Converting EBCDIC to ASCII file format- SerDe for hive. How to overcome issue of delimiter in EBCDIC file

Question

There is nothing like /n in ebcdic. There is no support for new line in ebcdic. How should I convert that like. There is no delimiter in ebcdic. So while converting this file. How shall I know that new line has come? Suggestions please.

Answer 1

Actually there is a new-line character (x'15'). Normal ZOS files do not use the New-Line character. ZOS is built around Fixed-Width, VB, VSAM etc files.

Options include:

1. If it is a Text file (unlikely) convert the file to ascii when it is transferred off the mainframe.
2. Convert the File to Text on the Mainframe and convert when transferring off the mainframe.
3. Use a commercial package. Sync-Sort has DMX-h, there is also Datameer
4. If you have Cobol copybook, look at these open source packages:
   • https://wiki.cask.co/display/CE/Plugin+for+COBOL+Copybook+Reader+-+Fixed+Length
   • https://index.pocketcluster.io/tmalaska-copybookinputformat.html
   • https://github.com/ianbuss/CopybookHadoop
   • https://sourceforge.net/projects/coboltocsv/
5. JRecord could be used with a Cobol Copybook, Plain Java Code or a Xml file description 6.