we need to transfer data from Unix platform to z/OS. Data contains DECIMAL packed, INTEGER and text data (with umlauts). Direct connection from/to DB2 databases is not possible, so data have to be transported via simple file transfer.
Does anyone has an idea of how to manage this problem?
We’ve already checked-out some options:
FTP with SBDATACONN=(IBM-273,ISO8859-1): does not help, because DECIMAL data gets corrupted due to codepage translation.
FTP with BIN / IMAGE option: does not help, since text data gets corrupted due to codepage translation. This could be fixed by using a new DFSORT TRANSlate option, which allows translation TRAN=ATOE (ascci to ebcdic and vice versa) column-wise. But this requires additional work and deep knowledge of structures trasfered.
IBM provides a tool, called EDCICONV at z/OS and iconv at USS and AIX, but these tools translate record-wise, so corrupting DECIMAL/INTEGER/… data.
Another option/work-around is first to load into a DB2 for z/OS table and then externalize/UNLOAD/DNSTIAUL the data again. DB2 differentiates between internal and external data formats (text). So, it protects decimal/integer/date/time/timestamp/… data for being translated.
Ansicht von 2 Beiträgen - 1 bis 2 (von insgesamt 2)