Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!usc!ucla-cs!henry.jpl.nasa.gov!elroy.jpl.nasa.gov!forsight!jato!jpl-devvax!lwall
From: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Newsgroups: comp.lang.perl
Subject: Re: How to zero a file
Message-ID: <6677@jpl-devvax.JPL.NASA.GOV>
Date: 27 Dec 89 22:46:48 GMT
References: <891@pan.UUCP>
Reply-To: lwall@jpl-devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA
Lines: 118

In article <891@pan.UUCP> jw@pan.UUCP (Jamie Watson) writes:
: What is the preferred way to zero a file in perl?  In shell scripts I
: use "> filename", so in perl I can 'exec "> filename"', but it seems
: to me that there should be a way to do this without using exec.

open(ZERO,'> filename') && close ZERO;

If you do this a lot you can always make it a subroutine:

sub zero {
    open(ZERO, "> $_[0]") && close ZERO;
}

&zero('filename');

: In a similar vein, what is the preferred way to copy a file?  I could
: obviously just crank up 'exec "cp old new"', but again I wonder if there
: is some "better" way to do this entirely within perl.  Copying a line at
: a time, with "while () print" doesn't seem to me to be an ideal solution.

There are several ways to do it, depending on what you're optimizing for.

# Most concise without defining a subroutine
# (and probably fastest on huge files)

system 'cp old new';

0.0u 0.2s 0:01 (timed on /etc/termcap, size 165500)

# Most concise without a subprocess, but slurps whole file

open(OLD,'old');
open(NEW,'>new');
print NEW <OLD>;
close NEW;
close OLD;

1.3u 0.6s 0:03

# much less space taken

open(OLD,'old');
open(NEW,'>new');
while (<OLD>) {
    print NEW;
}
close NEW;
close OLD;

0.8u 0.2s 0:02

# more space taken

open(OLD,'old');
open(NEW,'>new');
$/ = "\000";
while (<OLD>) {
    print NEW;
}
close NEW;
close OLD;

1.0u 0.3s 0:02

# fastest

open(OLD,'old');
open(NEW,'>new');
while (read(OLD,$foo,8192)) {
    print NEW $foo;
}
close NEW;
close OLD;

0.1u 0.1s 0:01

# almost as fast, maybe faster on multiple files

open(OLD,'old');
($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size) = stat OLD;
open(NEW,'>new');
read(OLD,$foo,$size);
print NEW $foo;
close NEW;
close OLD;

0.0u 0.3s 0:01

# slowest

open(OLD,'old');
open(NEW,'>new');
select(NEW); $| = 1;
while (($foo = getc(OLD)) ne '') {
    print NEW $foo;
}
close NEW;
close OLD;

66.1u 73.3s 2:39

# weirdest

#!/usr/local/bin/perl -i.bak

symlink('old','new');
@ARGV=('new');
while (<>) {
    print;
}
unlink 'new.bak';

0.8u 0.3s 0:02

Your mileage may vary, depending on how stdio is implemented on your system,
and in particular on how efficient your fread() and fwrite() are.

Larry