"Perl script is getting stuck for no reason"

Unlikely. Since you refuse to show the code in question or create an example which replicates your issue (see How do I post a question effectively?) I suggest you learn to profile and debug your code. Tutorials->Debugging and Optimization/Basic debugging checklist. Devel::NYTProf.

Update: added link.

Update 2: Please mark your posts as updated, this can't be your code, it doesn't even run. Ensure you don't miss parts when copy/pasting.

"Perl script is getting stuck for no reason"

There's always a reason. An apparent lack of reason is an indication of inadequate investigation.

"...it is getting stuck at random..."

That is almost impossible; computers are useful because they are deterministic. Of course you're probably aware that computers don't behave randomly, or you wouldn't be asking for assistance with the behavior you're seeing. But the sooner you abolish all notion that something is happening randomly, the sooner you'll be on your way to diagnosing the problem yourself.

The code you posted doesn't compile. But you're not asking about why it doesn't compile, you're asking why your code seems to hang. So that tells me that the code you provided for us to look at is not exactly the code you're running. That makes it difficult for us to know what's wrong.

Fixing the right curly bracket, I then get three warnings. Have you investigated why you're getting warnings? Are you getting warnings, or is that something that only happens in the code you provided us?

We can't debug this for you; we don't have the data, or the real code. But let me offer these suggestions, which with a little diligence, will probably give you an answer. For what it's worth, these are steps I would probably take.

• Add "warn" statements that tell you where you are within your code, and where you are within loops.
• Run this with "top" going in another terminal. Watch for memory growth. The "warn" statements previously mentioned will help you to see what parts of your script are consuming memory, as you follow the runtime with top.
• Test with a much smaller subset of your data sets. This will allow for quicker iterations of the debug/test cycle.
• Even with your much smaller subset, don't be satisfied with any substantial growth in memory usage as the script progresses. Keep in mind that you are dealing with very large files. If it becomes too costly to hold intermediate data in memory, put intermediate results into files or a database, and process those files rather than trying to hold the entire computation in memory at once.

If it seems like I'm talking a lot about memory, it's because without a running script and small test data set, I can only go with instinct, and my instinct is that when I see some data structures persisting throughout your script, and read that you're dealing with 800MB, 16GB, and 500MB files, and when I read that your complaint is the script grinding to a halt, you are consuming too much memory and bringing your system to its knees.

So, to sum it up, code that we don't know is operating on data we haven't seen and failing in an unspecified way. I'm sure I'm not the only monk shrugging helplessly now, thinking "how exactly do you think we're supposed to help you?"

Please give us something to work with. If your code's too large to post, whittle it down to the a simple test case that exhibits the problematic behavior. If your input files are too large to post (or confidential), provide some simpler ones that show the problem.

Attempting to do this may be instructive in itself: perhaps while simplifying, you'll realize what the problem is. Or perhaps you'll cross a threshold below which it'll work, which may provide clues as to what's going wrong (and why, and how to fix it). Even if you can successfully create a small, postable test case, perhaps it'll be simple enough that you'll see the issue yourself. If not, at least the monks will have something to work with.

Some general tips and links for asking questions, copied from my homenode:

• How to ask questions (based on ww, Re: Replace key pair value from one to other file):

  1. Read the instructions ("Asking questions effectively", "Formatting your write-up", below).
  2. Read the documentation.
  3. Show effort. Write some code; at the very least, try. Help is free; doing your job for you is not.
  4. Show us your code.
  5. Describe failures, expected results, and actual results.
  6. If applicable, show us verbatim (!) error messages/warnings.
  7. If applicable, give us some sample data.
  8. Give us the larger picture: tell us what you want to achieve, not just how you decided to go about achieving it. There may be better ways of doing it that you haven't contemplated.

• Asking questions effectively:

  • On asking for help
  • How do I post a question effectively?
  • How do I compose an effective node title?
  • I know what I mean. Why don't you?
  • How (Not) To Ask A Question
  • Clean your room
  • XY Problem
  • How to RTFM
  • The SSCCE: Short, Self Contained, Correct (Compilable) Example
  • A few guidelines for good code, from Laurent_R: Re: text files are printed after the end of second module
  • Where should I post X? / PerlMonks Sections

I suspect you mean "hang" in the sense of "doesn't seem to be producing any output even after a long time" as opposed to a more precise use of "hang" for "stops consuming CPU but doesn't exit".

The most likely explanation is that the performance of your script is not linear so the much bigger file just takes significantly longer to process. The most likely explanation for seriously non-linear performance given some of what you said is that your process ends up using more memory than your computer can efficiently provide it.

But such possibilities are easy for you to investigate, requiring, at most, some internet searching and minor research to learn a couple of commands and how to interpret part of their output (such as top and ps).

Of course, maybe you have already investigated those possibilities and eliminated them. But I can't tell that based on what little information you provided.

Identifying whether or not the problem is just "using too much memory" is a good step to do before diving into the search for a more specific source of the problem. Hence, I didn't dive into the code you have now added to your posting.

Good luck.

BTW, further updates to your question are better done as replies rather than as modifying the original question. The site does not provide a way for people to notice that you've made an update (while there are several heavily used way to notice new replies). Also, updates tend to make the thread confusing.

use Acme::Tasseography;

my $at = Acme::Tasseography->new(<<ISSUE);
oooh... 2 inputs files
oooh... of size 16gb & 800mb
oooh... 4 hashes
ooohh...
oooooohhhh... around 500mb
oooooooooohhhh... it stucks
ISSUE

say $at->tell_fortune;
[download]

Look at line 100...
[download]

hire a consultant...
[download]