Let's suppose for a moment that you wanted to split the data at the lines containing "INPUT SEQUENCE=" and use the number following the = as part of the file name for the output file, then you could:

use strict;
use warnings;

$/ = 'INPUT SEQUENCE=';

while (<DATA>) {
    chomp;
    next unless length;
    
    next unless s/^(\d+)\n//;
    print "Start of file seq$1.dat\n";
    print "$_\n";
}


__DATA__
[download]

Prints:

Start of file seq6618.dat
>P40757|ALN_RANCA Allantoinase, mitochondrial precursor - Rana catesbe
+iana (Bull frog)
MALKSKPGIMNITPGSKISVIRSKRVIQANTISSCDIIISDGKISSVLAWGKHVTSGAKLLDVGDLVVMA
GIIDPHVHVNEPGRTDWEGYRTATLAAAAGGITAIVDMPLNSLPPTTSVTNFHTKLQAAKRQCYVDVAFW
GGVIPDNQVELIPMLQAGVAGFKCFLINSGVPEFPHVSVTDLHTAMSELQGTNSVLLFHAELEIAKPAPE
IGDSTLYQTFLDSRPDDMEIAAVQLVADLCQQYKVRCHIVHLSSAQSLTIIRKAKEAGAPLTVETTHHYL
SLSSEHIPPGATYFKCCPPVRGHRNKEALWNALLQGHIDMVVSDHSPCTPDLKLLKEGDYMKAWGGISSL
QFGLPLFWTSARTRGFSLTDVSQLLSSNTAKLCGLGIVKEPLKWVMMLIWSSGILTKSFRCKKMIFITRI
SSPHIWDSFFKEKSWLLLFEGLLFISKGSMLPNQLENLFLYTLWSLVKPVHPVHPIIRKNLPHI
                                                                      
+                                                                     
+                  
Total Number of residues in the sequence =484
----------------------------------------------------------------------
+----------­-------------------------
Number of residues in the repeat = 5
AAAAG                                                               96
+ to 100
AAAGG                                                               97
+ to 101
----------------------------------------------------------------------
+----------­-------------------------
Number of residues in the repeat = 7
AGVAGFK                                                            157
+ to 163
GGISSLQ                                                            345
+ to 351
----------------------------------------------------------------------
+----------­-------------------------
Minimum number of amino-acids present in the distant repeat is  = 5
Maximum number of amino-acids present in the distant repeat is  = 7
Total number of distant repeats found = 62
______________________________________________________________________
+__________­_________________________

Start of file seq6619.dat
>Q9RKU5|ALN_STRCO Probable allantoinase - Streptomyces coelicolor
MSEAELVLRSTRVITPEGTRAASVAVTGEKITAVLPYDAPVPAGARLEDVGDHVVLPGLVDTHVHVNDPG
RTEWEGFWTATRAAAAGGITTLVDMPLNSIPPTTTVDNLRTKREVAADKAHIDVGFWGGALPDNVKDLRP
LHEAGVFGFKAFLSPSGVDEFPHLDQEQLARSLAEIAAFDGLLIVHAEDPHHLAAAPQQGGPKYTHFLAS
RPRDAEDTAIATLLAQAKRFNARVHVLHLSSSDALPLIAEARADGVRVTVETCPHYLTLTAEEVPDGASE
FKCCPPIREAANQDLLWQALADGTIDCVVTDHSPSTADLKTDDFATAWGGIAGLQLSLPAMWTAARGRGL
GLEDVVRWMSERTAALVGLDARKGAIAPGHDADFAVLAPDETFTVDPAALQHRNRVTAYAGKTLYGVVKS
TWLRGERIVADGAFTDPKGQLLDRA
                                                                      
+                                                                     
+                  
Total Number of residues in the sequence =445
----------------------------------------------------------------------
+----------­-------------------------
Number of residues in the repeat = 5
AAAAG                                                               83
+ to 87
AAAGG                                                               84
+ to 88
PSTAD                                                              314
+ to 318
----------------------------------------------------------------------
+----------­-------------------------
Number of residues in the repeat = 6
AAAGGI                                                              84
+ to 89
SSSDAL                                                             240
+ to 245
----------------------------------------------------------------------
+----------­-------------------------
Minimum number of amino-acids present in the distant repeat is  = 5
Maximum number of amino-acids present in the distant repeat is  = 8
Total number of distant repeats found = 94
______________________________________________________________________
+__________­_________________________
[download]

which may or may not be anything at all like what you had in mind, but then you haven't actually told us that so guessing is all we can do.

There Is More Than One Way To Do It - one bizarre way is

#!/usr/bin/perl
{
  local $/ = "INPUT SEQUENCE=";
  <>; # discard first chunk
  while(<>) {
    chomp;
    open O, '>', /^(\d+)/ and $1 or die "$!\n";
    print O $/,$_;
  }
}
[download]

update - oh. You want bunches of 1000 records in separate files?

#!/usr/bin/perl
{
  my $file = 'File00000';
  local $/ = "INPUT SEQUENCE=";
  <>;
  while(<>) {
    chomp;
    unless ($.-2 % 1000) {
        open O, '>', ++$file or die "cant write '$file': $!\n";
    }
    print O $/,$_;
  }
}
[download]

update - a bit of explanation: setting the input record separator $/ (see perlvar) to the token right after the file boundary lets the diamond operator <> (or readline) read up to and including that token as a single line into $_.

With chomp we remove $/ from the end; it is added up front when outputting. The ++$file is a string increment; doing that we get the next file name (File00001, File00002, ... ).

Since the first "line" (consisting of the record separator only) isn't interesting, we do a <> before the loop. Next line is then number 2 ($. - see perlvar), and $.-2 % 1000 (modulo 1000) is 0 at line 2, 1002, 2002, ... , so we (re-)open the output filehandle at that line count, which does an implicit close. See open.

_($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                              /\_¯/(q    /
----------------------------  \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}

If I understand correctly, you have a file with a number of records, each one beginning with a line:

INPUT SEQUENCE=XXX
[download]

and ending with:

______________________________________________________________________
+___________________________________
[download]

And you want to split the whole file in smaller files containing N or these records

The easiest way I can imagine doing this is using Tie::File. Try this script:

#!/usr/bin/perl

use strict;
use warnings;
use Tie::File;

my ($file,$recs_X_file) = @ARGV;
die "Usage: $0 <input_file> <recs x file>" if (@ARGV != 2);
tie my @arr, 'Tie::File', $file, recsep => "__________________________
+_____________________________________________________________________
+__________",autochomp=>0;

my $from=0;
my $to=$recs_X_file-1;
while ($from < $#arr){
  my $ofile = "file.$from-$to";
  open F,">",$ofile or die $!;
  print "printing records $from to $to in $ofile\n";
  print F @arr[$from..$to];
  $from=$to+1;
  $to = $from+$recs_X_file-1;
}
[download]

This script interfaces the file as an array, but in the way that you want it to do: Each record in the array corresponds with one logical record in the file. Once done, it splits the array (i.e. the records in the file) N by N records and outputs them in sub-files

For example, if you call the script "split_records.pl" you can invoke it with:

perl split_records.pl inputfile 10
[download]

Outputs:

printing records 0 to 9 in file.0-9
printing records 10 to 19 in file.10-19
printing records 20 to 29 in file.20-29
... and so on (depending on the number of records of the original file
[download]

Of course, the files containing the records are created too

Hope this helps!

citromatik

It is not quite clear to me what you want to do. Try to explain your algorithm that decides what should go into each file. There is helpful advice here on how to How (Not) To Ask A Question.

my $suffix = 'z';
my $sequence = 0;
my $maxseq = 1000;

my $input = shift @ARGV or die "usage";
open my $ifh, $input or die "$!: $input\n";
my $ofh;
while( <$ifh> ) {
    /\AINPUT\sSEQUENCE/
      and SwitchFile( $input, \$ofh, \$suffix, \$sequence, $maxseq );
    $ofh or die "Unexpected prelude: $_";
    print $ofh $_;
}
close $ofh;

sub SwitchFile {
    my ( $input, $oref, $sref, $qref, $max ) = @_;
    if ( defined( $$oref ) )
        ( ++$$qref < $max ) and return;
        $$qref = 0;
        close $$oref;
    }

    my $newfile = "$input." . ++$$sref;
    open my $ofh, ">$newfile" or die "$!: $newfile";
    $$oref = $ofh;
}
[download]