Recently i've been working on a script that parses a file and formats the data and spits it out (and it's a finished piece because it e-mails the results as well ; ) One of the odd things about the data is that if there is a duplicate line we want to keep it and insert some previous data before that line. What i came up with is something like:

# in some loop looking through files
if ($duplicate) {
   $data[$count]{VAl1} = $data[$count-1]{VAL1};
   $data[$count]{VAL2} = $data[$count-1]{VAL2};
   ...
   #flag data as being duplicate
   $data[$count]{VAL1} .= '/duplicate';
}
$data[$count]{VALX} = /(par)se_so(me)_(dat)a/;
next LOOP;
[download]

But the problem i have is that the previous data isn't getting sent into the array (of hashes) before the duplicate is sent into the array. When i run with the debugger it is putting the data into the array (and appending the duplicate flag), but apparently it's not being saved.

The reason i'm using $count to index @data is because i don't know which order the lines will go in, but i can check to make sure everything for one record has been pulled out (records aren't nested, thank whatever god you pray to : ) i never actually assign anything to an individual array spot, however, i'm always assigning to the hash inside the array spot. It's my understanding that this 'autovivifies' the spot in the array.

Since the changes are happening inside an if block my guess is that the newly autovivified entry is lost because it only gets lexical scope, maybe. But all of the data is getting set inside a while loop anyway, so why doesn't it all vanish after the while loop?

So the questions are (after that long intro):
a) am i correct that the changes are local even though the spot in the array is autovivified?
b) is there an elegant (or possibly kludgy-but-gets-job-done) way around this data being lost?

Please (un)confuse me,

jynx