Help with character sort

[grazinggoat]

I have 30 various channele split into groups

serv1g1s1
dbserv100g5s1
serv1g4s2
dbserv20g15s6

I want to be able to break the list down by the groups (ie g1s1, g4s2, g5s1, g15s6 )
how can i sort when there can be any number of characters before the g, inbetween the s?

I only want the list to chop down to show just the groups

g1s1
g4s2
g15s6

I have no idea on how to do this? I looked up sort an am not sure what could be done using sed or awk Any thoughts please?

 

[feherke]

Hi

Off-topic solution, but my time is limited now. If only the "g*s*" parts are needed, I would do this way :

grep -o 'g[0-9]\+s[0-9]\+$' /input/file | sort -k 1.2n

Feherke.
feherke.ga

 

[dioid]

You could use grep, sed, or awk to achieve this...

Sed:

sed 's/.*\(g.*\)$/\1/g' gg.data | sort -V

Awk:

awk 'match($0,/g.*$/) {print substr($0,RSTART,RLENGTH)}' gg.data | sort -V

output:
g1s1
g4s2
g5s1
g15s6

thread80-1459823