[SOLVED] sed match domain help

x0000000009 · 04-17-2012, 09:16 AM

I'm trying to come up with a sed pattern match that works to get me the top domain for filtering in procmail.

if the domain is 123.mydomain.org then i want to extract mydomain.org like
sed -e 's/[a-zA-Z0-9]*\.//;s/*\.[a-z]// ;s/ //g'

but if it is just mydomain.org then i would like it to either return mydomain.org or nothing is also fine.

but if it is 123.456.mydomain.com for example it returns
mydomain.com

I am still grasping sed but i think its suppose to be something like
sed -n 's/[A-Za-z0-9]+\(.*\)+.[a-zA-Z]{2,4}/\1/p'
however that does not work.

grail · 04-17-2012, 10:39 AM

Have you thought of using awk and setting the delimiter to a dot? Or are we stuck with using sed?

colucix · 04-17-2012, 10:46 AM

Code:

$ echo 123.456.mydomain.com | sed -r 's/.*\.([^.]+\.[^.]+)$/\1/'
mydomain.com
$ echo mydomain.com | sed -r 's/.*\.([^.]+\.[^.]+)$/\1/'
mydomain.com

or as grail suggested, using awk:

Code:

$ echo 123.456.pippo.mydomain.com | awk -F. '{print $(NF-1) FS $NF}'
mydomain.com

Hope this helps.