13

Possible Duplicate:
keep duplicates out of $PATH on source

Some of my scripts are located in $HOME/mydir. Unfortunately, they are not accessible from elsewhere if this dir is not part of PATH.

I would like to create a small scripts checking whether $HOME/mydir is part of PATH. If it is not, PATH should be updated with it and EXPORTED.

My question is: which command can I use to check whether $HOME/mydir is part of PATH? Thanks.

SOLUTION

For the records, I implemented the following:

echo "Before: $PATH"

echo $PATH | grep -q "$HOME/scripts"

if [ $? -eq 0 ]; then
  PATH=$PATH:$HOME/scripts
  export PATH
else
  export PATH
fi

echo ""
echo "After: $PATH"
Rui F Ribeiro
  • 55,929
  • 26
  • 146
  • 227
Jérôme Verstrynge
  • 1,699
  • 5
  • 19
  • 20
  • 1
    are you sure this works? I made a typo in my response, which I have now edited. The exit code is in the $? variable, not $0, which should be the name of your script. Change `$0 == 0` to `$? -eq 0`. You should try it with a some path that is not in your PATH – Rado Feb 18 '12 at 18:33
  • It worked on my limited tests, but I have updated my code. – Jérôme Verstrynge Feb 19 '12 at 09:43

2 Answers2

26

I would write

case ":$PATH:" in
  *:$HOME/mydir:*) echo it is in the path;;
  *) echo not there ;;
esac

I write ":$PATH:" to ensure that the pattern matches if the desired path is either first or last in your $PATH.

glenn jackman
  • 84,176
  • 15
  • 116
  • 168
9

You can write a simple bash script to do something like this. 

echo $PATH | grep -q  "/your/search/path"

Then check if $? is not 0, meaning no match, and if so add the path

export PATH=$PATH:/your/search/path
Rado
  • 256
  • 2
  • 5
  • $0 ?? is it right way to do? – Nikhil Mulley Feb 18 '12 at 17:47
  • @NikhilMulley, no it's not ... it's a typo, should be `$?`. – Rado Feb 18 '12 at 18:29
  • [glennjackman's answer](http://unix.stackexchange.com/a/32054/9537) is the more correct way to do this. – jw013 Feb 18 '12 at 18:54
  • 3
    @jw013: there is no correct, on incorrect way to do it, there is ways that work and ways that do not, and I think the grep way works. Maybe you meant that using grep in this case waste a process, but I think it is very clean and easy to understand, so it is fine if performance is not an issue. Maybe you also meant that it might break if a subdirectory of `/your/search/path` is in your path, but it can easily be fixed (thanks to glennjackman idea of `:`). It can even be simplified as: `grep -q ":/your/search/path:" <<< ":$PATH:" || export PATH=$PATH:/your/search/path`. – jfg956 Feb 19 '12 at 13:26
  • @jfgagne There are certainly better and worse ways to do things. The `case` statement is a shell built-in - it's better for at least two reasons: (1) no external dependencies and (2) it doesn't need to fork a new process (slightly better performance). So far you haven't proved your point as you've only pointed out that using `grep` can get similar results, but you have not mentioned any **advantages** of using `grep` over `case`. – jw013 Feb 19 '12 at 15:52
  • @jw013: I think the grep solution is more readable. Moreover, it fits on one line. I would not use it to add 10.000 elements to my PATH, but for a single addition, it fits most needs. ;-) – jfg956 Feb 19 '12 at 17:56