open SOURCE,"<source.file" or die "Error opening source.file: $!";
open DEST,">dest.file" or die "Error opening dest.file: $!";
while ($line = <SOURCE>)
{
  next if ( 1 .. ($line =~ /RE/) );
  print DEST $line;
}
[download]

First of all those die statements are very uninformative. At the least you should say the attempted action, the filename, and report the OS error with $!. For something this simple there is no reason to not just read from @ARGV with <>.

If you have ever had to get a broken script with useless error reporting working again at midnight when it turns out that someone (not me!) had been mucking around as root and left permissions wrong on one file in several thousand, you will appreciate why I am a stickler for this!

Secondly you will lose the end of the line with the RE, so you didn't quite answer the problem stated. OTOH I suspect that you did answer the question that really was meant.

#!/usr/bin/perl -w

use strict;

my $found = 0;

while (<>) {
        if ($found or /RE(.*?)/) {
                $found = 1;
                print;
        }
}
[download]

[ar0n@zendo perl]$ perl re.pl < INPUT > OUTPUT
[download]

#!/usr/local/bin/perl -w

open FH,"+<file_name" or die "go away";

@x = <FH>;

$keep_bit=0;
foreach (@x) {
    if (/^RE$/) {
        $keep_bit++;
    }

    if ($keep_bit) {
        push(@y,$_);
    }
}
seek (FH,0,0);
print FH @y;
truncate(FH, tell(FH));
close FH;
[download]

perl -i.bak -nwe'(/REx/ .. 0) && print'

Of course, that will print the line that matched the REx. What you really want is hard to tell -- whether you want to start with the line after the REx or whether you want the line with the REx on it, and just starting as soon as the end of the REx matches.

perl -i.bak -nwe'(s/.*REx// .. 0) && print'

That will provide option #2 -- you get everything directly following the REx match, including the text on the same line.

perl -i.bak -nwe'print if $match; /REx/ && $match++'

And this one gives you all of the lines immediately following the line where the REx matched.

All of these examples can also be wrapped in while (<>) { ... } instead of used with the command line switches if necessary.

-dlc

Hey! I like this, never crossed my mind to do a one liner. To be honest I don't care whether the RE line is in or out - it's only a theoretical problem (just a small diversion from writing bloody Sybase SB procs)


"We are all prompted by the same motives, all deceived by the same fallacies, all animated by hope, obstructed by danger, entangled by desire, and seduced by pleasure." - Samuel Johnson

OK, I'll give it a try:

• If your file fits in memory and you want the rest of the line after the RE try:
  perl -i -e 'undef $/; ($dump, $keep)= split /RE/, <>; print $keep;' file
• Now if you want to keep lines AFTER the RE this should work... but it does not! Anybody can point my mistake?:
  perl -i -p -e '<> until( $ok || m/RE/); $ok=1 ' foo

You need to set $_ explicitely here.

perl -i -p -e '$_ = <> until $ok || m/RE/; $ok=1' foo
[download]

Darn! I never realized that the <> did not assign $_. Why o why?

In any case thanks Fastolfe and dchetlin for pointing this out.

Wow ... peek into my brain as I look at your code above:

• Why would that work? What on earth are you trying to do?
• Oh, I see ... hmm, why doesn't that work? It's bizarre, but it should...
• Ah, duh. The magic diamond operator doesn't automatically assign to $_ -- it's the while magic that does that. So you're comparing the REx against the same $_ the whole time, causing an infinite loop.

Even if you did assign your diamond operator to a variable, you're still going to accidentally let the first line of the file through, because it's already been pulled out by the implicit while (<>) { ... } continue { print } that the -p switch gives you. Basically, using the diamond operator inside of a -p or -n script is going to do strange things. I'd avoid it. See my note further down on a non-confusing way to do this :-)

-dlc

I tried it and got something, but this spits out from the LAST line that contains a match to the end of the input file, and I can't see why it does that. How can I change this to spit out from the FIRST occurance of the RE?

perl -i -n -e "unless($ok || /RE/) {<>} else{print;$ok=1}" foo

deleting everything until the RE is found


"We are all prompted by the same motives, all deceived by the same fallacies, all animated by hope, obstructed by danger, entangled by desire, and seduced by pleasure." - Samuel Johnson

open FH, "<fh.txt" or die "$!\n";

while (<FH>) {

   push(@y, $_), next if $flag;

   /^RE$/ and push(@y, $_), $flag++;

}

close FH;
[download]

Grumble.... I didn't realize I wasn't logged in, that anonymous post was mine... ah well.