Perl - Memory Issues

[James B.]

I have written a perl script to parse some large flat-text logs. The
logs are bzipped and come in through a pipe to STDIN. The script then
performs some regular expressions to the incoming data then prints to
STDOUT.

The script works great, but the issue I have is that the script uses
as much memory as the data coming into it. Therefore if I pipe a 800MB
file into it, the memory usage grows and grows until it reaches
approximatly 800MB, and it doesn't appear to release any of the memory
until it is completely finished. Is there a way to have Perl not use
so much of the system memory? Here is a sample of what I am doing.

#!/usr/bin/perl -w

use strict;

if ($ARGV[0]){
open STDIN, "bzcat $ARGV[0]|" or die "Cant uncompress file as a
pipe\n$!\n";
}

foreach (<STDIN>) {
chomp;
if ($_ =~ /(somedata)/) {
print "$1\n";
}
}

[Jürgen Exner]

James B. wrote:

The script works great, but the issue I have is that the script uses
as much memory as the data coming into it. Therefore if I pipe a 800MB
file into it, the memory usage grows and grows until it reaches
approximatly 800MB, and it doesn't appear to release any of the memory
until it is completely finished. Is there a way to have Perl not use
so much of the system memory? Here is a sample of what I am doing.
[...] foreach (<STDIN>) {

Well, here you are computing the full array of all lines from STDIN, then
foreach loops through them.
Why don't you use the more typical
while (<STDIN>)
which will process line by line?

jue

[Shawn Zabel]

"James B." <ja************ @jsc.nasa.gov> wrote in message
news:75******** *************** ***@posting.goo gle.com...

I have written a perl script to parse some large flat-text logs. The
logs are bzipped and come in through a pipe to STDIN. The script then
performs some regular expressions to the incoming data then prints to
STDOUT.

The script works great, but the issue I have is that the script uses
as much memory as the data coming into it. Therefore if I pipe a 800MB
file into it, the memory usage grows and grows until it reaches
approximatly 800MB, and it doesn't appear to release any of the memory
until it is completely finished. Is there a way to have Perl not use
so much of the system memory? Here is a sample of what I am doing.

#!/usr/bin/perl -w

use strict;

if ($ARGV[0]){
open STDIN, "bzcat $ARGV[0]|" or die "Cant uncompress file as a
pipe\n$!\n";
}

foreach (<STDIN>) {
chomp;
if ($_ =~ /(somedata)/) {
print "$1\n";
}
}

Change

foreach (<STDIN>) {

to

while (<STDIN>) {

The foreach construct reads every line in and then loops through them one at
a time.
The while construct reads one line in at a time and executes the loop.

--
Shawn

[James B.]

Wow, what a simple soultion. Works great, Thanks.

James