If you just have a string with items separated by dots, split might be able to do the job:

use IO::File;
...

my %fd;
my %files = (
   'somestring' => 'somefilename',
   'someother'  => 'someother',
);

my @items = (
   'somestring.this.content.will.go.into.somefilename',
   'someother.this.will.go.into.someother',
   'notthere.this.will.generate.a.warning',
);

foreach @items {
   my @array = split(/\./);
   if (exists($files{$array[0]})) {
      $fd{$array[0]} ||= new IO::File "> $files{$array[0]}"
         or die "Couldn't open $files{$array[0]}: $!";
      $fd{$array[0]}->print;
   } else {
      warn "No filename defined for $array[0]";
   }
}

$fd{$_}->close for keys %fd;
[download]

my @array = (
   [ 'somefilename',  @data1 ],
   [ 'otherfilename', @data2 ],
);

foreach @array {
   my $filename = shift(@{$_});
   open(F, ">$filename") or die "Couldn't open $filename: $!";
   print F join("\n", @{$_}), "\n";
}
[download]

Is there any way that I can pull out that first entry, and then use that variable to create a filehandle to print to?

Perhaps you should look into using Storable, FreezeThaw and Data::Dumper. This would allow you to take all of your data and just store it in a file, with names that you predetermine (this prevents filename conflicts).

Cheers,
Ovid

Join the Perlmonks Setiathome Group or just go the the link and check out our stats.

Seems to me that FileHandle is a module that might help you out here. Roughly speaking, it lets you treat filehandles as regular scalars. That's very rough (see the module documentation for more info. It's a standard module)

Philosophy can be made out of anything. Or less -- Jerry A. Fodor

It is also an enormous speed hit, and you can already do that, either by using Symbol and then calling gensym, or with the following:

my $fh = do {local *FH};
my $file = "foo.txt";
open ($fh, "> $file") or die "Cannot write '$file': $!";
print $fh "Hello world\n";
[download]

And yes, you can keep filehandles in arrays, etc.

But note that if you try to abuse this, you could run out of filehandles in your program. If lots of programs all open lots of filehandles, you could run out of filehandles on your machine. So it is best to not leave a million of these open...

Its also out-dated. I beleive the standards now are the IO::blah modules (IO::File, IO::Handle, IO::Pipe, etc) which are also part of the standard, and that FileHandle.pm is only included to support old scripts. I've heard people complain about these as well, but if you try to use STDOUT as an object, perl demands IO::Handle

prompt>perl -we "STDOUT->print('test')"
Can't locate object method "print" via package "IO::Handle" at -e line
+ 1.

prompt>perl -MIO::Handle -we "STDOUT->print('test')"
test
[download]

That's using 5.6 (ActiveState 613) on NT4, so I'm not sure how far back this goes. (And yes, that example also works under strict)

Those are also slower.

With Perl 5.6 you don't need to generate symbols. Just use a scalar as a filehandle and it Just Works.

Data and example results:

wfAtmInterfaceGroup.wfAtmCommonGroup

wfAtmInterfaceGroup.wfAtmLinkModuleGroup

wfAtmGroup.wfAtm

wfAtmGroup.wfAtm.wfAtmDelete

wfAtmInterfaceGroup.wfAtmCellSwitchGroup

The result would be

A file named "wfAtmInterfaceGroup.mib" would contain the text:

wfAtmInterfaceGroup.wfAtmCommonGroup

wfAtmInterfaceGroup.wfAtmLinkModuleGroup

wfAtmInterfaceGroup.wfAtmCellSwitchGroup

Another file named "wfAtmGroup.mib" would contain the text:

wfAtmGroup.wfAtm

wfAtmGroup.wfAtmDelete

NOTES: This program is a total hack to get around some badly written MIBs files. I'm not worried about overwriting filenames, since I'll just dump the new files into an empty directory, and if they have the same base stem I *want* them going into the same file (obviously, all files will be opened for appending).

Thanks for the help you have given so far, I'm going to go try out some of those solutions. No matter how many times I see it, I'm always amazed when a new post gets half a dozen answers in under an hour. Thanks for the help.

The guy who asked the question

It looks like the first code snippet in my first response does precisely what you're wanting.