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

Hi folks,

I have a file which has the following format

245
record #1
record #2
.
.
record #245
====
165
record #1
record #2
.
.
record #165
====

As you can make out, the first line is the number or records in that set (till "===="). There are 20 such records. Instead of regexing, I used the following 

#initialize some variables
my $records; # No. of records
my $last;  # The last line
my $ini = 1; # set initial line
my $count = 1; # counter


for (my $count = 1; $count <= 20; $count++){
open (OLD, "< $old_file.txt") || die $!; # open file
# Now we need to split up the file into 20 files

open (NEW, "> $old_file$count.txt") || die $!; #open outputfile

while (<OLD>){
    if($. == $ini){
        chomp;
        $records = $_;
    }
}
close OLD;

$last = $ini + $records + 1;

open (OLD, "<$old_file.txt") || die $!; # I am more comfortable with r
+e-opening the file like this
while (<OLD>){
    next unless $. >= $ini;
    print NEW $_;
    last if $. >= $last;
}
close OLD;
close NEW;
$ini = $last + 1;
}

[download]

What I wanted to know was, that in a case like this, would it be better practise to use a regex (since there are well-defined patterns), or use, what to me at least, is a simpler approach.

mndoci

"What you do in this world is a matter of no consequence. The question is, what can you make people believe that you have done?"-Sherlock Holmes in 'A study in scarlet'

Replies are listed 'Best First'.
(jeffa) Re: To regex or not to regex
by jeffa (Bishop) on Mar 11, 2002 at 05:35 UTC
    Thanks for sharing. Here is what i came up with. Instead of setting a limit at 20 records, this version will keep adding new files until the input file is exhausted.

    I did use one regex to read the line that tells how many records to expect. I also took the liberty of changing the names of the output files with sprintf and a different prefix than the original file (just so i could delete my mistakes easier :D). 

    use strict;

open(IN,'old_file.txt') or die;

my $i = 1;
while (<IN>) {
   if (/^(\d+)\s*$/) {
      open(OUT,'>new_file'.sprintf("%02d",$i++).'.txt') or die;
      do { $_ = <IN>; print OUT; } for (1..$1);
   }
}

    [download]

    jeffa

    L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)
[reply]
[d/l]
      Here is a more newbie-friendly version of jeffa's code. Consider it a learning tool. 
      my $file_num = 1;
while (<IN>) {
   if (/^(\d+)\s*$/) {
      my $count = $1;
      my $fname = sprintf("new_file%02d.txt", $file_num);
      open(OUT,">$fname") or die "Cannot open file $fname: $!\n";
      for(1..$count){
         my $rec = <IN>; 
         print OUT $rec;
      }
      close OUT; 
      $file_num++;
   } else {
     # ignore the file separator =======
   }
}

      [download]
[reply]
[d/l]

      Thank you very much. Much more compact than my blurb. There is one thing missing though, and I am sure I can fix that (I have not tried) without too much fuss. I want to include the first line (number of records), and the last line (====) as well. Otherwise, works like a charm.

      mndoci

      "What you do in this world is a matter of no consequence. The question is, what can you make people believe that you have done?"-Sherlock Holmes in 'A study in scarlet'

[reply]
Re: To regex or not to regex
by Anonymous Monk on Mar 11, 2002 at 06:00 UTC 
    {local $/="====\n";
while( <> ){
  open (NEW, "> $old_file$..txt") || die $!; #open output
  chomp;
  print NEW;
}
}

    [download]
[reply]
[d/l]
      This code is slightly broken because it ignores a requirement that was specified by the problem (every section starts with a line containing a number which is the number of records) and relies on an assumption instead (every section is terminated by ==== followed by "\n"). There are at least three cases where this would break, and they do happen more often than not.
      • When there is a record ending with "====", this code breaks the section into more files than it should.
      • When there is a space after the "====", then it won't match the input record separator, thus 2 sections will be merged in one file. This space would not be visible to the eye so debugging this problem is not easy.
      • When there is an empty line after the last "====", you will have an additional empty file.
      So, the wisdom behind this story: follow the specs carefully and don't golf when you don't need to.

      Hope this helps,,,

      Aziz,,,

[reply]
        Just realized that your post was targeted at Anonymous' post and not mine. DUH!!!!
        That said, I almost used '====' as the separator, but my first test went into a weird loop, so I did not continue along that path.

        mndoci

        "What you do in this world is a matter of no consequence. The question is, what can you make people believe that you have done?"-Sherlock Holmes in 'A study in scarlet'

[reply]
        Yes it does rely on that assumption. I did have lines in my code that removed leading whitespace, but I didn't think about trailing whitespace. As for records between the # and '====', they will nto have any '====', so I guess I am safe, but I agree, my solution is not general. Thanks for the warning!!

        mndoci

        "What you do in this world is a matter of no consequence. The question is, what can you make people believe that you have done?"-Sherlock Holmes in 'A study in scarlet'

[reply]
Re: To regex or not to regex
by jryan (Vicar) on Mar 11, 2002 at 18:32 UTC

    Algorithmically golfed: (golfed without employing space-saving techniques, otherwise it would look scary :P) 

    open (OLD, "<$old_file.txt") || die $!;
my $i;
foreach (split (/====\s*\n(?=\d+)/, do{local$/;<OLD>}))
{
    open (NEW,">$old_file$i.txt") || die $!;
    print NEW $_;
    close (NEW);
    $i++;
}

    [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom