for $_ (%edict) {    # for entries with more than one keyword
[download]

in other words, given the hash %edict = ( a => 1, b => 2); $_ would contain 'a', then 1, then 'b' and then 2.

Also you're splitting whatever it is you're splitting on "; " (that's semicolon followed by a space), instead of a semicolon as you claim you want.

Update: as a first attempt at clarifying this, please explain what the top-level for loop is supposed to do and how that loop's body is effected by the loop (and especially, what you're doing with $_).

The following part of the script should look if there two words seperated by ";", split them,

@foo = split (/; /, $edict{$count}{english});

Note the extra space in the regex - it doesn't do what you described it should. If there's no space, then split will not match.

Anyway, your example uses quite some nested data structures of which we know nothing. We also don't know what the string looks like that you want to split. If you want more or better help, give us some data (and a simpler piece of code that does the same thing). See also How (Not) To Ask A Question.

When you do for $_ (@foo), $_ will sequentially contain the content of each element of the array @foo. Next you do $foo[$_], in other words you index into the array with the value of the element. This is almost certainly wrong, unless you are trying to implement some sort of linked list. I think you just want to use the content of each element and that is already in $_.

Sorry for the obscurities. As I said, I use hashes of hashes to store dictionary entries. So %edict{0} contains the first entry (0 is the key to the embedded hash) and &edict{0}{english} contains the english translation. That's mostly one word like "house" etc. but sometimes there are entries like "office; court; ....". That's why the split looks for "; ", since there is always a whitespace after the semicolon.

What I'm trying to do is to split the entry with more than one english words, put the them in the array @foo and then take the first word from @foo, put it in another array @engdict with other information belongs to it, so that I will have two seperate scalars in @engdict: One is office : xyz, and the other is court : xyz.

So that's what the code must do: 1. Find the entry about english translation with more than one word inside, seperated by "; " 2. Put them seperately in the array @foo 3. Take the first word from @foo and put it in $engdict[$no] with other information 4. Add one to $no 5. Take the second word and put it in $engdict[$no+1] and so on.

The problem is this is not working as I thought. The $foo[$_] is always empty so I think I'm doing something wrong with the split() command but I can't see the mistake. So here is my question: Can you tell me the mistake here?

Thank you

I think your problem is that you do too much at once without testing the result from the intermediate steps.

Most of my scripts start with

# always use those:
use strict;
use warnings;

use Data::Dumper;
[download]

When you suspect that split isn't working the way you think it is, try adding the line print Dumper \@foo; immediately after the split, and look if the results are what you execpted it to be.

The $foo[$_] is always empty so

for iterates over the values in the array, not the indexes. So instead of $foo[$_] just write $_.

Changing $foo[$_] with $_ solved the problem. Thanks a lot. Also thank you for the advice about Data::Dumper