使用 shell 脚本遍历 $PATH 变量

Question

If I type echo $PATH in the terminal I get the following result:

/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/home/fnorbert/.local/bin:/home/fnorbert/bin

I want to iterate over these paths using a shell script, but I do not know how to do that.
I tried the following:

for i in 1 2 3  
do
    echo $PATH | cut -d':' -f$i
done

This prints the first three paths, but I want to represent every path with the variable i if it is possible.

Answer 1

You can use read with delimiter set as :

while read -d ':' p; do
   echo "$p"
done <<< "$PATH:"

Answer 2

Simplest way is probably to change IFS to a colon and let the word splitting do it:

IFS=:
for p in $PATH ; do 
    echo "$p"
done

But that might trigger file name globbing, in the weird case that your PATH contained characters like *?[] . You'd need to use set -f to avoid that.

Though changing IFS might be considered problematic anyway, since it affects the rest of the script. So in Bash, we could just split the paths to an array with read -a , this doesn't have a problem with glob characters either:

IFS=: read -a paths <<< "$PATH"
for p in "${paths[@]}" ; do
    echo "$p"
done

Answer 3

with echo:

echo "${PATH//:/$'\n'}"

sed:

sed 's/:/\n/g' <<< "$PATH"

tr:

tr ':' '\n' <<< "$PATH"

python:

python -c "import os; print os.environ['PATH'].replace(':', '\n')"

for iterate use for:

for i in ${PATH//:/ }; do echo $i; done

Answer 4

Here is a trivial solution, that extends your attempt a little

for i in $( echo "$PATH" | cut -d: -f 1- --output-delimiter=" " ) ; 
do 
   echo "$i" ; 
done

One trick that is used here is -f 1- to specify all fields instead of just one.

Another trick is to use the --output-delimiter option.

This solution is suffers from sensitivity to special characters in directory names.

Consider the following example

PATH="help me":\*:now
for i in $( echo "$PATH" | cut -d: -f 1- --output-delimiter=" " ) ; 
do 
   echo "$i" ; 
done

This would output

help
me
foo
now

That is,

• spaces in directory names will not be treated correctly
• special characters, such as * will be expanded by the shell into a list of files in the current directory ( foo is the name of a file residing in the current directory)

But if your PATH does not contain anything special, this would work. Otherwise or rather in all cases take the solution that uses read .

Answer 5

You can use awk for the requirement,

echo $PATH | awk -F: '{ for (i=1; i<=NF; i++) {print $i}}'

This is well efficient in performance.