Karyn Williams am Donnerstag, 29. März 2007 20:27:
[snip]
> It varies greatly. Often one or two. Sometimes 70. The size of the files is
> fairly substantial:
>
> -rw-r--r-- 1 root other 249595095 Sep 1 2006 maillog.200608
>
> I am getting Out of memory errors while running the script.
[snip]
> I have hashed and slurped files in other scripts. I could try it here and
> see what happens.
[snip]
> >> >> open MAILLOG,
> >>
> >> "/export/home/archives/maillog.$flist[1]" or die
> >>
> >> >> "couldn't open maillog.$flist[1] : $!\n";
> >> >> if ($count = grep /user=$k/o, <MAILLOG> ) {
> >> >> print "$k checked mail $count times in
> >> >> maillog.$flist[1].\n"; next ;
> >> >> } else {
> >> >> close MAILLOG;
[snip]
Hello Karyn
(Didn't quite got the overview: Formatting, combined top/inline posting,
unstripped postings. So the following may be inappropriate)
Try to read the logfiles line per line (instead of slurping the whole log file
into memory), using a while loop, along the lines:
while (<MAILLOG>) {
# handle log line (contained in $_) here
}
# handle result of parsing the whole file here
This will use minimal memory resources.
> # ./tt.pl karyn smurphy root
> karyn checked mail 2864 times in /var/adm/maillog.
> smurphy checked mail 2864 times in /var/adm/maillog.
> root checked mail 2864 times in /var/adm/maillog.
> #
>
> Yes, it goes to the next arg but as you can see, it returns the same count
> for each subsequent arg.
The reason for this probably is the o modifier in /user=$k/o regexes: this
will fix the value of $k within the regex. Try to define a precompiled regex
every time $k changes - this will compile a regex once for every different
value of $k:
# value assigned to $k here
my $regex=qr/user=$k/;
# ...
$variable=~/$regex/ && count++;
This is the alternate way to speed up matching.
Dani
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
