#!/usr/bin/perl my @input = ( [1, 2, 3], [4, 5, 6] ); # [ 1, 2, 3 ] is an anonymous array. # I could have this instead: #@line1 = ( 1, 2, 3 ); #@line2 = ( 4, 5, 6 ); #my @input = ( \@line1, \@line2 ); my @data; my $i = 0; # The 'my' in the loop is imperative, otherwise, each row will erase the # previous one. # The braces in @{ ... } are imperative because of operator precedence # against the [ ] brackets while (my @dataRow = @{$input[$i++]}) { print @dataRow, "\n"; # This stores a pointer to the @dataRow array, i.e. its address or # reference. # So essentially, you get an array of addresses of arrays, which # effectively gives you an array of arrays. Note that each sub-array # (each row) can be different sizes unlike traditional 2-dimensional # arrays. push @data, \@dataRow; } # Now how do we get the data out? # So in our array of arrays, if you say @data[5] you're saying # data[5] which you want to be an array. but that's not the way we # stored things; we stored things as addresses of (or references to) arrays. # So you have to say $data[5] to get the address of the array, and use @ to # get the array, i.e. @{$data[5]} # This is called dereferencing. print "Output loop\n"; foreach (@data) { print @{$_}, "\n"; # I could have this instead, but less clear: #print @$_, "\n"; print ${$_}[0], "\n"; # I could have this instead, but less clear: #print $$_[0], "\n"; } # Or equivalently print "Explicit Output\n"; print @{$data[0]}, "\n"; print ${$data[0]}[0], "\n"; print @{$data[1]}, "\n"; print ${$data[1]}[0], "\n";