Question

I have a tab-delimited text file. I have split this into columns. Each of the first 2 columns contains an ID number.

I want to keep all lines with ID number starting with P or Q, and remove any other lines where column 1 or 2 has any other ID or is blank.

eg. so columns to be kept will be like this: P12345 or Q12345. Columns to get rid of will be GAG123, CH123 etc. or just blank.

I can’t work out how to do this. I have tried splitting lines into arrays and grep /^[PQ]/elements [0] and [1], and various other things, but I must be doing something wrong.

I’ve tried the follwoing code below from TLP, but it won’t work, I know I must be doing something fundamentally wrong:

#!/usr/bin/perl  

use warnings;
use strict;

open(FILE,"<myfile.txt"); 
my @LINES = <FILE>; 
open(my $outfile, '>', 'changedtxt');
my @wanted;

while (<FILE>) {
    my @fields = split('\t', $_);
    if ( $fields[0] =~ /^[PQ]/ and $fields[1] =~ /^[PQ]/ ) {
        push @wanted, $_;  
        print {$outfile} $_;    
    }
}
exit:

Answer 1

use strict;
use warnings;

my @wanted;
while (<$fh>) {
    my @fields = split /\t/, $_;
    if ( $fields[0] =~ /^[PQ]/ or $fields[1] =~ /^[PQ]/ ) {
        push @wanted, $_;
    }
}

If you want both IDs to begin with P or Q, exchange or for and.

If you simply want to move the wanted lines to another file, simply do:

perl -wnae 'print if (($F[0] =~ /^[PQ]/) or ($F[1] =~ /^[PQ]/))' input.txt > output.txt

Or as a script, use with script.pl input.txt > output.txt:

use warnings;
use strict;

while (<>) {
    my @fields = split(/\t/, $_);
    print if ( $fields[0] =~ /^[PQ]/ and $fields[1] =~ /^[PQ]/ );

} 

Note that you can’t use '\t' as a split pattern.