Re: how to check for a word in a file and if found remove that line, the above line and the below line from the file.

You need to figure out if you're worried about overlapping matches, and specify your problem accordingly. There are solutions which will check excluded lines to see if there are additional lines to exclude after those. There are also solutions that will exclude a set of lines and move on as if the excluded lines after (but not before of course) the matched line can't possibly match. Here's an example of the difference.:

print me
exclude me
match me
match me
exclude me
print me
[download]

vs:

print me
exclude me
match me
would match me but the above line excluded me
print me anyway because the previous line wasn't matched after exclusi
+on
print me
[download]

Here's a filter that will do either, based on a command-line argument. It also is configurable for how many additional lines to exclude (before, after, or both).:

#!/usr/bin/perl
use strict;
use warnings;
use Getopt::Long ();

process( init() );
exit;

sub process {
    my $opts = shift;

    my @buffer = ();
    my $wait = 0;
    while ( <> ) {

        if ( ( scalar @buffer ) > $opts->{ 'b' } ) {
            print ( shift @buffer );
        }

        if ( $wait ) {
            $wait--;
            next unless $opts->{ 'nest' };
        } else {
            push @buffer, $_;
        }

        if ( /$opts->{ 'pattern' }/ ) {
            @buffer = ();
            $wait = $opts->{ 'a' };
        }
    }
    print @buffer;
}

sub init {
    my %options = (
        'help'    => 0,
        'a'       => 1,
        'b'       => 1,
        'c'       => 0,
        'nest'    => 0,
        'pattern' => '',
    );
    Getopt::Long::Configure( 'gnu_getopt' );
    Getopt::Long::GetOptions( \%options, 'help+', 'a=i', 'b=i', 'c=i',
+ 'nest+', 'pattern=s' );
    if ( $options{ 'c' } ) {
        $options{ 'a' } = $options{ 'b' } = $options{ 'c' };
    }
    if ( $options{ 'help' } ) {
        warn "Usage: $0 [[--help]|[[-a <n>] [-b <n>]|[-c <n>]] [--nest
+] --pattern <s>\n\n"
        . "Print the input file excluding the matched line provided by
+ the -p argument and as many lines before and after that line as spec
+ified.\n\n"
        . "\t--help\t\t\tthis help message\n"
        . "\t-a <number>\t\texclude <number> lines after the matched l
+ine\n"
        . "\t-b <number>\t\texclude <number> lines before the matched 
+line\n"
        . "\t-c <number>\t\texclude <number> lines before and after th
+e matched line, overriding -a and -b in the process\n"
        . "\t--nest\t\t\tmatch lines already excluded by a preceding m
+atch, and exclude the following lines accordingly\n"
        . "\t--pattern <string>\texclude the line matching this patter
+n (may be a regular expression) and other lines as specified by the o
+ther options\n\n";
        exit 0;
    }

    return \%options;
}
[download]

I made it a filter as the file handling itself is kind of secondary to the problem at hand. It's possible to add input and output arguments either positionally or via flags. My preference if I were to flesh it out a bit more would be to default to STDIN and STDOUT but allow flags to override one, the other, or both.

Comment on Re: how to check for a word in a file and if found remove that line, the above line and the below line from the file. Select or Download Code


Perl Monk, Perl Meditation
	PerlMonks