Could you copy the kanji only from the two sources to two text files with one kanji on each line, sort them, then do a line by line comparison?

It depends on your OS, but on Windows, for example, I believe you could do something like save them in Notepad in UTF8 text format, then sort each text file with sort.exe, then diff them textually via fc.exe (file compare) or via some free text comparison tool like BeyondCompare.

When doing things similar to this before, I've found that only minor issue to keep in mind is to remember to preserve the little UTF-8 file marker at the beginning of the text files when managing the output...

Learn SQL!

(I'd tell you how, but I've forgotten exactly how myself...)

If you have an Apple computer then here's a simple AppleScript to do that using the TextEdit word processor application.

The script will go through each character of a TextEdit document (named docOne.rtf) and see if it is present anywhere in another document (named docTwo.rtf). It will make a list of all those that are not in docTwo. (First set up the two documents then paste everything below this paragraph into the AppleScript Editor and click "Run". The result will appear in the bottom pane. If there are thousands of characters it may take a few minutes.)


tell application "TextEdit"
    set totalList to ""
    set text1 to text of document "docOne.rtf"
    set text2 to text of document "docTwo.rtf"
    set totalChar to count of characters of text1

    repeat with characX from 1 to totalChar
        set nexChar to character characX of text1
        if nexChar is not in text2 then set totalList to totalList & nexChar
    end repeat
    say "I have finished"
    totalList
end tell

On OS X you could also save the lists to plain text files and run this in Terminal:

comm -23 <(sort 1.txt) <(sort 2.txt)

Another option using Ruby:

ruby -e 'puts File.readlines("1.txt") - File.readlines("2.txt")'