UNIX for Dummies Questions & Answers

Ok it works with bash but not csh. Do you know if you have to declare variables in csh?

Two things.

1. Avoid using c shell. It is horrible for coding.
2. Use the following script (rename.sh):

Code:

#!/bin/ksh
if [ -e $1 ]; then
  cp "$1" "$1"_fails
fi

as

Code:

ksh rename.sh abcd

---------- Post updated at 07:40 PM ---------- Previous update was at 07:39 PM ----------

I have not worked a lot with c shell. But I guess it is something like:

Code:

set var_name value

Thanks jaws using the set before the variable name worked. However if looks like with the c shell that the syntax is different for a lot of things. For the if statement it needs round brackets instead of square ones for example.

By the way your code produces the wrong output if you have a suffix at the end of the filename. I need the _fails before the suffix.

---------- Post updated at 03:51 PM ---------- Previous update was at 03:32 PM ----------

Hi Ranga,

How exactly does the last part of the following line work

Code:

    mv ${base} ${base/./_fails.}

It works fine with bash but not with csh. If I understood how it worked then maybe I could rewrite it for csh.

It takes the input filename test1.txt.
It then somehow puts _fails before the .txt.

My bad. This one inserts the "_failed" command before the extension.

Code:

#!/bin/ksh
if [ -e $1 ]; then
  ofile=`echo $1 | sed 's/\(.*\)\.\(.*\)/\1_failed.\2/g'`
  cp "$1" "${ofile}"
fi

Hope this helps.

http://www.grymoire.com/Unix/CshTop10.txt

Thanks Jaws. That works exactly as I wanted to but with the ksh. Unfortunately I must use csh or tcsh. I can't get it work with either. Do you mind explaining the logic you used with sed? I'm familiar with sed (I understand the basic use of sed using substitution). But I'm a bit lost with the syntax you have. Hopefully then I can convert it to csh syntax. Thanks again.

Your basic global substitution syntax for sed is:

's/search_term/replace_term/g'

This sed command says that:

1. search for any character combination, i.e. \(.*\). Here .* represents any number of occurrences of '.' which is wildcard character and the enclosing \( and \) are used to "mark" the pattern (as explained later)
2. followed by a single dot character, i.e. \. - I used the back slash so sed does not mistake the dot for the wildcard '.' which matches any character
3. followed by - again - any combination of characters - \(.*\) which is marked again.

Now in the replacement area I say:

\1_failed.\2

This basically means that the entire search term will be replaced by the term marked in point 1 (\1) followed by the string "_failed" followed by the term matched in point 3 (\2). This is the benefit of marking (i.e. enclosing search terms in \( and \) ) - you can manipulate the matched patterns around easily. Also notice that I don't need to backslash the '.' character in the replacement area.

Coming to your second problem (converting it to csh syntax), I think the syntax for sed is pretty much universal. You just need to change other parts of the script like the if-else statements to make it work.

While I understand that code in c-shell exists in a lot of systems and has to be maintained - I would like to reiterate: If you have any say in the matter, switch to some other better shell like bash or korn. It is easier to script in, and as neutronscott points out above - there are a lot of reasons to "not" use c shell.