Joined: 23 Nov 2006 Posts: 19270 Location: Inside the Matrix
Hello Hari and welcome to the forums,
Do some one has a resolution for this.
When data is ftp'd from the mainframe to the unix system, it is converted from ebcdic to ascii. Your "latin" characters have a hex value that translates to the unwanted ascii character.
Do you know which hex value on the unix system represents the "Latin N"? Are there many?
You might talk with your network support people to see if there is a code page that provides the conversion you need. Otherwise, you may want to convert the Latin N (etc) to standard characters before downloading.
You could also transmit the file in binary and handle the conversion on the unix system.
Joined: 06 Jun 2008 Posts: 8217 Location: Dubuque, Iowa, USA
CR/LF are PC concepts that are not applicable to mainframe. Unix, by the way, uses LF but not the CR. Probably the CR and LF aren't getting transferred because they don't exist on the mainframe file.
If your file is fixed record size on the mainframe, it's a fairly trivial process on a PC to chop the file back into records. If the file is variable record size on the mainframe, binary transfers are problematic due to the different way record sizes are determined between platforms.
Suggestion: find the Communications Server User's Guide manual and read about the SBDATACONN option of FTP. This allows you to change the code page while FTP is transferring the data.