Any Script to Change One Line in a File

Wow, nice ozo!  Could explain the first script?
perl -i -pe 's/^(?=three$)/> /' example.txt


perl -i allows

Wow, nice ozo!  Could explain the first script?
perl -i -pe 's/^(?=three$)/> /' example.txt


-i: allows editting of files in place
-e: allows one line of code on the command line
-p:   assume loop like -n but print line also, like sed (don't quite understand what this means, perform the line of code for every line in the file?)

Then the actual line of code:
s/^(?=three$)/> /

I thought I knew Regex inside and out but apparently not.  An online source said:
(?=pattern)
    A zero-width positive lookahead assertion. For example, /\w+(?=\t)/ matches a word followed by a tab, without including the tab in $&.

Does that mean that you replace the "nothing" in front of the word "three" with  '> ', because of the allow editting files in place?

Really great script, I am going to study your sed example now!

Joe P

My friend pointed out:
file_put_contents(str_replace($word, "> ".$word, file_get_contents($file)));
a nice php one-liner if you have php on the linux box.

Also this website repeats ozo's perl one liner search and replace: and my understanding of the script is better after studing it and playing around with it.
http://joseph.randomnetworks.com/archives/2005/08/18/perl-oneliner-recursive-search-and-replace/

Is it possible to perform the equivalent using the map function?  I figured it would be like this:
open(IN,'example.txt') or die("$!");
@lines = map { chomp; s/^(?=three$)/> / } <IN>
...

But from a failed attempt and some research shows that s/// does not work well inside map because it doesn't return $_, instead it returns a 0 or a 1.

I'll keep looking for somre more creative examples but I like ozo's perl script.

Keem em coming,
Joe P

@lines = map { chomp; s/^(?=three$)/>; $_ } <IN>

Small typo, this worked for me:
my @lines = map { chomp; s/^(?=three$)/> /; $_ } <IN>;

Thanks ozo, I'll leave this open until tommorrow afternoon and accept all the solutions.
Thanks for your help.  I'm still learning, didn't know I could do that in the map() functin.

Joe P

Excellent Perl_Driver, I confirmed it works with testing and also this website uses similar constructs:
http://www.thescripts.com/forum/thread591710.html

Speaking of which I have a few questions about your script after researching:
local @ARGV = ($FILE);
local $^I = '.bac';

I found that:
$^I -- Holds the file extension used to create a backup file for the in-place editing specified by the -i command line option. For example, it could be equal to ".bak."
http://www.cs.cf.ac.uk/Dave/PERL/node104.html
  -- and --
If $^I is set (either from within the program, or by the -i commandline switch), opening files with the diamond operator (<>) automatically performs a similar operation (allowing editing in place by holding a backup file. Notice I didn't need the loop for @ARGV; that's implicit in the diamond operator. The value set for $^I (normally undef) is added to the names of the files in @ARGV to create backup files.
http://www.stonehenge.com/merlyn/UnixReview/col23.html

So my question:
local @ARGV = ($FILE);
Just to make sure I have grasped it: This is making a list of one element (the file name) and putting into @ARGV solely for the block ({ ... }) so that the while(<>) will read from the $FILE as if it was on the command line because the <> reads $ARGV[0] like a file?  That is freaking brilliant!


ozo: yours is more complicated and I have more questions but I spent about 20-30 minutes studing Perll_Driver's script so I will have to give yours a fair amount of time as well.  Especially researching perl's command line arguments which I have just started looking at because of your excellent one-line scripts.

It is nice to see I went back and shortened my script a little cutting of the if(/^$WORD$/) { ... } down to s/^($WORD)$/> $1/ and I see that your new script uses this regex as well.

Thanks again both of you!
Joe P

Perl_Driver shortened script: (maybe microseconds slower, but no branching could speed it up?)
#!/usr/bin/perl
use warnings;
use strict;

# Vars (could be command line args)
my $FILE = $ARGV[0] ||  'example.txt';
my $WORD = $ARGV[1] || 'three';

{
   local @ARGV = ($FILE);
   local $^I = '.bac';
   while (<>) {
      s/^($WORD)$/> $1/;
      print;
   }
}

BogoJoker:

you are correct in all the comments you posted directed towards me.

I don't know if your shortened script would be any faster or slower than what I posted. But it is a good way to do the same thing.

I think you will enjoy this link too:

http://www.perl.com/pub/a/2004/10/14/file_editing.html

Excellent link Perl_Driver.  I have heard about the FMTYEWTK pages because I just started following this tutorial here: (and the introduction mentioned those pages)
http://www.perl.org/books/beginning-perl/

Thanks for the link, it will certainly help me, and I bookmarked it immediatly.

Again thanks to both you and ozo for helping me out and for suffering through my always long questions and comments =):

I'll get there yet!
perl -e '$points=$ARGV[0]; @scores=split(/_/, $points); print join("\n",@scores), "\n";' 250_250

Joe P

Hmm, Accepted Solution I guess is the one that received the most points.  Overall I liked the first one-liner and would have taken that as the "Accepted" but they were all great and I thank you guys once again.

Joe P