Preventing "foo" from matching "foo-bar" with grep -w - regex

Question

I am using grep inside my Perl script and I am trying to grep the exact keyword that I am giving. The problem is that "-w" doesn't recognize the "-" symbol as a separator.
example:
Let's say that I have these two records:
A1BG 0.0767377011073753
A1BG-AS1 0.233775553296782
if I give
grep -w "A1BG"
it returns both of them but I want only the exact one.
Any suggestions?
Many thanks in advance.
PS.
Here is my whole code.
The input file is a two-columns tab separated. So, I want to keep a unique value for each gene. In cases that I have more than one record, I calculate the average.
#!/usr/bin/perl
use strict;
use warnings;
#Find the average fc between common genes
sub avg {
my $total;
$total += $_ foreach #_;
return $total / #_;
}
my #mykeys = `cat G13_T.txt| awk '{print \$1}'| sort -u`;
foreach (#mykeys)
{
my #TSS = ();
my $op1 = 0;
my $key = $_;
chomp($key);
#print "$key\n";
my $command = "cat G13_T.txt|grep -E '([[:space:]]|^)$key([[:space:]]|\$)'";
#my $command = "cat Unique_Genes/G13_T.txt|grep -w $key";
my #belongs= `$command`;
chomp(#belongs);
my $count = scalar(#belongs);
if ($count == 1) {
print "$belongs[0]\n";
}
else {
for (my $i = 0; $i < $count; $i++) {
my #token = split('\t', $belongs[$i]);
my $lfc = $token[1];
push (#TSS, $lfc);
}
$op1 = avg(#TSS);
print $key ."\t". $op1. "\n";
}
}

Answer 1

If I got clarifications in comments right, the objective is to find the average of values (second column) for unique names in the first column. Then there is no need for external tools.
Read the file line by line and add up values for each name. The name uniqueness is granted by using a hash, with names being keys. Along with this also track their counts
use warnings;
use strict;
use feature 'say';
my $file = shift // die "Usage: $0 filename\n";
open my $fh, '<', $file or die "Can't open $file: $!";
my %results;
while (<$fh>) {
#my ($name, $value) = split /\t/;
my ($name, $value) = split /\s+/; # used for easier testing
$results{$name}{value} += $value;
++$results{$name}{count};
}
foreach my $name (sort keys %results) {
$results{$name}{value} /= $results{$name}{count}
if $results{$name}{count} > 1;
say "$name => $results{$name}{value}";
}
After the file is processed each accumulated value is divided by its count and overwritten by that, so by its average (/= divides and assigns), if count > 1 (as a small measure of efficiency).
If there is any use in knowing all values that were found for each name, then store them in an arrayref for each key instead of adding them
while (<$fh>) {
#my ($name, $value) = split /\t/;
my ($name, $value) = split /\s+/; # used for easier testing
push #{$results{$name}}, $value;
}
where now we don't need the count as it is given by the number of elements in the array(ref)
use List::Util qw(sum);
foreach my $name (sort keys %results) {
say "$name => ", sum(#{$results{$name}}) / #{$results{$name}};
}
Note that a hash built this way needs memory comparable to the file size (or may even exceed it), since all values are stored.
This was tested using the shown two lines of sample data, repeated and changed in a file. The code does not test the input in any way, but expects the second field to always be a number.
Notice that there is no reason to ever step out of our program and use external commands.

Answer 2

You may use a POSIX ERE regex with grep like this:
grep -E '([[:space:]]|^)A1BG([[:space:]]|$)' file
To return matches (not matching lines) only:
grep -Eo '([[:space:]]|^)A1BG([[:space:]]|$)' file
Details
([[:space:]]|^) - Group 1: a whitespace or start of line
A1BG - a substring
([[:space:]]|$) - Group 2: a whitespace or end of line