OK, this is so odd I know it's something I'm going to be something dumb that I'll be embarrassed about. When I created the testcase, it works fine. When I run the complete program, there's a problem.
But here's why that's odd -- instead of posting a lot of code, Excel spreadsheets, etc, I just put a Data::Dumper command for %categories right before I did the print in the original program. Then I created the testcase, copied the output from Data::Dumper and just set %categories to the content of Data::Dumper, then did the print using the same code. Works fine!
Now if I run the original program, I see the output from Data::Dumper isn't changed -- but the final printed output is screwed up. I'm confused since the content of %categories looks EXACTLY the same!
(I do agree, would be better/more flexible to use recursion! However, the testcase works)
Here's the testcase and it's output:
So.. that's correct. No issues with how I'm reading the hash.
However in the main program, the output looks like this. Look at accessories->bookends -- a whole new level of hierarchy at the third level with lots of bogus info!
<etc... rest of the errors snipped>
| [reply]
[d/l]
[select] |
I get Can't use string ("1") as a HASH ref while "strict refs" in use. You seem to be hiding errors by not using use strict; use warnings;. no autovivification; would also be useful to you.
This kind of mistake elsewhere in your code would definitely explain your symptoms. You're placing all your data in one hash (%1), and you are linking to that hash throughout the tree.
I think your data structure is equivalent to the one below. It would also explain the infinite looping you are experiencing with the recursive solutions.
%{"1"} = (
cat2 => \%{"1"},
cat21 => \%{"1"},
cat22 => \%{"1"},
cat3 => \%{"1"},
);
my %category = (
cat1 => \%{"1"},
);
| [reply]
[d/l]
[select] |