Iterating over a complex data structure - perl

Question

I have what appears to be a hash of a hash of an array of hashes. I'm trying to pull some values out and I'm stumped (this is way deeper than I would go with a structure. It looks like this.....
%htest = (
8569 => {
4587 => [
{
date=> "2011-01-15",
approved=> 1,
},
{
date=> "2011-01-12",
approved=> 1,
},
],
1254 => [
{
date=> "2011-01-12",
approved=> "",
},
{
date=> "",
approved=> 1,
},
],
},
);
Trying to iterate over this thing is giving me a massive headache. I'm trying to access the number of elements under the second hash value (4587 and 1254). The number of those elements where approved="1" and the number of elements where the date contains a value.
If I could iterate over them I sure I could shove what I need into a less complex structure but so far I'm at a loss.
I got this far...
while (my ($id, $surveyhash) = each %{ $htest{'8569'} } ){
print "$enumid = $subhash\n";
print Dumper $subhash."\n";
}
That gives me the "4587" and "1254" but trying to do a dumper on $subhash just gives me....
4587 = ARRAY(0x9a9ffb0)
$VAR1 = 'ARRAY(0x9a9ffb0)
';
1254 = ARRAY(0x9a91788)
$VAR1 = 'ARRAY(0x9a91788)
';
Any idea how to iterate over this monstrosity?
Janie

Answer 1

Your structure has typos, you need commas between the innermost hashes and a parenthesis at the end (rather than a curly bracket)
Once you fix it you can use something like this:
my $approved = 0, my $date_has_value = 0;
while ( my ($k,$vref) = each %htest ) {
while ( my ($k,$v) = each %$vref ) {
# Now you're inside the inner hash, so there will be 2 iterations
# with $k 4587 and 1254
foreach my $item (#$v) {
# Now each $item is a reference to the innermost hashes
$approved++ if $item->{approved} == 1;
$date_has_value++ if $item->{date};
}
}
}

Answer 2

Here's a fairly explicit iteration that should help get you started
my ($num_approved, $num_date) = (0, 0);
# outer hash
while (my ($ka, $va) = each %htest)
{
# inner hash
while (my ($kb, $vb) = each %{$va})
{
# each hash inside the array
foreach my $h (#{$vb})
{
$num_approved += ${$h}{"approved"} == 1;
$num_date += length(${$h}{"date"}) > 0;
}
}
}

Answer 3

Counting match cases can be done with "scalar grep".
my ($approved, $date_has_value) = (0, 0);
for my $v1 (values %htest) {
for my $v2 (values %$v1) {
$approved += scalar grep { $$_{approved} eq '1' } #$v2;
$date_has_value += scalar grep { $$_{date} ne '' } #$v2;
}
}