in reply to Re^3: need explanation of @foo{@bar} = (); (hash slice)
in thread need explanation of @foo{@bar} = (); (hash slice)

But defined doesn't do anything to prevent autovivification either. Look at this: 


my %hash = ( Key1 => { john => 1, pete => 2 },
             Key2 => { frank => 3, howard => 4 } );
print "Exists!\n" if defined $hash{Key7}{ted};
print "Key7 exists!\n" if exists $hash{Key7};

[download]

Here you're testing the definedness of a key that doesn't exist: $hash{Key7}{ted}, and then you're going to check to see if $hash{Key7} autovivified. Guess what? It did, just the same as when we used exists to check for the existance of $hash{Key7}{ted}. The same conditions that will cause exists to autovivify a hash element will also cause defined to autovivify a hash element. exists doesn't have any more possibility to do that than any other function.


Dave

Replies are listed 'Best First'.
Re^5: need explanation of @foo{@bar} = (); (hash slice)
by tlm (Prior) on May 09, 2005 at 02:24 UTC

    As I said, I don't claim that exists is a more efficacious way to induce autovivification than defined (or than anything else for that matter), but rather that it is a bad test to use, since there is the possibility of unintended autovivification happening anywhere else, and when this happens exists gives misleading results. In other words, using a slight modification of your example (s/defined/exists/): 

    my %hash = ( Key1 => { john => 1, pete => 2 },
             Key2 => { frank => 3, howard => 4 } );
print "Exists!\n" if exists $hash{Key7}{ted};
print "Key7 exists!\n" if exists $hash{Key7};

    [download]
    it is the second exists that I find problematic is not the first one, because it is the second exists that gives the misleading result (yes "Key7" exists but for the wrong reason). I'd just as soon have both tests fail, by using defined instead of exists for both.

    the lowliest monk

[reply]
[d/l]

In Section Seekers of Perl Wisdom