Deleting hash entry

hotshot has asked for the wisdom of the Perl Monks concerning the following question:

hello guys!

I have a hash of hashes, for example:

%hash = {
         abc => {
                 def => 1,
                },
         ghi => {
                 jkl => 1,
                },
};
[download]

What happen if I do:

delete($hash{abc}{def});
[download]

am I staying with?:

%hash = {
         abc => undef,
         ghi => {
                 jkl => 1,
                 mno => 1,
                },
};
[download]

That was the first question.
My second one is, I'm looping over the hash and delete keys as shown above if fulfilled a condition. If the ansewer to the first question is yes, is there a short way to delete also the first level key ('abc' in the example) when deleted all second level keys under it. In the example, when performing the above delete, i would have like to stay with:?:

%hash = {
         ghi => {
                 jkl => 1,
                 mno => 1,
                },
};
[download]

Thanks

Hotshot

Comment on Deleting hash entry Select or Download Code

Replies are listed 'Best First'.
Re: Deleting hash entry by broquaint (Abbot) on Dec 16, 2003 at 15:05 UTC
Firstly, the first level of that hash assignment should be using parens and not curly braces. Now in answer to your first question, you will `delete` the `def` key in the hash pointed to by `$hash{abc}`. As for the second question, just test that the hash is empty with the use of `keys` in a scalar context e.g `use Data::Dumper; my %hash = ( abc => {}, ghi =>{ qw/ two key value pairs / } ); print "before: ", Dumper \%hash; delete $hash{abc} unless keys %{ $hash{abc} }; print "after: ", Dumper \%hash; __output__ before: $VAR1 = { 'abc' => {}, 'ghi' => { 'two' => 'key', 'value' => 'pairs' } }; after: $VAR1 = { 'ghi' => { 'two' => 'key', 'value' => 'pairs' } };` [download] HTH `_________ broquaint`	[reply] [d/l]
Re: Deleting hash entry by talexb (Chancellor) on Dec 16, 2003 at 15:05 UTC
What happen if I do: `delete($hash{abc}{def});` I would hazard a guess that the hash key value pair you have specified would be deleted. What happens when you try it? (I did not.) .. is there a short way to delete also the first level key ('abc' in the example) when deleted all second level keys under it. Sure, I would use `delete($hash{abc})` This would delete the top level and everything underneath it. And now, an editorial comment .. Please, I would strongly urge you (and anyone else) to go into the Perl debugger and try these things out rather than posting here. Wait until you've tried a bunch of things and are well and truly stumped. We love interesting, challenging questions. Questions that show no effort leave us uninspired. --t. alex Life is short: get busy!	[reply] [d/l] [select]
Re^2: Deleting hash entry by Anonymous Monk on Oct 19, 2012 at 12:57 UTC
Alex, your forgetting the usefulness of this forum as a reference.	[reply]
Re: Deleting hash entry by tcf22 (Priest) on Dec 16, 2003 at 15:01 UTC
I think you meant: `%hash = ( abc => { def => 1, }, ghi => { jkl => 1, }, );` [download] Hashes use (), hash refs use {}. `delete($hash{abc}{def});` [download] should work for deleting a second level key. - Tom	[reply] [d/l] [select]
Re: Deleting hash entry by Aragorn (Curate) on Dec 16, 2003 at 15:22 UTC
First of all, the notation `%hash = { ...` is wrong. You either assign a hash with `%hash = ( ...` or a hash reference with `$hash = { ...`. Watch those parentheses and curly braces. See perlreftut and perlref for more information. As for deleting: `#!/usr/bin/perl use Data::Dumper; $hash = { abc => { 'def' => 1, }, ghi => { jkl => 1, mno => 1, }, }; print Dumper(\$hash); delete($hash->{abc}->{def}); print Dumper(\$hash); __END__ $VAR1 = \{ 'abc' => { 'def' => 1 }, 'ghi' => { 'mno' => 1, 'jkl' => 1 } }; $VAR1 = \{ 'abc' => {}, 'ghi' => { 'mno' => 1, 'jkl' => 1 } };` [download] The second question: I don't know :-). Having an undefined hash value is perfectly valid, so the upper level is not automatically deleted when the lower level becomes undefined. You could add a statement like `delete $hash->{$key} if not keys %{$hash->{$key}}`. This deletes the key only if the hash it references is empty. Arjen	[reply] [d/l] [select]

And now, an editorial comment ..