Problem with binary data type in CFF stage

[hobocamp]

Hello -

This one has me perplexed so I'm hoping someone might have a suggestion.

I have a number of ebcidic files I'm reading via CFF stage, and outputting to flat files. Going pretty smoothly, except I've hit a snag with one of the files whose definition is thus:

05 SPARMSGU-RECORD.
10 PART-NB PIC X(4).
10 CLIENT-NBR PIC X(4).
10 PART-CLNT-GRNLT-CD PIC X(3).
10 USER-KEY-TXT PIC X(10).
10 SPAD-SCRTY-CD PIC X(4).
10 APPL-TYPE-CD PIC X(4).
10 MSSG-EXPTN-DT PIC X(10).
10 MSSG-LINE-NBR PIC S9(2) COMP.
10 ALPHA-KEY-TXT PIC X(20).

The problem occurs in the PIC S9 record (highlighted in red).

For the field in question, the raw data values are '00 01', '00 02', '00 03', etc. In the output, numbers 1, 2 and 3 display correctly, as 1, 2, 3, etc. But 00 04 always displays 63, and 00 05 displays as 26.

Any ideas as to why this might be occurring?

Here are the other details for my job set-up.
Record type: Fixed
Byte order: Native-endian
Data format: Binary
Character Set: EBCDIC

I've tried the setting variations that I can think of, including changing the Byte-order selections, so far to no avail.
Thanks in advance for any suggestions.

Thomas Smith

[hobocamp]

Solved the issue. I was splitting the file into two parts before processing (via a job that read the file as a whole string of data into two separate cff targets). Somehow in the splitting the binary data was getting mangled.

Marked as resolved.