Creating members of hash of hashes parallelly

[utku]

Hi,

I cannot figure out how to create a HoH (or even HoHoH) from a file in
+ following format:

path       start                 end                 delay
----------------------------------------------------------
my_path    /path/3434/some_txt   wodo                -445
your_path  /path/3434/some_txt   wodo_modo           41
my_path2   /path/232/wge45       wodo_modo           240
my_path    /t43/annow23x         /a/b/c/43           NA
...

Each line above corresponds a "record" and belongs to a "path".
So a "path" can include several "records".
Above, "my_path" includes 2 records, ie. 1st and 4th lines.

I just want to store all lines above into a single hash
in the form of:

$db_hash{$path}{<index_for_corresponding_path>}{START}
$db_hash{$path}{<index_for_corresponding_path>}{END}
$db_hash{$path}{<index_for_corresponding_path>}{DELAY}

For ex, 1st line will be:
$db_hash{my_path}{1}{START} = "/path/3434/some_txt"
$db_hash{my_path}{1}{END} = "wodo"
$db_hash{my_path}{1}{DELAY} = "-445"
...

$db_hash{my_path}{2} must be generated from the 4th line and so forth.

The number of "paths" are determined during runtime,
thus I must also determine indices for each "path" on runtime.
The number of members in each "path" can be different.

My question is, how to create index for each unique $path key?
I shall create a separate array of indices, shan't I?

[japhy]

You don't want a hash in that middle level, then, you want an array: $db_hash{$path}[$idx]{$key}. To do that, you'd do:

while (<>) {
  my ($path, $start, $end, $delay) = split;  # or however you get them
  push @{ $db_hash{$path} }, {
    START => $start,
    END => $end,
    DELAY => $delay,
  };
}

[liverpole]

Hi utku,

If I understand you correctly, the following should work:

use strict;
use warnings;
use Data::Dumper;

my %db_hash;

while (<DATA>) {
    chomp;
    my ($path, $start, $end, $delay) = split /\s+/, $_;

    # Get index for this path (if undefined, set it to 1)
    my $index = ++$db_hash{$path}{'index'};
    $db_hash{$path}{$index}{START} = $start;
    $db_hash{$path}{$index}{END}   = $end;
    $db_hash{$path}{$index}{DELAY} = $delay;
}

printf "Contents of db_hash => %s\n", Dumper(\%db_hash);

__DATA__
my_path    /path/3434/some_txt   wodo                -445
your_path  /path/3434/some_txt   wodo_modo           41
my_path2   /path/232/wge45       wodo_modo           240
my_path    /t43/annow23x         /a/b/c/43           NA

Which will yield:

Contents of db_hash => $VAR1 = {
          'my_path2' => {
                          '1' => {
                                   'START' => '/path/232/wge45',
                                   'DELAY' => '240',
                                   'END' => 'wodo_modo'
                                 },
                          'index' => 1
                        },
          'my_path' => {
                         '1' => {
                                  'START' => '/path/3434/some_txt',
                                  'DELAY' => '-445',
                                  'END' => 'wodo'
                                },
                         'index' => 2,
                         '2' => {
                                  'START' => '/t43/annow23x',
                                  'DELAY' => 'NA',
                                  'END' => '/a/b/c/43'
                                }
                       },
          'your_path' => {
                           '1' => {
                                    'START' => '/path/3434/some_txt',
                                    'DELAY' => '41',
                                    'END' => 'wodo_modo'
                                  },
                           'index' => 1
                         }
        };

Update:  Changed print to the printf it was supposed to be (which reveals a mistake I make all the time).  Thanks ysth, for catching that.

[utku]

Thank you very much folks, you have enlightened the points I could not figure out.

Both solutions (HoAoH by robotics and japhy; HoHoH by liverpole) worked.
HoHoH solution has the drawback that the interpreter generates an initialization error during initialization of indices.

Thanks again for your ultra-fast answers!

[ysth]

HoHoH solution has the drawback that the interpreter generates an initialization error during initialization of indices.

How so? Can you show us what you mean?

[roboticus]

The last sentence in your question provides the clue: You need an index, so rather than a HoH or HoHoH, what you really want is a HoAoH. The first H key being your path, the A being an array of your records, the last H being your attributes.

Roboticus