Ruby libxml parser that I use to process large xml files in SAX mode refused to process a file that looked perfectly valid, throwing ‘Extra content at the end of the document’ error somewhere in the middle of the file. It turned out that it disliked control character \x0B (vertical tab), which is not allowed in XML according to the spec.
To simply remove the vertical tabs from the file (or, rather, replace them with spaces), I tried using sed like this
sed s/\x0B/\ /g file.xml
but I found out that \xXX syntax is not supported by OSX sed version, which is a shame, so I used a ruby script, which, to my surprise, was quick enough to process a 800 MB file.
output = File.open("out.xml", 'w+')
File.open('file.xml').each{|p| output.puts p.gsub(/\x0B/, ' ')}
A Big Thanks ! I was stumped as to why it was terminating with that error in the _middle_ of the file !
How did you discover the rogue character? I tried gsubbing away that character in my xml but it doesn’t appear to be the culprit
I was processing an xml file in sax mode, so I knew which record threw an exception. Then I looked in hex mode at the offending record, looking for anything below 0x1F. I also found out that things like  also raise an exception inside my parser.