Question

One part of my problem was solved with this answer:
Threadlink
, but an important part of my problem was unsolved!

After using

diff a.csv b.csv | grep -E -A1 '^[0-9]+d[0-9]+$' | grep -v '^--$' | sed -n '0~2 p' | sed -re 's,^< (.*)$,\1,g'

several times i found something left.
Sometimes multible following lines are deleted.
If only one line was deleted there are something like this found:

3663d3661

For multible lines it is:

3724,3725d3718

So i changed the diff call to:

diff a.csv b.csv | grep -E -A1 '^[0-9]+\,*[0-9]*d[0-9]+$' | grep -v '^--$' | sed -n '0~2 p' | sed -re 's,^< (.*)$,\1,g'

This works for the first of multiple deleted lines.

My question is:
How could i get all deletet lines (maybe 5 following lines) in such a case?
What did i have to change in the diff call?

Answer 1

diff a.csv b.csv | sed -n '/^[0-9]\+d[0-9]*/,/^[0-9]\+[^d]*$/{/^[0-9]\+/d;s/^< //;p}'

will do that.

/^[0-9]\+d[0-9]*/,/^[0-9]\+[^d]*$/

will find the string range which are deleted

/^[0-9]\+/d

will delete all 6842d6844,6772

s/^< //

will replace all < in the beginning of lines

and p will print the line.