Hash creation problem

[agentorange]

I'll say straight up I struggle with hashes and this is proving especially so with this challenge.

My _DATA_ is shown below - it's the output from a system command and the columns are space seperated. Pseudo hash structure shown below as to how I believe I need the output stored in order for me to make sense of the data in my code. Which is to be able to read the status for each subsystem/daemon and also sort all daemons by their status.

But I cannot get me head around how to construct the required hash from my _DATA_ can anyone assist please?

$VAR1 = {
          'inted' => {
                            'Group' => 'tcpip',
                            'PID' =>   '2424886',
                            'Status' => 'active'
                      }
      'xntpd' => {
                            'Group' => 'tcpip',
                            'PID' =>   '3473550',
                            'Status' => 'active'
                      }
       'rwhod' => {
                            'Group' => 'tcpip',
                            'PID' =>   '',
                            'Status' => 'inoperative'
                      }
        };


_DATA_
Subsystem         Group            PID          Status
 inetd            tcpip            2424886      active
 xntpd            tcpip            3473550      active
 rwhod            tcpip                         inoperative
 snmpd            tcpip                         inoperative
 aixmibd          tcpip                         inoperative
 hostmibd         tcpip                         inoperative
 snmpmibd         tcpip                         inoperative

[BrowserUk]

#! perl -slw
use strict;
use Data::Dump qw[ pp ];

my( undef, @keys ) = unpack 'A18 A17 A13 A*', <DATA>;
my %hash;
while( <DATA> ) {
    chomp;
    my( $id, @fields ) = unpack 'A18 A17 A13 A*';
    @{ $hash{ $id } }{ @keys } = @fields;
}

pp \%hash;

__DATA__
Subsystem         Group            PID          Status
 inetd            tcpip            2424886      active
 xntpd            tcpip            3473550      active
 rwhod            tcpip                         inoperative
 snmpd            tcpip                         inoperative
 aixmibd          tcpip                         inoperative
 hostmibd         tcpip                         inoperative
 snmpmibd         tcpip                         inoperative

Output:

C:\test>junk17
{
  " aixmibd"  => { Group => "tcpip", PID => "", Status => "inoperative
+" },
  " hostmibd" => { Group => "tcpip", PID => "", Status => "inoperative
+" },
  " inetd"    => { Group => "tcpip", PID => 2424886, Status => "active
+" },
  " rwhod"    => { Group => "tcpip", PID => "", Status => "inoperative
+" },
  " snmpd"    => { Group => "tcpip", PID => "", Status => "inoperative
+" },
  " snmpmibd" => { Group => "tcpip", PID => "", Status => "inoperative
+" },
  " xntpd"    => { Group => "tcpip", PID => 3473550, Status => "active
+" },
}

[hippo]

Your structure there is a hash of hashes (HoH), see perldsc for details of data structures like this. To construct it:

1. Read first line of data, split on whitespace and store in an array. These are your field names.
2. Read next line of data and split on whitespace. First column is outer key, columns 2 .. n are the values to be associated with the equivalent key names in the array from step 1 above.
3. Repeat step 2 until eof.

Try that and follow-up here with your code if you get stuck. Note that arrays start from zero in Perl, so the second column will actually be $array[1] in your code, etc. HTH.

Update: Splitting on whitespace could give you problems with the missing fields. Unless you have well defined field delimiters, follow BrowserUK's fixed-column approach above when reading instead.

[kevbot]

Others have given you good suggestions already. I just wanted to show that you could use the Text::xSV::Slurp module to create the HoH structure for you. The whitespace separated data would be a problem for this module, so you would need your input data to be in a delimited format for this approach to work (like csv shown below). BrowserUK has shown how to deal with the whitespace separated data that you have. I would encourage you to become very familiar with how to use hashes in perl before resorting to a module to do the work for you, since they are such an important feature of perl.

This code:

#!/usr/bin/env perl

use strict;
use warnings;

use Text::xSV::Slurp;

my $hoh = xsv_slurp(
    \*DATA,
    shape    => 'hoh',
    key      => 'Subsystem',
);

use Data::Dumper;

print Dumper $hoh;

exit;

__DATA__
Subsystem,Group,PID,Status
inetd,tcpip,2424886,active
xntpd,tcpip,3473550,active
rwhod,tcpip,,inoperative
snmpd,tcpip,,inoperative
aixmibd,tcpip,,inoperative
hostmibd,tcpip,,inoperative
snmpmibd,tcpip,,inoperative

Gives the output:

$VAR1 = {
          'inetd' => {
                       'Group' => 'tcpip',
                       'Status' => 'active',
                       'PID' => '2424886'
                     },
          'hostmibd' => {
                          'Group' => 'tcpip',
                          'Status' => 'inoperative',
                          'PID' => ''
                        },
          'snmpmibd' => {
                          'Status' => 'inoperative',
                          'PID' => '',
                          'Group' => 'tcpip'
                        },
          'aixmibd' => {
                         'PID' => '',
                         'Status' => 'inoperative',
                         'Group' => 'tcpip'
                       },
          'snmpd' => {
                       'Group' => 'tcpip',
                       'Status' => 'inoperative',
                       'PID' => ''
                     },
          'xntpd' => {
                       'Group' => 'tcpip',
                       'Status' => 'active',
                       'PID' => '3473550'
                     },
          'rwhod' => {
                       'Group' => 'tcpip',
                       'Status' => 'inoperative',
                       'PID' => ''
                     }
        };

[BrowserUk]

I would encourage you to now write the program that will take the OP's data in the form he has it and then convert it to the form that your "solution" requires it to be in to work.

And then see how much time you saved using a module that solved a different problem.

[kevbot]

I just wanted to show an alternative. I think the solution you posted is the most useful for the specific question at hand. Perhaps the OP has some control over the format of the input data. What if they had the means to quickly obtain the input data in delimited format (e.g. a command line switch)? Then maybe using Text::xSV::Slurp makes sense. Also, the module has additional features that may prove useful for the OP in the future.

[sundialsvc4]

There is a Perl feature that is specifically designed to deal with this sort of thing:   it goes by the somewhat fancy-pants moniker, auto-vivification, which is “to spring to life automatically.”   Consider these two statements:

 my $foo = undef;
$foo->{'bar'}{'bletch'}{'sneeze'} ++;

A lot of automatic voodoo “just happens” here.   $foo immediately becomes a hashref; it acquires the key 'bar' which also becomes a hashref, which itself now has the key 'bletch'; ditto 'sneeze'; and this element automatically becomes an integer, automatically zero, which is incremented to 1.

Therefore, in Perl it requires very little code at all in order to construct arbitrarily-complex reference based data structures in memory.   Or, thanks to a powerful reference-count based memory manager, to maintain those structures without leaks.   There are many aspects of Perl that are “what all the fuss is about,” and this is certainly one of them.

[agentorange]

Unfortunately it's this voodoo that I just have failed to get my head around.

BrowserUK I really appreciate that and having read up on unpack can understand it too. hippo - thank you that helps.