concating files together

[Andrew Wilson]

On Fri, May 23, 2003 at 04:09:24PM +0100, Boyle Bernadette wrote:
> Andrew,
> 
> I was using it in my perl script as exec `more /var/log/maillog* >
> /tmp/output` but would have preferred to use Perl syntax.  Your suggestion
> is very cool!!

Thanks
 
>   perl -p -e'' /var/log/maillog* > /tmp/output

In a script, I wouldn't do that.  The @ARGV array in perl is special, it
holds the command line arguments.  when you use the -p option, perl
treats each entry as a filename, it opens each in turn, gives you one
line at a time then moves to the next file.  In a perl script, you can
use <> to get the same special behaviour, so you can do.

while (<>) {
  print;
}

And get a similar effect to the -p option.  You can set the @ARGV array
at any point, so you could do:

{
  @ARGV = qw{file1 file2 file3};
  while (<>) {
    print;
  }
}

But that destroys @ARGV.  You can avoid that problem by localising
@ARGV:

{
  local @ARGV = qw{file1 file2 file3};
  while (<>) {
    print;
  }
}

And since you know what output file you want (lets say you assign it to
$maillog) then you can do:

my $maillog = '/tmp/maillog';

{
  open LOG, $maillog or die "Can't open $maillog $!";
  local @ARGV = qw{file1 file2 file3};
  while (<>) {
    print OUT $_;
  }
}

The only remaining problem is that the list of files you're going to
concat is fixed.  You can get a list that changes like the /log/maillog*
in your original example using the glob function.

my $maillog = '/tmp/maillog';

{
  open LOG, $maillog or die "Can't open $maillog $!";
  local @ARGV = glob('/log/maillog*');
  while (<>) {
    print OUT $_;
  }
}

Hope that helps.

andrew
-- 
Virgo: (Aug. 23 - Sept. 22)
If there's one thing you should try to learn from next week's events,
it's the precise melting point of aluminium.