filtering text in this manner, is it possible ?

[javanoob]

hi all,

i have a text file which goes like this..


text1..
text2..
text3...
text4..
session1
text 5..
.

text1..
text2..
session2
text5...
.

text1
text2
text3
session1
text4..

-------------------

1) each paragraph is seperated by a . follow by a linefeed
2) each paragraph of text belong to a session
3) a session might span 2 different paragraph (e.g. session 1)
4) the amount of line(text1,2,3) etc is variable for each paragraph (sometime 10 lines, sometime 4lines)

================

how do i grep only paragraph belonging to a particular session ?

e.g. (you can see paragraph belonging to session 2 is omitted)

text1..
text2..
text3...
text4..
session1
text 5..
.

text1
text2
text3
session1
text4..
.

========

Regards,
Noob

[atreyu]

You could certainly do this with awk and grep, but here's a perl script that I think will do it:

#!/usr/bin/perl
use strict;
use warnings;
my $file = shift || die "Give me a file\n";
my $cnt = 0;
my %hash = ();
open(FH,'<',$file) or die "can't read '$file': $!\n";
while(<FH>){
  chomp;
  if(/./ .. /^\.$/){
    push(@{$hash{$cnt}},$_) unless(/^\.$/);
  }else{
    $cnt+=1;
  }
}
close(FH);

my %sessions = ();
for my $par(sort { $a <=> $b } keys %hash){
  my $session;
  my @array = ();
  for(@{$hash{$par}}){
    if(/^session([0-9]*)$/){
      $session = $1;
    }else{
      push(@array,$_);
    }
  }
  die "No session defined for paragraph $par\n" unless(defined($session));
  for(@array){
    push(@{$sessions{$session}},$_);
  }
}

for my $session(sort { $a <=> $b } keys %sessions){
  print "Session $session\n";
  my @lines = @{$sessions{$session}};
  for(@lines){
    print "  ",$_,"\n";
  }
}

Create a file with this code in it (call it "read-file.pl" or whatever), make it executable ("chmod +x ./read-file.pl") and pass your text file as an argument to it:

./read-file.pl text.txt