Fastload, Multiload and TPump all allow you to specify the character set of the input file using the -c option.
TPT aso allows you to specify input chracterset.
If your input data is all character format, you can also use your FTP package on the mainframe to do the convert - you dont need a separate mainframe step.
Note if your output has mixed datatypes (eg characters, decimal and integer/ real/ smallint numbers in non-character form) FTP cannot convert - use the EBCDIC characterset.
So drop it down and load away!
According to the MLOAD manual EBCDIC charset not supported unless you have a channel
Table 13: Character Sets Supported by MultiLoad
Character Set Name Description Configuration
ASCII Latin Network-attached
EBCDIC Latin Channel-attached
HANGULEBCDIC933_1II Korean Channel-attached
HANGULKSC5601_2R4 Korean Network-attached
SCHEBCDIC935_2IJ Simplified Chinese Channel-attached
SCHGB2312_1T0 Simplified Chinese Network-attached
TCHBIG5_1R0 Traditional Chinese Network-attached
TCHEBCDIC937_31B Traditional Chinese Channel-attached
KATAKANAEBCDIC Japanese Channel-attached
KANJIEBCDIC5026_01 Japanese Channel-attached
KANJIEBCDIC5035_01 Japanese Channel-attached
KANJIEUC_0U Japanese Network-attached
KANJISJIS_0S Japanese Network-attached
UTF8 Unicode Channel-attached
Network-attached
UTF16 Unicode Network-attached
Lets say you have a Mainframe EBCDIC file, however you have NO Channel connection. You have a remote located Teradata machine with a LINUX ETL
server.
I say in order to load this file you must first convert it while on the Mainframe to
an ASCII/Text file, FTP the file to the LINUX ETL server and then load from there.
I don't think you can just drop this File on a server and load away !!!!
Is there anything I'm over looking here ? what about TPT ?
Any other idea's ???