use IO::Uncompress::Unzip qw(unzip $UnzipError);
my $file = '2.zip';
my $inner;
unzip $file, \$inner, Name => '1.zip';
[download]

my $txt;
unzip \$inner, \$txt, Name => '1.txt';
[download]

Apparently the "Name => $name" part is optional if the ZIP archive contains only one file (what they call a "member file"):

Say you have a zip file, file1.zip, that only contains a single member, you can read it and write the uncompressed data to the file file1.txt like this.

unzip $input => $output
        or die "unzip failed: $UnzipError\n";
[download]

If you have a zip file that contains multiple members and want to read a specific member from the file, say "data1", use the Name option

unzip $input => $output, Name => "data1"
        or die "unzip failed: $UnzipError\n";
[download]

That code will work fine, but here is a variation that means you don't need to uncompress the complete outer zip into memory before accessing the inner one.

use warnings;
use strict;

use IO::Uncompress::Unzip qw(:all);

my $file1 = "all.zip";
my $file2 = "1.zip";

my $first = new IO::Uncompress::Unzip $file1
    or die "Cannot open $file1: $UnzipError\n";

my $output;
unzip $first => \$output
    or die "Cannot open $file2: $UnzipError\n";
[download]

Thanks bart for code. It is working! I do have another question regarding the following script: unzip $file, \$inner, Name => '1.zip'; What does \$inner mean? Unzipping file into a temp location called inner?

\$inner is a reference to the scalar $inner. It allows programmers to change the content of the variable $inner without explicitly mentioning it by name in the source. You can just as well use a different variable and it'll work the same. For example:

$apples = 10;
$pears = 6;

take_one(\$apples);   # take one apple
take_one(\$pears);    # take one pear

print "I've still got $apples apples and $pears pears.\n";

sub take_one {
    my $ref = shift;  # a reference
    $$ref--           # access the value of the referenced scalar
}
[download]

This prints:

I've still got 9 apples and 5 pears.
[download]

For more info, check out perlreftut and perlref.

It's easy for a programmer to detect if a value is a reference: just use ref: it'll return an empty string for a normal string, and "SCALAR" for a reference to a scalar.

It's a convention among module authors to use a reference to a scalar if you want to actually use the string value as data, instead of as a file name from which you'll read the contents; for example in the various HTML parsing modules.

When I wrote unzip I wanted to allow flexibility on where it obtained it's input and where it sent its output. Currently the input and output paramers can be a filename, filehandle or in-memory.

So, for example

my $inner;
unzip "my.zip" => \$inner;

# remove all whitespace
$inner =~ s/\s+//g;

print $inner;
[download]

This will uncompress the contents of the file "my.zip" and write the uncompressed data into the variable $inner.

The code then makes use of the uncompressed data in this case to remove all whitespace, then print it.

Or, at a minimum, doing so and demonstrating that you have with some question not clearly answered there, would at least belie your chosen nick.

I have written a code to read a file in one level of compression. I have been searching through to look at how I can read a file under 2 level of compression. Its the methodology I required which can be advance such as hash tracking. The nick is more for fun ;)