CODEWITHSUNDEEP
linuxstringbashecho
user3473750
user3473750

Reputation: 129

Basename like string manipulation when using pipes

Hello and thank you in advance for your answers.

When I use the following code, I get output such as:

  /home/pony/IOSO/test/A
  /home/pony/IOSO/test/B


FILE=last.cfg
DIR=$(realpath "$2") #or something else    
grep $DIR $FILE | awk '{ print $2 }' | sort | uniq # | basename does not work

How do I make my pipe work with basename or something like this:

$ s=/the/path/foo.txt
$ echo ${s##*/}
foo.txt

Upvotes: 2

Views: 1021

Answers (2)

nu11p01n73R
nu11p01n73R

Reputation: 26667

You can do this using awk so that you can avoid the unecessary usage of grep awk and sed sequance

The correct method can be

awk -v dir="$DIR" -F"/" '$0 ~ dir{print $NF}'

Test

$ DIR="/home/pony/IOSO/test"

$ cat input
/home/pony/IOSO/test/A
/home/pony/IOSO/test/B
/home/test/test

$ awk -v dir="$DIR" -F"/" '$0 ~ dir{print $NF}' input
A
B

What it does

  • -v dir="$DIR" creates an awk variable with the value of shell variable DIR

  • -F"/" set the field delimiter as the /. This is done as in input file the directories are separated by / and we need to get the last field, filename.

  • '$0 ~ dir checks for each input line if it matches the pattern in variable dir

  • print $NF print the last field, here the the filename. NF stands for number of fields.

Upvotes: 1

Avinash Raj
Avinash Raj

Reputation: 174796

Just pass the uniq command's output to a sed command like below to remove all the characters upto the last / symbol.

grep $DIR $FILE | awk '{ print $2 }' | sort | uniq | sed 's~.*/~~'

Upvotes: 1

Related Questions