File one is:

I_pron
would_mod
like_v
to_to
learn_v
this_pron
._sent
OK_ok
?_quest
[download]

and the second one is in this format:

I would like to learn this .
OK ?
[download]

Okay... let me be sure I understand what you're asking...

I'm guessing that your goal is to parse file two, and match each word against a lookup file (file one, which is in word_article format) and output the lookup file version of file two.

So, a more generalized way of asking is: Take a file/sentence, and lookup each token of the file/sentence and output the look-up version of the tokens.

In other words, an instant Perl sentence diagrammer :)

Assuming that's true, here's what I come up with (keeping the basic structure of your code intact... and fully admitting I'm not the uber coder some here are):

$DOC1="1.txt";
$DOC2="2.txt";

# Create an associative lookup array from file 1

open (DOC1,$DOC1);
while (<DOC1>) {
  chomp($_);
  $ind = (split /_/)[0];
  $lookup{$ind} = $_;
}
close (DOC1);

open (DOC2,$DOC2);
@lines=<DOC2>;
close (DOC2);

$outln = '';
foreach $line (@lines) {
  @words = split(/ /, $line);
  foreach $word (@words) {
    chomp($word);    # Get rid of stray carriage return
    $outln = $outln . " " . $lookup{$word};
  }
}
print "$outln\n";
[download]

That close to what you're after?

Trek