were performing looked like ASCII data was being transfered into Teradata.
My understanding is that Teradata recognises the "host character set" and will internally translate into UNICODE (in V2R3 and later) for storage. This is how Teradata can support ASCII, UNICODE, Shift-JIS, EUC, EBCDIC and other host operating systems' character sets.
Again in relation to your second question, do you have a specific problem that you have encountered?
hope this is of assistance
|Copyright 2016 - All Rights Reserved
|Last Modified: 15 Jun 2023