0

We have used the below process in loading the file(with comp3 data) to informatica from Mainframe.

  1. FTPd the mainframe file in Binary mode to Unix server
  2. Imported the source cobol copybook as VSAM file type
  3. Set the source file properties as fixed width and codepage as IBM EBCDIC US English IBM1047.
  4. Verified the USAGE in source as COMP3 for the relevent fields

Even with above steps we could not map the comp3 fields correct. When code page was changed to ASCII we could only map string data.

Could you please let us know what we could be missing?

user2677856
  • 1
  • 1

2 Answers2

0
When code page was changed to ASCII we could only map string data.

That's because character encoding only applies to string data. Bytes of data such as 'packed' bytes or binary data bytes have no meaning after being converted to a different character encoding unless first converted to character representations.

What platform is running the COBOL? In COBOL, the meanings of the various "COMP" data types change depending on hardware platform and potentially even OS. (I suppose even by specific compiler though I don't know of examples.) Without knowing how your specific COBOL maps data types to memory bit patterns, there's no good way to guess how "mainframe 'packed'" data should properly be handled.

user2338816
  • 2,163
  • 11
  • 11
0

comp-3 compress 2 numbers in 1 character, you should map it to a numeric field, then you can put it in your string data . or use PIC Z data

user2453063
  • 1
  • 1