Whenever a client application talks to the Teradata Database, it indicates its character set (form-of-use for character data) and the Teradata Database returns all character types to the application in that form.
Note that in application programs (within ANSI C programs, for example), Kanji strings can be represented as character strings with mixed single-byte character/multibyte character form-of-use or as strings with fixed width form-of-use.
Suppose that table_1 has three columns named cunicode, csjis, and clatin with the respective corresponding repertoires UNICODE, KANJISJIS, and LATIN. Suppose further that the current client character set is KanjiEUC.
If you submit the following statement, then the data from all columns will arrive into the application in the KanjiEUC form-of-use.
SELECT cunicode, csjis, clatin FROM table_1;
Because of this, the application should use variables declared as mixed single-byte character/multibyte character strings in order to store the data from the Teradata Database.
Consider a BTEQ session with the ASCII client character set.
The data from the cunicode, csjis, and clatin columns are returned to the BTEQ session in ASCII, as illustrated by the following figure.
Note that your application should declare its variables as mixed single-byte strings to store the data from the Teradata Database.
Because the ASCII repertoire is much smaller than the repertoires of cunicode or csjis columns, some characters might be converted to error characters on input, some might be converted to something else on export, or the transaction might abort.