In shell, print only the specific line/word from command's output

Question

This is the output snippet of cloc command

C/C++ Header
 same                            0              0          42812          19019
 modified                       33              0            128            434
 added                           3             63            367            207
 removed                         0            105            265            479
Perl
 same                            0              0              0              0
 modified                        0              0              0              0
 added                           2              0              0              0
 removed                         0              0              0              0
-------------------------------------------------------------------------------
SUM:
  same                            0              0          54150          81007
  modified                       82              0            252           2137
  added                           7            184            802           1281
  removed                         0            234            734           1499
-------------------------------------------------------------------------------

Using Shell command, how to print only 82, 2137 (in row modified) and 7, 1281(in row added) ? I need to print only these 4 numbers under "SUM:" and ignore the rest.

Answer 1

If you want to get the number of added/modified files/lines between two versions of a project, you can do it correctly and robustly with cloc's XML output:

 cloc --xml --out=report.xml --diff  bash-4.2.tar.gz bash-4.3.tar.gz 

 xmlstarlet < report.xml sel -t \
    -m /diff_results/modified/total -v @sum_files -o ' ' -v @code -nl \
    -m /diff_results/added/total    -v @sum_files -o ' ' -v @code -nl 

xmlstarlet can output the numbers in any format you want. This specific example uses this format (modified files/lines followed by added files/lines):

15 11157
203 6058

The benefit of processing computer readable output rather than human readable output is that it won't break if cloc slightly changes its output formatting in the future.