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

Dear all,
The solution proposed in posting 7633 (How do I extract all text between two keywords like start and end?) is not working for me. I have tried it and failed. (probably because I'm a perl newbie) A try with grep failed also (the "s" switch didn't work). Neither do I find other good examples on the net. Can somebody solve the problem?
My problem is the following:
  1. a 9 MB text file
  2. Multiple parts that must be extracted between "labels"
  3. The parts are not nested but successive
  4. The substracted parts must be in a CSV form
  5. It would be nice to have the label names in the csv
e.g. 

start\n
blabla\n
...
blabla\n </i>
begin:\n
usefull information1\n
usefull information2\n
usefull information3\n
end:\n
blabla\n
...
blabla\n
\n
start
blabla\n
...
blabla\n
begin:\n
usefull information1\n
end:\n
blabla\n
etc etc

[download]

I hope you can help. Thanks in advance.
Firewall

edited by ybiC: my best effort at intended formatting, with <code> instead of unbalanced & unopened </p>s

  • Comment on Selecting text between given words (multi-lines (/n) and multiple occurrences)
  • Download Code

Replies are listed 'Best First'.
Re: Selecting text between given words (multi-lines (/n) and multiple occurrences)
by Paladin (Vicar) on Aug 25, 2003 at 17:29 UTC
    Sounds like a good use for the .. operator in scalar context. 
    open FH, "file.txt" or die "Couldn't open file.txt: $!";
while (<FH>) {
    if (/^begin:$/ .. /^end:$/) {
        print;  # or do whatever else you want here.
    }
}

    [download]
[reply]
[d/l]
[select]
      First things first: Paladin thx :-)

      How do I get the actual data between begin & end in an output file? When I try, the file stays empty. 
      #-----------------------------------------------------------
#!/usr/bin/perl (This code does not work !!!)
#
# between.pl <output file><input file><begin word><end word>
#

open (FH_output, ">> $ARGV[0]") or die "Couldn't open file.txt: $!";
open FH_input, "< $ARGV[1]" or die "Couldn't open file.txt: $!";
while (<FH_input>) {
    if (/^$ARGV[2]$/ .. /^$ARGV[3]$/) {
        $info = $_ ;
        print FH_output  $info;  
    }
}
close (FH_input) ;
close (FH_output) ;

      [download]
      Thanks,

      Firewall (Perl is fun. I'm going to study it some more. But still a long way to go :-)
[reply]
[d/l]
Re: Selecting text between given words (multi-lines (/n) and multiple occurrences)
by zentara (Cardinal) on Aug 26, 2003 at 15:57 UTC
    Maybe you could use a flag to print in-between lines. Since it's a 9 meg file you may not want to slurp the whole file in, so detecting multi-line matches may be troublesome. 
    #!/usr/bin/perl
$goprint=0;
     while (<>){  
  if ($_ =~ /^Start(.*)/){$goprint=1}
  if ($_ =~ /^END(.*)/){$goprint=0;next}
print "$_" if $goprint == 1;
}

    [download]
    or if you can slurp 
    #Here's code that finds everything  
#between START and END in a paragraph: 
undef $/;  # read in whole file, not just one line  
while ( <> ) {
while (/START(.*?)END/sgm) {
# /s makes . cross line boundaries 
print "$1\n";
}
}

    [download]
[reply]
[d/l]
[select]
      Zentare, Thanks for the help. I created a little script that works fine with your code but if I enters a begin/end-string with spaces (e.g "begin log1", "end log1") it takes the two first words (e.g begin & log1) as the begin and the end words. The use of quotation marks does not work. Thx, Firewall
[reply]
        Solved the problem :-) I just used getopt with "". That works fine.

        Thx everyone,

        Nylon
[reply]

Back to Seekers of Perl Wisdom