http://qs321.pair.com?node_id=11100503


in reply to how to avoid full scan in file.

I tried to get some of what you need but didn't get the use of $i, $j, $k or $modified.

This isn't a complete solution, but it uses the suggestion by LanX to create a hash of the smaller B file and loop through the A file just once. This should speed up your program considerably.

I used 2 pseudo files to stand in for your real files. I hope this gives you some direction for your problem.

#!/usr/bin/perl use strict; use warnings; $|=1; my $fileA =<<EOF; l100101,aaaaaaa,a_0100,loc,10,1 l100101,aaaaaaa,a_0100,loc,11,1 l100101,aaaaaaa,a_0100,loc,12,6 EOF my $fileB =<<EOF; l103709,bbbbbbb,c_0200,929 l100109,bbbbbbb,b_0100,442 l100107,bbbbbbb,c_0300,389 EOF my $filea = $ARGV[0]; my $fileb = $ARGV[1]; my $FileC = "result.csv"; open ( FA, '<', \$fileA) || die ( "File $filea Not Found!" ); open ( FB, '<', \$fileB) || die ( "File $fileb Not Found!" ); #open ( FC, ">", $FileC) || die ( "File $FileC Not Found!" ); my %B; while ( <FB> ) { chomp; my($look, $sec, $cls, $max) = split ","; $B{"$look,$sec,$cls"} = $max; } my @A; while ( <FA> ) { chomp; my($look, $sec, $cls, $att, $idx, $qtd) = split ","; my $keyA = "$look,$sec,$cls"; if (exists $B{$keyA}) { my $max = $B{$keyA}; my $tot = $qtd - 1; if ($tot >= 0) { print join(",", $look, $sec, $cls, $att, $idx, $max), "\n +"; } } }