1 of 1 people found this helpful
The actual issue handling these multibytes characters with fix-width is that at certain point a decision about how to handle the extra bytes has to be made.
When a program writes or reads, an extra byte has to determine either to truncate the next character, or truncate those at the end of the record, given the fact that the total length is fixed. The issues arise when the program that writes the record and the program that reads that record treat the extra characters in a different way. In worst case scenarios where a record contains only special characters (all having extra bytes), say that a normal character uses 1 byte and a special character uses 2 bytes, half of the record would be truncated.
Extract of documentation PowerCenter > PowerCenter Designer Guide > Working with Flat Files > Working with Multibyte Data in Fixed-Width Targets:
To load multibyte data into a fixed-width flat file target, configure the precision to accommodate the multibyte data. Fixed-width files are byte-oriented, not character-oriented. When you configure the precision for a fixed-width target, you need to consider the number of bytes you load into the target, rather than the number of characters. The Integration Service writes the row to the reject file if the precision is not large enough to accommodate the multibyte data. When the Integration Service writes the row to the reject file, it writes a message in the session log.
Note: Delimited files are character-oriented, and you do not need to allow for additional precision for multibyte data.