I want to diff two sets of mod_rewrite rules. The set of lines are about 90% identical, but the order is so different that diff basically says they are completely different.

How can I see which lines are truly different between two files, regardless of their line number?

  • 3
    Pass them both through sort fist. – Shawn J. Goff Oct 25 '11 at 21:45
  • @Shawn Can I do this without creating (and deleting afterwards) two one-off files? – user394 Oct 25 '11 at 21:52

sort can be used to get the files into the same order so diff can compare them and identify the differences. If you have process substitution, you can use that and avoid creating new sorted files.

diff <(sort file1) <(sort file2)

Made a script for this which keeps the line sequence intact. Here's an annotated version of the important lines:

# Strip all context lines
diff_lines="$(grep '^[><+-] ' | sed 's/^+/>/;s/^-/</')" || exit 0

# For each line, count the number of lines with the same content in the
# "left" and "right" diffs. If the numbers are not the same, then the line
# was either not moved or it's not obvious where it was moved, so the line
# is printed.
while IFS= read -r line
    count_removes="$(grep -cFxe "< $contents" <<< "$diff_lines" || true)"
    count_adds="$(grep -cFxe "> $contents" <<< "$diff_lines" || true)"
    if [[ "$count_removes" -eq "$count_adds" ]]
        # Line has been moved; skip it.

    echo "$line"
done <<< "$diff_lines"

if [ "${line+defined}" = defined ]
    printf "$line"

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.