Why can't my Perl subroutine see the value for the variable in the foreach loop that called it?

Question

I hope this is something straightforward that I'm doing wrong. I saw something online about "variable suicide" that looked good, but it was for an older version and I'm on 5.10.1.

Anyway - a variable that I declared - $RootDirectory - just suddenly loses its value, and I can't figure out why.

Here's a script to reproduce the problem. When I run through the script in debug mode (perl -d) I can get it to print out the $RootDirectory at line 21 and 26. But it's gone by line 30.

use strict;
my $RootDirectory; 
my @RootDirectories; 

@RootDirectories = (
   'c:\\P4\\EDW\\PRODEDW\\EDWDM\\main\\db\\'
   ,'c:\\P4\\EDW\\PRODEDW\\EDWADS\\main\\db\\'
   ,'c:\\P4\\EDW\\PRODEDW\\FJE\\main\\db\\'
   );

foreach $RootDirectory (@RootDirectories) { 
   # $RootDirectory = 'c:\\P4\\EDW\\PRODEDW\\EDWDM\\main\\db\\';
   # print ' In foreach ' . $RootDirectory. "\n";
   RunSchema ();
} 

exit(0);

sub RunSchema() { 
   # print ' In RunSchema ' . $RootDirectory. "\n";
   CreateTables ();
} 

sub CreateTables() { 
   # print ' In CreateTables ' . $RootDirectory. "\n";
   SQLExecFolder ('tbl');
} 

sub SQLExecFolder() { 
   print ' In SQLExecFolder ' . $RootDirectory. "\n";       # Variable $RootDirectory value is gone by now
} 

EDIT Thanks for all the comments! I think for now I'll use the "our" keyword which appears to work well - thanks Nathan. Also thanks toolic about the Use Warnings - I think I'm sold on that one!

The thing that continues to confuse me is why, when I did debug mode (perl -d), and stepped through the code, doing "p $RootDirectory" I got the expected output at lines 21 and 26, but not line 30. How is the situation different at line 30?

Also, I appreciate the comments about best practice being to pass $RootDirectory as a function parameter. I wanted to avoid that because I have so many functions following that - i.e. RunSchema calls CreateTables which calls SQLExecFolder. All of them would have to have the same parameter passed. Does it still make sense in this case, or are there any better ways to structure this?

Answer 1

What Nathan said is correct. That aside, why don't you pass in the value? It's better practice anyway:

foreach $RootDirectory (@RootDirectories) { 
   # $RootDirectory = 'c:\\P4\\EDW\\PRODEDW\\EDWDM\\main\\db\\';
   # print ' In foreach ' . $RootDirectory. "\n";
   RunSchema ($RootDirectory);
} 

sub SQLExecFolder { 
   my $RootDirectory = shift;
   print ' In SQLExecFolder ' . $RootDirectory. "\n";
} 

Answer 2

You're declaring $RootDirectory as the loop variable in a foreach loop. As far as I understand, that means that its value is localized to the loop, and its value is restored to its previous value at the end of the loop.

In your case the variable was never assigned, so at the end of the loop it returns to its previous value of undef.

Edit: Actually, the problem is that $RootDirectory is declared with my, so it is undefined in other scopes. In the functions RunSchema, CreateTables and SQLExecFolder the variable is undefined, regardless of the localization of the foreach.

If you want the variable to be declared for strictness, but want it to be global, declare $RootDirectory with our:

our $RootDirectory;

Edit: That being said, it's not always a good idea to use a global variable. You're better off passing the variable as a parameter to the functions as others have suggested.