mjhjmtu

I have a file (*.ses) which contains following line

$ rea ses '../../../../abcdefgh/abcd_efgh-A20_ABC-abcdefgh-Abcdefgh_Abcdef_123er_vb001.ses'

When I use this command:

cat a4.ses | grep -im1 'rea ses' | awk -F'[/]' 'print $NF'

the output is:

abcd_efgh-A20_ABC-abcdefgh-Abcdefgh_Abcdef_123er_vb001.ses'

I just want to output to be:

abcd_efgh-A20_ABC-abcdefgh-Abcdefgh_Abcdef_123er_vb001

without the extension.

How can I do that?

If .ses' is a static extension, simply hard-code the removal of those trailing 5 characters into awk by printing the string from the beginning until 5 from the end:

awk -F/ 'print substr($NF, 1, length($NF)-5)'

If the length of the extension can vary, then replace it with the empty string before printing it:

awk -F/ 'gsub(/..+$/, "", $NF); print $NF'

If your grep supports perl compatible regular expression (PCRE) syntax:

$ grep -Po 'rea ses.*/K[^.]*' file
abcd_efgh-A20_ABC-abcdefgh-Abcdefgh_Abcdef_123er_vb001

Explanation:

• match rea ses and then greedily everything up to / inclusive; then


• match the longest sequence of non-period characters


• discard (K) the left portion and output only what remains of the match (-o)

You can dispense with that pipeline and use sed instead

sed -n '/rea ses/s!^.*/(.*).[^.]*$!1!p' a4.ses

Output

abcd_efgh-A20_ABC-abcdefgh-Abcdefgh_Abcdef_123er_vb001

What that sed command does can be described as follows

1. 
   -n do not print anything unless a match is made


2. 
   /rea ses/ only consider lines that match this RE


3. 
   s!...!...!p substitute the RE within the first two exclamation marks (!) for the following string, but only print the line if a match occurs


4. 
   

   The RE ^.*/(.*).[^.]*$ matches

   
   
   
   • Everything up to the last slash /
     
   
   
   • Everything from there to the last dot . (remembered as pattern 1)
   
   
   • Everything else
   
   



5. The substitution of the pattern described in #4 is made with pattern 1, i.e. your filename without the trailing dotted extension

You can use basename to remove a trailing extension:

cat a4.ses | grep -im1 'rea ses' | awk -F'[/]' 'print $NF' | xargs basename -s .ses' 

(submitted for completeness, given your process, @steeldriver's answer is better)