Compare comandline arguments in bash

[anjoz]

how do you compare comandline arguments in bash example i have 5 or more arguments and 2 of the arguments are the same how can do it any suggestions? my problem is that i need to compare 2 or more arguments if any of the arguments are the same sleep for 1 sec

[thekashyap]

May be there is an easier way, but here is what I can think of..

total_params=`echo $* | sed 's/ */\n/g' | sort | sed '/^$/d' | wc -l`
unique_params=`echo $* | sed 's/ */\n/g' | sort -u | sed '/^$/d' | wc -l`
[ $total_params -ne $unique_params ] && echo "repeat" || echo "no repeat"

PS: There is a bit of catch with sort as it adds some newline chars by itself, that's why the /^$/d at end.

[L7Sqr]

To add to what thekashyap provided, you already have the count of arguments in the $# variable and, in bash at least, you do not need to worry about multiple spaces so you can get away without both sed calls.
For instance:

[ $# -ne `echo $* | tr ' ' '\n' | sort -u | wc -l` ] && echo "Duplicates"

Which is pretty terse, you may want to break it up a bit like

UNIQ=`echo $* | tr ' ' '\n' | sort -u | wc -l`
if [[ $# -ne $UNIQ ]]; then
   echo "Duplicates"
fi

[thekashyap]

To add to what thekashyap provided, you already have the count of arguments in the $# variable and, in bash at least, you do not need to worry about multiple spaces so you can get away without both sed calls.
For instance:

[ $# -ne `echo $* | tr ' ' '\n' | sort -u | wc -l` ] && echo "Duplicates"

Which is pretty terse, you may want to break it up a bit like

UNIQ=`echo $* | tr ' ' '\n' | sort -u | wc -l`
if [[ $# -ne $UNIQ ]]; then
   echo "Duplicates"
fi

Did you test your code? (I haven't)
I ask because as I said in my post:
"
PS: There is a bit of catch with sort as it adds some newline chars by itself, that's why the /^$/d at end.
"

[L7Sqr]

Did you test your code? (I haven't)

Yes.
Example run:

$ cat t.sh 
#!/bin/bash
[ $# -ne `echo $* | tr ' ' '\n' | sort -u | wc -l` ] && echo "Duplicates"
$ ./t.sh a
$ ./t.sh a b
$ ./t.sh a b a
Duplicates
$ ./t.sh a b a      d
Duplicates
$ ./t.sh a b c      d
$ ./t.sh a b c      d          3
$ ./t.sh a b c      d          3        "a  v"
$ ./t.sh a b c      d          3        "a  v"       "a  v"
Duplicates
$ 
$
$ echo "a b c d e" | tr ' ' '\n' | sort -u
a
b
c
d
e
$

As you can see I dont experience the problems you do with sort .

[Edit] You actually introduce the problem yourself by using sed in the first place. For instance:

$ echo "a b c d e" | sed 's/ */\n/g'

a
b
c
d
e

$

[anjoz]

what if example in the command line i wrote ./t.sh txt.txt /home/txt.txt how can i make it work when other arguments are paths

[thekashyap]

You actually introduce the problem yourself by using sed in the first place.

Thanks. Problem is with patter instead of sed . Following works like yours.

unique_params=`echo $* | sed 's/ /\n/g' | sort -u | wc -l`
[ $# -ne $unique_params ] && echo "repeat" || echo "no repeat"

PS: Now that you posted the examples, I couldn't help but notice that both our solns would fail for ./t.sh a b c d 3 "a v" .. :)
[EDIT] Just realized the simple soln is NOT to use $#..

[root@forssa tmp]# cat xxx
#!/bin/bash

uniq=`echo $* | sed 's/ /\n/g' | sort -u | wc -l`
tota=`echo $* | sed 's/ /\n/g' | sort | wc -l`
[ $uniq -ne $tota ] && echo "repeat" || echo "no repeat"

[ $# -ne `echo $* | tr ' ' '\n' | sort -u | wc -l` ] && echo "Duplicates" || echo "No duplicates"

exit 0

[root@forssa tmp]# ./xxx a b
no repeat
No duplicates
[root@forssa tmp]# ./xxx a b "c d"
no repeat
Duplicates
[root@forssa tmp]# ./xxx a b "c d" b
repeat
No duplicates
[root@forssa tmp]# ./xxx a b "c d" "c d"
repeat
No duplicates
[root@forssa tmp]#

what if example in the command line i wrote ./t.sh txt.txt /home/txt.txt how can i make it work when other arguments are paths

As is invariably teh case with YOUR homework, YOU would also have to do something. :)
1. txt.txt and /home/txt.txt: Are duplicates only if pwd=/home
2. You can use basename. If you want a dumb comparison. I.e. if two files have same name but are in different directory they still considered same.
3. For smarter comparision: You can get the inode number (using stat) of each path after following links (if any) and find duplicates in the list of inode numbers using the solns posted here.