OK. So that is basically the opposite of what I understood in the first
place: the -D option of 'uniq' must be used instead of its -u option:
$ awk 'NR > 2 { print $2, $1 }' selection.txt | uniq -Df 1 | awk '{ print $2
"\t" $1 }'
alnikino.ru 212.109.192.55
alnikino.ru 212.109.197.49
alolika.su 82.202.160.112
alolika.su 92.63.110.215
Using the same command on selection-longer.txt and redirecting the output to
a file named out, I can then compare with your manual work (after removing
the headers and "normalizing" the number of tabs to one). I find three
differences:
$ tail -n +3 selection-longer-dups.txt | tr -s '\t' | diff - out
11a12
> a214947.fvds.ru (149.154.69.180)
76a78,79
> a920222.fvds.ru (188.120.251.198)
> a920222.fvds.ru (188.120.251.50)
So, well, either I still have not understood the task or you forgot those
three lines. In any case, you should never mechanically process by hand long
files.
