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

I am rewriting the split utility in perl, mostly just to do it (plagiarizing a version in awk in the gawk manual).It requires creating new files to write to every thousand lines (or so) of input.

I have read the relevant passages in perl cookbook and learning perl, but I am unsure how to deal with the anonymous fileglob produced from an indirect filehandle. How do I close the file?

The code is unfinished, but you can see where I am unsure at the end of it:

# split-unix.pl -- do split in perl
#
# usage:split [-num] file [outname]

my $outfile = "x";  #default
my $limit = 1000;
my $readfile;
my $s1= "a";

if (scalar(@ARGV) > 3){
  usage;
}
if($ARGV[0] =~/^[0-9]+/){
  $limit = -$ARGV[0];
  shift @ARGV;
}
$readfile = $ARGV[0];
shift @ARGV;
if (defined $ARGV[0]){
  $outfile = $ARG[0];
}
$outfile.=$s1;
  
open ($write, ">", $outfile) or die "$0: can't open $outfile $!";  
open ($fh, "<", $readfile) or die "$0: can't open $readfile $!";
while(<>){
  if ($. < $limit){ 
    print;
  }
  else{
    $outfile++;
    $write=""; # I think this answers my question
    open ($write, ">", $outfile) or die "$0: can't open $outfile $!";
  }
}

[download]
cleaned up the syntax errors, but now problem is error:"can't open -100 No file or directory ..."

Replies are listed 'Best First'.
Re: dynamic filenames, indirect filehandles
by rupesh (Hermit) on May 19, 2005 at 04:07 UTC
    IMHO, I'm not sure what you're trying to achieve here, but here are a few pointers:

    • Always  use strict;
    • Why confuse yourself with the ARGuments? Why not use $ARGV[0], $ARGV[1]..etc?
    • open FH, '<', '<path of file>';
    • Why are you appending a file handle ($outfile)with  $sl ?
    • Indent


    Update:Use split;

    Cheers,
    Rupesh.
[reply]
[d/l]
[select]
      Thanks rupesh
      I don't mean to append the filehandle, I mean to append the name ("path") of the file. Incrementing the name is just what I want to achieve. This is the core of my question: For a 20,000 line $readfile, I want to produce 20 separate $outfiles, files, named (for example) "outa", "outb", etc.
[reply]
      Why are you appending a file handle ($outfile)with $sl ?

      So that he can say $outfile++ and generate unique file names for each split file.

      Lou

[reply]
Re: dynamic filenames, indirect filehandles
by jdporter (Paladin) on May 19, 2005 at 04:14 UTC 
    if($ARGV[0] =~/^[0-9]+/){

    [download]
    I think that should have a dash in it: 
    if($ARGV[0] =~/^-[0-9]+/){

    [download]
[reply]
[d/l]
[select]
      I tried that; my program print the error:
      C:\scripts>perl split-unix.pl -100 tinysat tmp Can't open -100: No such file or directory at split-unix.pl line 27.
      <-- contents of the file-->
      Can't open tmp: No such file or directory at split-unix.pl line 28, <> line 16.
      It also writes an empty file "tmpa".
[reply]
        The problem is solved; thanks for your patience with my ill-formed question that turned out not to be the problem at all. 
        # split-unix.pl -- do split in perl
#
# usage:split [-num] file [outname]

use strict;

my $outfile = "x";  #default
my $limit = 1000;
my $readfile;
my $s1= "a";

if($ARGV[0] =~/^-?[0-9]+/){
  $limit = 0- $ARGV[0];
  shift @ARGV;
}
$readfile = $ARGV[0];
shift @ARGV;
if (defined $ARGV[0]){
  $outfile = $ARGV[0];
}
$outfile.=$s1;
  
open (OUT, ">", $outfile) or die "$0: can't open $outfile $!";  
open (IN, "<", $readfile) or die "$0: can't open $readfile $!";
print OUT $limit;

#=pod
while(<IN>){
  if ($. > $limit ){
    close OUT;
    $outfile++;
    open(OUT, ">", $outfile) or die "$0: can't open $outfile $!";
  }
  print OUT "$_";
} 

#=cut
close OUT;

        [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom