use strict;
use warnings;
use Data::Dumper;

my @array=('GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : tsm
+c11_wld(sxfatd12j)) GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSPA
+N04] : tspan04-dfdint-wld(sxfamd6f)) POS_WLI02 - POS_WLI02 Web Logic 
+Integrator');

my @newarray;
while ($array[0] =~ /(\w+) - /g) {
    push @newarray, $1;
}
print Dumper(\@newarray);

__END__

$VAR1 = [
          'GAP_SPAN09',
          'GAP_SPAN03',
          'POS_WLI02'
        ];
[download]

• perlintro
• perlre
• perlretut

Using your updated data which is now an array:-

knoppix@Microknoppix:~$ perl -Mstrict -wE '
> my @array = (
>    q{GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : tsmc11_w
+ld(sxfatd12j))},
>    q{GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSPAN04] : tspan04
+-dfdint-wld(sxfamd6f))},
>    q{OS_WLI02 - POS_WLI02 Web Logic Integrator},
>    );
> my @newArray = 
>    map m{^(.*?)\s-\s},
>    @array;
> say for @newArray;'
GAP_SPAN09
GAP_SPAN03
OS_WLI02
knoppix@Microknoppix:~$
[download]

I hope this is helpful.

You realize @array has only one element, the long single-quoted string?

Perhaps this:

use strict;
my @array=('GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : tsm
+c11
+_wld(sxfatd12j)) GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSPAN04
+] : tspan04-dfdint-wld(sxfamd6f)) POS_WLI02 - POS_WLI02 Web Logic Int
+egrator');
my @newarray = $array[0] =~ /(\S+) - /g;
print "$_\n" for @newarray;
[download]

Apologies the example given should have looked like this

my @array=('GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : tsm
+c11_wld(sxfatd12j))','GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TS
+PAN04] : tspan04-dfdint-wld(sxfamd6f))','POS_WLI02 - POS_WLI02 Web Lo
+gic Int
+egrator');
[download]

So there are 3 elements. This was purely an example though since there could be upto 20 elements in the original array

Ahh, the question becomes clearer though the logic is largely similar

use strict;
use warnings;
my @records=('GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : t
+smc11_wld(sxfatd12j))',
    'GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSPAN04] : tspan04-d
+fdint-wld(sxfamd6f))',
    'POS_WLI02 - POS_WLI02 Web Logic Integrator)');
my @record_ids;
for my $record (@records){
    $record=~m/^(\S+)(?= - )/g;
    push @record_ids , $1;
}
for my $record_id (@record_ids){
    print "$record_id\n";
}
__END__
GAP_SPAN09
GAP_SPAN03
POS_WLI02
[download]

print "Good ",qw(night morning afternoon evening)[(localtime)[2]/6]," fellow monks."

my @newarray = map /^(\S+) +- /, @array;
[download]

use strict;
use warnings;
use 5.012;

my @array = (
   'GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : tsmc11_wld(
+sxfatd12j))',
   'GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSPAN04] : tspan04-df
+dint-wld(sxfamd6f))',
   'OS_WLI02 - POS_WLI02 Web Logic Integrator',
);

my @results = map{(split / - /)[0]} @array;
say for @results;
[download]

GAP_SPAN09
GAP_SPAN03
OS_WLI02
[download]

Thanks for all the responses. It just goes to show how versatile Perl is !

Here are some tips that might help you on your way:

1. In perldoc perlre, carefully review the //g, //c, //s and //m modifiers.   Also consider the difference between using a regex in a scalar vs. a list context.
2. Yes, there is such a thing as “regex golf,” which is the amusing geek-pasttime of trying to Name That Tune In One Note.   But clarity usually wins the race, and if that means writing several regexes and breaking down the string in several stages ... do so.   After all, after you write the thing, you will also be maintaining it.
3. The join operator is very handy for constructing “comma-delimited strings.”
4. There are gobs of “regular expression test sites” on the Internet.   Nearly all regex engines are “Perl compatible.”
5. (Just in case...)   Far from being “dismissive,” we are seriously endeavoring to help you.   Attempt to solve the problem and show us your attempts.   Instead of “handing you a fish,” we’d like to “teach you to fish.”   This is an extremely frequently-done task that is one of the key reasons for “what is the fuss all about” with regards to Perl.   String-mangling is one of the reasons why Perl is referred to as a Swiss Army Knife.®

Thus @array=$string=~/PATTERN/g is the syntax you need.

The PATTERN is to capture a series of non space characters if they are followed by the ' - ' pattern.

If you need more help than that:

~/$ perl -e '$string=q(GAP_SPAN09 - GAP SPAN base (Scratch Testing [TS
+MC11] : tsmc11_wld(sxfatd12j)) GAP_SPAN03 - GAP SPAN base (DFD E2E Te
+sting [TSPAN04] : tspan04-dfdint-wld(sxfamd6f)) POS_WLI02 - POS_WLI02
+ Web Logic Integrator);@array=$string=~m/(\S+)(?= - )/g;for $record_t
+itle (@array){print "$record_title\n";}'
GAP_SPAN09
GAP_SPAN03
POS_WLI02
[download]

use strict;
use warnings;

my $string ="GAP_SPAN09 - GAP SPAN base (Scratch Testing [TSMC11] : ts
+mc11_wld(sxfatd12j)) GAP_SPAN03 - GAP SPAN base (DFD E2E Testing [TSP
+AN04: tspan04-dfdint-wld(sxfamd6f)) POS_WLI02 - POS_WLI02 Web Logic I
+ntegrator";

my @array;
while ($string =~ m/(\w+\s?)-/g)
{
    push (@array, $1);
}


print join("\t", @array);
[download]

 foreach (@array) {
     @pieces = split / - /;
     foreach (@pieces) {
       #then just grab the last full word 
       # do regex stuff here.
       if (/(\w+)$/) {
         push @newarray, $1;
       } 
     }
 }
[download]