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

Dear Monks

In my attempts in trying to improve my understanding of Regex, I have the following data from a HTML dump; 
1  adriaanf Europe Local    _Default Different Owner For Target Machin
+e HA050069  OYWVM1237  LN-CS/06 Technology XP DESKTOP Dell OptiPlex G
+X270 (0151) Pentium 4 (1 x 2793) 1023 38146 11147 N/A  N/A
 OYWVM1237  Technology LN-OY/02 VIRTUAL (OYWVH161) VMWare VMWare For D
+esktop Not Defined (1 x 3065) 767 20473 8209 23/11/2004 11:10:02  N/A
+  N/A
 N/A N/A 
2  adriaanf Europe Local    _Default Different Owner For Target Machin
+e HA050069  OYWVM1262  LN-CS/06 Technology XP DESKTOP Dell OptiPlex G
+X270 (0151) Pentium 4 (1 x 2793) 1023 38146 11147 N/A  N/A
 OYWVM1262  Technology LN-OY/06 VIRTUAL (OYWVH159) VMWare VMWare For D
+esktop Not Defined (1 x 3064) 767 20473 7800 07/12/2004 10:50:32  N/A
+  N/A
 N/A N/A 
5  adrianst Europe Local    ER_LN_WAR Different Owner For Target Machi
+ne CW041698  OYWVM1263  LN-CW/04 Research XP DESKTOP Compaq Evo D510 
+(07E8h) Small Form Factor Pentium 4 (1 x 2259) 511 38154 10740 N/A  N
+/A
 OYWVM1263  Technology LN-OY/02 VIRTUAL (OYWVH138) VMWare VMWare For D
+esktop Not Defined (1 x 3065) 767 20473 7788 06/12/2004 18:24:34  N/A
+  N/A
 N/A N/A 
6  adrianst Europe Local    ER_LN_WAR Different Owner For Target Machi
+ne CW041698  OYWVM1230  LN-CW/04 Research XP DESKTOP Compaq Evo D510 
+(07E8h) Small Form Factor Pentium 4 (1 x 2259) 511 38154 10740 N/A  N
+/A
 OYWVM1230  Technology LN-OY/06 VIRTUAL (OYWVH133) VMWare VMWare For D
+esktop Not Defined (1 x 3065) 767 20473 6921 06/12/2004 17:48:37  N/A
+  N/A
 N/A N/A

[download]
From that data, I need to grab it in a record form. Each record starts with a record number 1,2,3,...50000..etc

I am not sure on how to do it, so I started with this code; 
#! c:/perl/bin/perl.exe -slw
$|++;
use strict;
use vars qw/%data/;

open (LST, "$ARGV[0]") or die "\n$0 Error => $^E\n";
chomp (my @unclean = <LST>);
print"size : $#unclean";
for (@unclean)
{
    if ($_ =~ /^\d+\s+/)
    {
        print "First 1 : $_\n";
        #print "Line 2
        #print "Line 3
        #print "Line 4
        print "____________________________________\n";
    }    
}

[download]
It does grab the first line of that data, but I am not sure on how to code the regex so that it grabs all lines until the next record number, where start of a new record begins.

I have spent some time trying all sorts of different combinations of regex to no avail. I would appreciate if any of you divine beings can inspire and guide me through this.

Thanks

UPDATE : The 'n/a's are empty field values reserved for dates. its not always that records end with n/a.
Blackadder

Replies are listed 'Best First'.
Re: Capturing Multiple lined data with regex.
by duff (Parson) on Dec 08, 2004 at 15:25 UTC

    I'm not so sure you should be doing all of the work inside of a regular expression. Here are two ways to do what you're after:

    Example #1 

    open(my $fh, "<", $ARGV[0]) or die "\n$0 Error => $^E\n";
my $data = do { local $/; <$fh> };
close $fh;
my @records = split /(?=\n\d+\s+)/, $data;
# Now each item of @records has the lines you're looking for

    [download]
    Example #2 
    open(my $fh, "<", $ARGV[0]) or die "\n$0 Error => $^E\n";
my (@records,$rec);
while (<$fh>) {
   unless (/^\d+\s+/) {  $rec .= $_; next }
   push @records, $rec if $rec; $rec = $_
}
push @records, $rec if $rec;
close $fh;
# Now each item of @records has the lines you're looking for

    [download]

    Can you guess which example I like best? :-)

    You can use various print statements if you don't want to populate an array of course.

    duff

[reply]
[d/l]
[select]
      Kool stuff,...Thanks a lot.

      But let me guess, Example #1. right?

      :-)
      Blackadder
[reply]
Re: Capturing Multiple lined data with regex.
by conrad (Beadle) on Dec 08, 2004 at 15:30 UTC

    The best thing would be to set $/ to some kind of record separator such that your my @unclean = <LST> actually sucks in one record per array entry rather than one line ($/ is the input record separator, usually "\n", so <LST> reads in a line at a time). However, without more information it's difficult to tell what a good value would be.

    It looks as if every record is separated by a line consisting of "N/A N/A\n"; if that's true, then try setting $/ = "N/A N/A\n" ("\nN/A N/A\n" might be safer — matches a line containing exactly N/A N/A, not simply ending in N/A N/A) and taking a look at what you get in @unclean (more info in perlvar docs). If that works then you might also want to investigate using /s and/or /m as modifiers on your regexp (these modify how regexps process newlines - more info in perlop docs).

    HTH…

[reply]
[d/l]
[select]
      It all came flooding back to me! All them hours of studying! Reading this and this.

      Many thanks indeed
      Blackadder
[reply]

Back to Seekers of Perl Wisdom