align data by word into column
How can I align data into pretty columns relative to given word?
For example, I have output of the route -n command:
default via 172.20.99.254 dev eth0
87.33.17.71 dev tun0 scope link
89.223.15.12 via 172.20.99.254 dev eth0 src 172.20.99.74
172.20.9.0/24 dev eth0 proto kernel scope link src 172.20.99.74
65.46.5.89 dev tun0 scope link
192.168.11.0/24 dev tun0 scope link
45.211.111.7 dev tun0 scope link
and I would like to align it by the word dev, so that the column containing the word dev is aligned:
default via 172.20.99.254 dev eth0
87.33.17.71 dev tun0 scope link
89.223.15.12 via 172.20.99.254 dev eth0 src 172.20.99.74
172.20.9.0/24 dev eth0 proto kernel scope link src 172.20.99.74
65.46.5.89 dev tun0 scope link
192.168.11.0/24 dev tun0 scope link
45.211.111.7 dev tun0 scope link
I cannot just naively replace fist space character with a tab, because sometimes I need 1 tab, other times I need 3 tabs.
I would use a character that I am sure it doesn’t exist into file, as a dummy separator for the column command, like:
$ sed 's/dev/@dev/' file | column -ts@
default via 172.20.99.254 dev eth0
87.33.17.71 dev tun0 scope link
89.223.15.12 via 172.20.99.254 dev eth0 src 172.20.99.74
172.20.9.0/24 dev eth0 proto kernel scope link src 172.20.99.74
65.46.5.89 dev tun0 scope link
192.168.11.0/24 dev tun0 scope link
45.211.111.7 dev tun0 scope link
You can also use -o (specify the output separator) together with column, you can try with -o "" or -o " " to see behaviour.
Also the above sed substitutes only the first occurence, and a simple matching was used, maybe you ‘d need a more strict matching for other cases, like matching surrounding whitespaces.