Memory use when reading a text file

[abingham]

Below is my code snippet...

...simple job, open a text log file, read through it line by line and then close it.

I am actually trying to extract statistics from our mail logs, so the final code will count up the number of messages to from each address and store this in a hash to be output to CSV file at the end. However, that's getting beyond myself.

The problem is when I run this with a 46Mb file, my WinXp workstation uses up all 1GB of RAM + Swap and goes ape. This seems alot of memory usage to just read through a file, even if the machine caches the entire file in memory. With this code, I am not storing any of the data or manipulating it in any way!

Any ideas from the Perl Monks?

Is there a more efficient way of reading through files?

# process.pl
# andy bingham
# 14 April 2002
#

open(INP, "log.txt") or die "cannot open input";

foreach $line (<INP>) { # read in each line of INP
        #processing code goes here
}
close(INP);
# 
# DONE 
#

[Abigail-II]

You are using <INP> in list context. That will make Perl slurp in the entire file, and turn it into a list of lines.

The standard idiom of processing a file line by line is:

    while (<IN>) {
        # Processing goes here; the current line is in $_
    }

That should cut your memory usage as it doesn't read in the entire file in memory (unless there's just one line in the file...)

Abigail

[ichimunki]

This is related to your use of for(each), which tries to store the whole file in a temporary array in memory.

Try:

while( my $line = <INP> ){
    ...
}

instead. This will simply copy each line into memory as needed.