another unix script

alright i feel crappy askin for more help, but ive read all the man pages i can take, and im still left with some difficulties. I know i should either use sed or awk or maybe even vi, but can anyone help me out some of those to scane a file (.profile) for a line and then rewrite that whole line? an example would be to add a new printer or something, in the .profile would be a line say like "printerdest=" and after that line could be xxxxxx or whatever, i want vi awk or sed to scan the file for "printerdest=" and replace it with what im specifying in the script. is there a specific command in awk or sed that im not seeing? help or guidance would be much appreciated. thanks.

Perl and regular expressions.

Sed will work just fine. Say I have a file called users.txt. It looks like this:
Al
Ben
Jon

Now say Jon's name is spelled wrong. To change it to John, we would do:
sed s/Jon/John/ users.txt >users.temp
mv users.temp users.txt

The sed command uses a regular expression substitution (s/). It looks for Jon and replaces it with John in the file called users.txt Since sed is a stream editor it sends the output to stdout, or the terminal, so use > to redirect to a temp file, then rename the temp file to the original.

Apply this to your .profile and it should work fine.

ok, see that seems fine but sometimes the original strings are different, what if it was something like "printerdest=xxxxxxx" and sometimes "printerdest=printer1", is there a way it will just scan for "printerdest=" and it will rewrite the whole line, regardless if there is a printer there or not?

I haven't done a good simple Q&D in perl lately, so here you go...

Code:

---

#!/usr/bin/perl -w

$print = $ARGV[0] ? $ARGV[0] : "new_printer";

$file = "text.txt";
open FOO, "<$file" or die "Can't open $file: $!\n";

$new_file = "test2.txt";
open BAR, ">$new_file" or die "Can't open $new_file: $!\n";

while (<FOO>) {
  $string = $_;

  if (/^printerdest=/i) {
    $new_line = "printerdest=$printer\n";
  } else {
    $new_line = $string;
  }

  print BAR "$new_line";
}

---

The above is a perl script that will rewrite the given file test.txt into test2.txt with the new printer destination. Take the below example:

If test.txt has this:

Quote:

---

This is a test file that has
some information relating
to a profile and some
printerdest=blahblahblah
user information for you
to peruse...

---

Then if you ran './foo 12345', and did a 'cat test2.txt', you would see the following:

Quote:

---

This is a test file that has
some information relating
to a profile and some
printerdest=12345
user information for you
to peruse...

---

Otherwise, if you ran './foo', you would get the following:

Quote:

---

This is a test file that has
some information relating
to a profile and some
printerdest=new_printer
user information for you
to peruse...

---

Hope this helps!

I am along with Juridan an Vorlin on this one, perl is the way to go.

I am not sure that I get your question right, so I don't know if Vorlin's script will work. What I think your are asking for is just a script that will scan a file and check for printer* and if it scans and gets that it will rewrite that to another file, with just that info.

Vorlins script does just about that just a little more also.

hmm...

Let me know if Vorlins is what you are looking for, if it isn't I will write something up for you real quick.

Though if you don't like using perl you could always write up a bash with some grep in it.

like so.. (did something quite similar a while back)

Code:

---

#!/bin/bash

#
#      (c) 2004 the_JinX ( Anne Jan Brouwer )
#                      GNU GPL
#

if [[ -z $1 ]]||[[ -z $2 ]]||[[ $1 == "--help" ]]; then
    echo
    echo "    Parameter replacer v0.1"
    echo "Usage:    $0 keyword value < input"
    echo "Optional: $0 keyword value < input > output"
    echo
    exit
fi

while read line; do
    if [[ `expr match "$line" ".*$1=*."` -gt 0 ]]; then
        echo ${line%%=*}=$2
    else
        echo $line
    fi
done

---

as you can see the actual code is only about 7 lines..

To match vorlins example..

./foo printerdest 12345 < test.txt > test2.txt

I use this script on /etc files sometimes (placed this script in the file /usr/local/bin/chparams on my main box)

My script will only change one line, and that's any line beginning with 'printerdest=', otherwise it simply writes out line for line into another file.

To make the prior script a lot more beefier, one could do this:

Code:

---

#!/usr/bin/perl -w

# Set parameters and test stuff...

$printer = $ARGV[0] ? $ARGV[0] : "new_printer";
$curr_file = $ARGV[1] ? $ARGV[1] : NULL;
$new_file = $ARGV[2] ? $ARGV[2] : NULL;

# Set counters and other integer variables

$count = @ARGV;
$i = 0;
$found = 0;

if ($count < 3) {
        print "Usage: $0 <printer> <file> <new_file>\n";
        exit 1;
} elsif (! -e $curr_file || (-e $curr_file && ! -r $curr_file)) {
        print "Error: file '$curr_file' doesn't exist or can't be read, aborting.\n";
        exit 1;
}

# Got this far so now open files and such...

open FOO, "<$curr_file";
open BAR, ">$new_file";

while (<FOO>) {
        $i++;
        chomp($string = $_);

        if (/^printerdest=/i) {
                $found = $i;
                $old_printer = $string;
                $old_printer =~ s/printerdest=//i;
                $new_line = "printerdest=$printer";
        } else {
                $new_line = $string;
        }
       
        print BAR "$new_line\n";
}

if ($found > 0) {
        print "Current file: $curr_file\n";
        print "New file: $new_file\n";
        print "Old printer: $old_printer\n";
        print "New printer: $printer\n";
        print "Found on line: $found\n";
} else {
        print "No printer destination line found in '$curr_file'.\n";
        print "Unlinking new file '$new_file'.\n";
        unlink($new_file);
}

---

Now, when you do it, say with './foo blarghorama test.txt test2.txt', and your test.txt has the following:

Quote:

---

This is a test of some stuff in a file to check to
see if it has a printerdest= in it somewhere and
I hope it does because then this works and some
printerdest=foobar
test test test

---

Then the output would be as such:

Quote:

---

Current file: test.txt
New file: test2.txt
Old printer: foobar
New printer: blarghorama
Found on line: 4

---

If file 'test.txt' didn't have a printerdest= line, it would say this:

Quote:

---

No printer destination line found in 'test.txt'.
Unlinking new file 'test2.txt'.

---

Sure, we could check for null printerdest fields, and do some other stuff like formatted output, but who's counting at this point, hehe...

OK Vorlin..
You win the pissing contest, your script is bigger then mine ;)

I think you are compensating :P

Thx for the previous version idea Vorlin..
I've updated to version 0.2 ;)
version two now adds the old keyword and parameter(s) with a # in front of it, to comment it out :D

Code:

---

#!/bin/bash

#
#      (c) 2004 the_JinX ( Anne Jan Brouwer )
#                      GNU GPL
#

if [[ -z $1 ]]||[[ -z $2 ]]||[[ $1 == "--help" ]]; then
    echo
    echo "    Parameter replacer v0.2"
    echo "Usage:    $0 keyword value < input"
    echo "Optional: $0 keyword value < input > output"
    echo
    exit
fi

while read line; do
    if [[ `expr match "$line" ".*$1=*."` -gt 0 ]]; then
        echo ${line%%=*}=$2
        echo "#"$line
    else
        echo $line
    fi
done

---

ROFL, nah...yours is much more concise as far as brevity is concerned. I generally like to comment on sections of code, do a lot of error checking and print out a lot of stuff for the end-user's benefit (having done customized output for departments and the like). It also makes it easier for me to debug later.

That's the beauty of programming...as Larry Wall said about perl oh-so-long ago, "There's more than one way to do it" and that definitely applies here.

That's what opensource is all about..

That's what I like about unixes so much !!
You could do this that and thensome or something totally different and get the exact same end result..
Some like it one way, some the other..

If you know exactly what you want to find and replace it with, then this is a quick fix...
perl -e 's/TextToFind/NewText/gi' -p -i Filename

It will find all instances of "TextToFind" and replace it with "NewText" in file "Filename"

yup.. that's one way, but why use perl..

sed -e 's/TextToFind/NewText/gi' -i.old Filename

this also makes a backup of the origional, called Filename.old

lol thanks guys. i wasnt plannin on using perl but now i might. thanks alot, especially for the bash script jinx. appreciate it.