EBCDIC format not coming right

[shartyag]

Hi Guys,
I am working on a problem for last 1 week and even after trying almost everything its not getting solved, now my last hope is DSXchange.
Here it goes.

Requirement:
I need two o/p files in the end of process (it involves around 20 jobs) , one in text format (ascii) and other one in EBCDIC format which will be ftped to mainframes.

Process:
I have a process in which at some point say job no. 13 and 15 is creating 5 complex flat files (intermediate files) .
All these CFF append the data one by one to two files
1) to a seq ebcdic file with NLS IBM500
2) to a seq text file with NLS 8859
After all the CFFs are appended to both files (ascii and ebsdic) these two go to the final job where data gets sorted and final files get created.

Problem:
wherever there is a field as "NATIVE BINARY" in CFF stage I am setting it to decimal with the same length in the transformer after CFF but its data is not coming in the right format.
I have tried CHAR also with the same length but it also not worked.
Also for these fields I am getting a warning in the director as follows:
When checking operator: When binding input interface field "Filler" to field "Filler": Implicit conversion from source type "int16" to result type "ustring[3]": Converting number to ustring."

I am really bugged up with this problem, Please suggest if anyone has any idea ASAP.

[battaliou]

Got a possible solution to CFF COBOL write problem.

At the CFF level stage, Record Options tab, Set the Byte Order as Big-endian...