mjhjmtu

I try to display two statistics of my nginx-server side-by-side. My first approach for this is column and process substitution.

column <(
 echo "sorted by busiest access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 sort -nk 1 |
 tail -25
) <(
 echo "recent access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 tail -25
)

Becomes:

sorted by busiest access times: recent access times:
80 2018-03-28 20:xx h 6 2018-03-29 23:xx h
81 2018-03-31 15:xx h 1 2018-03-30 01:xx h
82 2018-02-25 16:xx h 23 2018-03-30 08:xx h
83 2018-03-15 19:xx h 15 2018-03-30 09:xx h
84 2018-03-25 18:xx h 17 2018-03-30 11:xx h

Yikes, ugly. Where did the leading whitespaces go? OK, workaround with paste:

paste -d '#' <(
 echo "sorted by busiest access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 sort -nk 1 |
 tail -25
) <(
 echo "recent access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 tail -25
) | column -ts '#'

Better. However, the leading whitespaces are still missing:

sorted by busiest access times: recent access times:
80 2018-03-28 20:xx h 6 2018-03-29 23:xx h
81 2018-03-31 15:xx h 1 2018-03-30 01:xx h
82 2018-02-25 16:xx h 23 2018-03-30 08:xx h
83 2018-03-15 19:xx h 15 2018-03-30 09:xx h
84 2018-03-25 18:xx h 17 2018-03-30 11:xx h
...

Now I add a <(echo) as the first process substitution:

paste -d '#' <(echo) <(
 echo "sorted by busiest access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 sort -nk 1 |
 tail -25
) <(
 echo "recent access times:"
 egrep -v " /hs/|GET /favicon|GET / " $file |
 sed -e 's/^[^+[|:..:.. .*//g; s/:/ /; s///-/g' |
 date -f - +'%Y-%m-%d %H:xx h' |
 uniq -c |
 tail -25
) | column -ts '#'

The output becomes:

 sorted by busiest access times: recent access times:
 80 2018-03-28 20:xx h 6 2018-03-29 23:xx h
 81 2018-03-31 15:xx h 1 2018-03-30 01:xx h
 82 2018-02-25 16:xx h 23 2018-03-30 08:xx h
 83 2018-03-15 19:xx h 15 2018-03-30 09:xx h
 84 2018-03-25 18:xx h 17 2018-03-30 11:xx h
...

Even better, but now I have two spaces at the beginning of each line, which shouldn't be there. Well, I could remove those two with an additional | sed -e 's/^..//' after column, but I'm starting to think that I'm missing the actual problem here.

What's a better approach or am I actually doing it right? Or can I make column keep the leading whitespaces?

I'm using Manjaro if that's of significance.