Shell Programming and Scripting

export only last two fields?

here is my test file, with variable length of fields, each line is composed of letter only, and each line has at least 2 fields, fields were separated by dot (.)

abcd.abc.mlm
dft.dfasdf.bmk.kdjlka
ksdfalksdjfsl.tyu.ml
kasdjf.asdfkja.asdfj.asdf.df.m
lja.ml
aslkjdfjasldkf.sadlfkasdjlf.sdfasdfsa.asdfasdf.fdlm

I am trying to use sed to parse the test file, and only display the last two fields, like

abc.mlm
bmk.kdjlka
tyu.ml
df.m

At first my idea was something like ++ sed 's/!\.[a-z][a-z]*\.[a-z][a-z]*$/&/' test ++, and sure it did not work.

Any help? thank you!

fields.txt:

Code:

abcd.abc.mlm                                       
dft.dfasdf.bmk.kdjlka                              
ksdfalksdjfsl.tyu.ml                               
kasdjf.asdfkja.asdfj.asdf.df.m                     
lja.ml                                             
aslkjdfjasldkf.sadlfkasdjlf.sdfasdfsa.asdfasdf.fdlm

Code:

awk 'BEGIN { FS="."; OFS="." } { print $(NF-1),$NF }' fields.txt

Code:

abc.mlm      
bmk.kdjlka   
tyu.ml       
df.m         
lja.ml       
asdfasdf.fdlm

Really cool and thanks Glenn Arndt!


But still I have great interest in whether sed can do this

I think "pattern keeping method" should work, as following

sed 's/^.*\(\.[a-z][a-z]*\.[a-z][a-z]*$\)/\1/'

but apparently it does not. I am reading more sed help file trying to figure out why.

Code:

echo 'dft.dfasdf.bmk.kdjlka' | sed 's/.*[.]\(.*\)[.]\(.*\)/\1.\2/'

Thank you vgersh99!