perlquestion
PrimeLord
Monks I come to you seeking your wisdom once again. I am having a bit of a memory management issue I was hoping you could help me with. I have written a script that produces a daily report on a system scan it does. It first reads in a benchmark file from the scan the day before into a hash. It then runs the daily scan dumping the info it finds into the new benchmark file and also compares the information to the data from the day before in the hash.
<BR><BR>
The problem I am running into is the data being read into the hash can be several hunder megabytes in size. I need to find a more efficent way to handle this data. Here is an example of the code I have written.
<BR>
<CODE>use strict;
sub _read_benchmark {
my %yesterday;
open BENCH, "benchmark_file" or die "$!";
while (<BENCH>) {
chomp;
$yesterday{$_}++;
}
close BENCH or warn "$!";
return \%yesterday;
}
sub _scan_system {
my $yesterday = shift;
my %today;
open BENCH, "> benchmark_file" or die "$!";
open IN, "find / $search_files -print |" or die "$!";
while (<IN>) {
chomp;
print BENCH "$_\n";
if (exists $yesterday->{$_}) {
delete $yesterday->{$_};
} else {
$today{$_}++;
}
}
close IN or warn "$!";
close BENCH or warn "$!";
return \%today, $yesterday;
}
sub _print_report {
...
}
my $yesterday = _read_benchmark;
my ($today, $yesterday) = _scan_system($yesterday);
_print_report($today, $yesterday);</CODE>
I believe there is a way I can tie the benchmark file to a hash and that would probably be a huge improvement, but I am not sure how to do that. Any suggestions on how to make this less of a memory hog would be very appreciated. Thanks!
<BR><BR>
-Prime