Something like this should do it (Obviously, the comments should be removed before use!):

#! perl -slw
use strict;

my $pair = 'A25 A15';

## Assumes the 6 of 125 you need are the 3rd, 33rd, 50th, 75th, 100th,
+ and 123rd,

my $tmplNeeded = join ' ', <<"EOT";
    x[($pair)2]     # skip 2
    $pair           # grab the 3rd
    x[($pair)29]    # skip 29
    $pair           # grab the 33rd
    x[($pair)16]    # skip 16
    $pair           # Grab the 50th
    x[($pair)24]    # ...
    $pair
    x[($pair)24]
    $pair
    x[($pair)22]
    $pair
EOT

open FILE, '<', ...;

while( <FILE> ) {
    my @required = unpack $tmplNeeded, $_;
    ## $required[ 0 ] is the 3rd field
    ## $required[ 1 ] is the 33rd field etc.
}
[download]

Say that you read a line in $line, then this code:

populates the hash %vars with the variable/value couples from $line.

I don't remember where/when I saw the benchmark but this is surprisingly fast (fixed width regexes -- you'd probably want the '\A' in front too). I'm sorry I don't have tuits to do a new chart and can't find the old script now but I remember being shocked at how fast this was and that it was faster than most of the alternatives. Perhaps some kind Monk remembers what I saw or is feeling charitable with Benchmark.

The problem with that comes if you do not know what the keys are, or what order they appear in, how do you pick out the nth field from the hash?

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

"Too many [] have been sedated by an oppressive environment of political correctness and risk aversion."

From the word choice in the original node, I understood that the OP knows the keys ("variables") and not necessarily the positions.

If the position are fixed and known, it could work this:

my @vars=$line=~/.{25}(.{15})/g;

or this, that keeps the keys interleaved with the values:

my @vars=$line=~/(.{25})(.{15})/g;

Not tested (I've not a perl available right now) but both of them should work.

Rule One: "Do not act incautiously when confronting a little bald wrinkly smiling man."

use strict;
use warnings;
#define a regexp matching the interesting variable names
my $interesting_vars = 
  qr(a111111111111111111111111|c222222222222222222222222);
#sample input rows
my @rows = ('a111111111111111111111111              1b8888888888888888
+888888888            15x222222222222222222222222              2', 
            'd999999999999999999999999              4b3333333333333333
+333333333            15c222222222222222222222222            123');

for (@rows) {
  #split by variable value pairs  
  for (/.{40}/g) {
    #split variable and value
    /(.{25})(.{15})/;
    #since I am doing an additional match, I have to
    #save my submatches
    my $var = $1;
    my $val = $2;
    print "'$var' = '$val'\n" 
      if $var =~ $interesting_vars;   
  }
}
[download]

'a111111111111111111111111' = '              1'
'c222222222222222222222222' = '            123'
[download]