Oracle Database Discussions

Hi All,

I am facing a character set issue.
This is 11.2 on Redhat 5.

I receive pipe delimited data files from other system. I don't know what OS and/or character set the sending system is on.

Sometimes some strange characters come in with the data (most of the file is plain English/Roman characters) e.g. en España|BAN

Now, when this is loaded with sql*loader, somehow the pipe is not recognised. It is like the characters before pipe 'eat up' that pipe character. Effectively, a line with 10 pipe delimited fields will be treated as line with 9 fields.

I wonder how I can address this?
can this be dealt with by setting NLS_LANG?
or it is something at OS level?
How the OS character set and Oracle character set tie up?
How can I find out the various character sets (i.e. OS character set, file character set, Oracle character set) involved?

If anybody has any such previous experience, please share.

Thanks in advance.