Help on multiline regex

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

Hi Monks, I need a little help here

How to parse this bit of multi-line data.

object-group service DM_SERVICE_7
 service-object tcp eq 1433 
 service-object tcp eq 49160 
 service-object tcp eq 8086 
object-group network Employees
 description Employees
 network-object 10.10.12.0 255.255.255.0
 network-object 10.11.12.0 255.255.255.0
[download]

Should parse into a hash like

'DM_SERVICE_7' => { 
    'type' => 'service ',
    'array_of_entries' => ["tcp eq 1433", "tcp eq 49160", "tcp eq 8086
+" ]
},
'Employees' => {
    'type' => 'network',
    'array_of_entries' => ["10.10.12.0 255.255.255.0", "10.11.12.0 255
+.255.255.0"]
}
[download]

I have got as far as here, but not sure how to proceed

undef $/;

while (<>) {
    @array = split /object-group/; #array of all interesting data
}
[download]

Comment on Help on multiline regex Select or Download Code

Replies are listed 'Best First'.
Re: Help on multiline regex by ikegami (Patriarch) on Jun 01, 2011 at 17:50 UTC
`my %objects; my $object; while (<>) { chomp; if ( my ($object_type, $object_id) = /^\sobject-group\s+(\S+)\s+(\S+)/ ) { $objects{$object_id} = $object = { type => $object_type, entries => [], }; } elsif (my ($entry) = /^\s\S+-object\s+(.*)/) { push @{ $object->{entries} }, $entry; } }` [download]	[reply] [d/l]
Re^2: Help on multiline regex by LanX (Saint) on Jun 01, 2011 at 18:05 UTC
better `elsif (my ($entry) = /^\s${object_type}-object\s+(.)/) {` Cheers Rolf	[reply] [d/l]
Re^3: Help on multiline regex by ikegami (Patriarch) on Jun 01, 2011 at 18:40 UTC
That should be `elsif (my ($entry) = /^\s\Q$object->{type}\E-object\s+(.)/) {` [download] but I fixed it with something simpler.	[reply] [d/l]
Re^4: Help on multiline regex by LanX (Saint) on Jun 01, 2011 at 18:50 UTC
Re^5: Help on multiline regex by ikegami (Patriarch) on Jun 01, 2011 at 19:17 UTC
Some notes below your chosen depth have not been shown here
Re: Help on multiline regex by wind (Priest) on Jun 01, 2011 at 20:38 UTC
use strict; use warnings; my %hash; my $key; while (<DATA>) { if (/object-group\s+(\w+)\s+(.)/) { $key = $2; $hash{$key}{type} = $1; } elsif (/\Q$hash{$key}{type}\E-object\s+(.)/) { push @{$hash{$key}{array_of_entries}}, $1; } } use Data::Dumper; print Dumper(\%hash); __DATA__ object-group service DM_SERVICE_7 service-object tcp eq 1433 service-object tcp eq 49160 service-object tcp eq 8086 object-group network Employees description Employees network-object 10.10.12.0 255.255.255.0 network-object 10.11.12.0 255.255.255.0 [download]	[reply] [d/l]
Re^2: Help on multiline regex by Anonymous Monk on Jun 02, 2011 at 00:31 UTC
Thanks for the complete code!	[reply]
Re: Help on multiline regex by LanX (Saint) on Jun 01, 2011 at 17:29 UTC
I don't think your mapping of input to output is unambiguous. I would loop over the input and try to match the head-lines and fieldelements of the object-groups. Then either call specialized subroutines which do the parsing of the following indented lines of the recognized group or raise a warning for an unknown group. Like this your data will also be evaluated for correctness. Sorry no code. Cheers Rolf	[reply]
Re: Help on multiline regex by Cristoforo (Curate) on Jun 02, 2011 at 00:56 UTC
Setting the INPUT_RECORD_SEPARATOR to 'object-group' #!/usr/bin/perl use strict; use warnings; use 5.012; use Data::Dumper; my %data; { local $/ = 'object-group'; while (<DATA>) { chomp; next unless $_; my ($type, $key) = /\A (\S+) (\S+)$/m; my @data = /^ $type-object (.+)$/mg; $data{ $key } = {type => $type, array_of_entries => \@data}; } } print Dumper \%data; __DATA__ object-group service DM_SERVICE_7 service-object tcp eq 1433 service-object tcp eq 49160 service-object tcp eq 8086 object-group network Employees description Employees network-object 10.10.12.0 255.255.255.0 network-object 10.11.12.0 255.255.255.0 [download] Update: edited line of code. `my ($type, $key) = /\A (\S+) (\S+)$/m;`	[reply] [d/l] [select]
Re^2: Help on multiline regex by Anonymous Monk on Jun 04, 2011 at 03:11 UTC
I like this method!	[reply]