For Loop Issues in creating nested array - arrays

Question

Creating a matrix of products for three element arrays. I understand Perl does not have multi-dimensional arrays and are flattened. I have been using refs but I can't seem to get past the for loop issue in getting three products into a single array and pushing that array into a different single array. And I could be way off too. Be nice, but I've spent too many hours on this.
I have moved values inside and out of various places i.e. { }, printed out variables until I'm blue and used $last all over for debugging. I'm likely fried at this point.
use strict;
use warnings;
my #array1 = (1, 2, 3);
my #array2 = (2, 4, 6);
my #matrixArray = ();
my $matrixArray;
my #row;
my #finalArray = maths(\#array1, \#array2);
print #finalArray;
sub maths{
my $array1ref = shift;
my $array2ref = shift;
my $value1;
my $value2;
my $maths;
my #row = ();
my #array1 = #{$array1ref};
my #array2 = #{$array2ref};
my $len1 = #array1;
my $len2 = #array2;
for my $x (0 ..($len1 -1)){
#iterate through first array at each value
$value1 = $array1[$x];
#print $value1, " value1 \n";
for my $y (0 .. ($len2 -1)){
#iterate through second array at each value
$value2 = $array2[$y];
#print $value2, " value2 \n";
#calculate new values
$maths = $value1 * $value2;
#exactly right here
#print $maths, " maths \n" ;
push #row, $maths;
}
}
#and exactly right here but not set of arrays
#print #row, "\n";
return #row;
}
Currently I'm able to get this: 246481261218. Which is the correct dumb math but...
it should appear as a matrix:
2 4 6
4 8 12
6 12 18
I am not passing three arrays so it seems my issue is up in the sub routine before I can get on with anything else. This seems to be a theme that I often miss. So sorry if I sound inept.
EDIT***
This was working but I couldn't unpack it
use strict;
use warnings;
my #array1 = (1, 2, 3);
my #array2 = (2, 4, 6);
my #matrixArray = ();
maths(\#array1, \#array2);
foreach my $x (#matrixArray){
print "$x \n";
}
sub maths{
my $array1ref = shift;
my $array2ref = shift;
my $value1;
my $value2;
my $maths;
my #row = ();
my $row;
my #array1 = #{$array1ref};
my #array2 = #{$array2ref};
my $len1 = #array1;
my $len2 = #array2;
for my $x (0 ..($len1 -1)){
#iterate through first array at each value
$value1 = $array1[$x];
for my $y (0 .. ($len2 -1)){
#iterate through second array at each value
$value2 = $array2[$y];
#calculate new values
$maths = $value1 * $value2;
push #row, $maths;
$row = \#row;
}
push #matrixArray, $row;
}
return #matrixArray;
}
The output right after the function call is this:
ARRAY(0x55bbe2c667b0)
ARRAY(0x55bbe2c667b0)
ARRAY(0x55bbe2c667b0)
which would be the (line 10) print of $x.
****EDIT
This Works (almost):
print join(" ", #{$_}), "\n" for #matrixArray;
Output is a bit wrong...
2 4 6 4 8 12 6 12 18
2 4 6 4 8 12 6 12 18
2 4 6 4 8 12 6 12 18
And of note: I knew $x was an array but I seemed to run into trouble trying to unpack it correctly. And I'm no longer a fan of Perl. I'm pining for the fjords of Python.
And *****EDIT
This is working great and I get three arrays out of it:
sub maths{
my ($array1, $array2) = #_;
my #res;
for my $x (#$array1) {
my #row;
for my $y (#$array2) {
push #row, $x * $y;
}
push #res, \#row;
}
#This is the correct structure on print #res!
return #res;
}
But, though it's putting it together correctly, I have no output after the call
maths(\#array1, \#array2);
NOTHING HERE...
print #res;
print join(" ", #{$_}), "\n" for #res;
foreach my $x (#res){
print join(" ", #{$x}), "\n";
}
And of course a million thanks! I regret taking this stupid course and fear my grade will eventually do me in. Still pining for Python!

Answer 1

It appears that you need a matrix with rows obtained by multiplying an array by elements of another.
One way
use warnings;
use strict;
use Data::Dump qw(dd);
my #ary = (2, 4, 6);
my #factors = (1, 2, 3);
my #matrix = map {
my $factor = $_;
[ map { $_ * $factor } #ary ]
} #factors;
dd #matrix;
The array #matrix, formed by the outer map, has array references for each element and is thus (at least) a two-dimensional structure (a "matrix"). Those arrayrefs are built with [ ], which creates an anonymous array out of a list inside. That list is generated by map over the #ary.
I use Data::Dump to nicely print complex data. In the core there is Data::Dumper.
With a lot of work like this, and with large data, efficiency may matter. The common wisdom would have it that direct iteration should be a bit faster than map, but here is a benchmark. This also serves to show more basic ways as well.
use warnings;
use strict;
use feature 'say';
use Benchmark qw(cmpthese);
my $runfor = shift // 5; # run each case for these many seconds
sub outer_map {
my ($ary, $fact) = #_;
my #matrix = map {
my $factor = $_;
[ map { $_ * $factor } #$ary ]
} #$fact;
return \#matrix;
}
sub outer {
my ($ary, $fact) = #_;
my #matrix;
foreach my $factor (#$fact) {
push #matrix, [];
foreach my $elem (#$ary) {
push #{$matrix[-1]}, $elem * $factor;
}
}
return \#matrix;
}
sub outer_tmp {
my ($ary, $fact) = #_;
my #matrix;
foreach my $factor (#$fact) {
my #tmp;
foreach my $elem (#$ary) {
push #tmp, $elem * $factor;
}
push #matrix, \#tmp;
}
return \#matrix;
}
my #a1 = map { 2*$_ } 1..1_000; # worth comparing only for large data
my #f1 = 1..1_000;
cmpthese( -$runfor, {
direct => sub { my $r1 = outer(\#a1, \#f1) },
w_tmp => sub { my $r2 = outer_tmp(\#a1, \#f1) },
w_map => sub { my $r3 = outer_map(\#a1, \#f1) },
});
On a nice machine with v5.16 this prints
Rate direct w_map w_tmp
direct 11.0/s -- -3% -20%
w_map 11.4/s 3% -- -17%
w_tmp 13.8/s 25% 21% --
The results are rather similar on v5.29.2, and on an oldish laptop.
So map is a touch faster than building a matrix directly, and 15-20% slower than the method using a temporary array for rows, which I'd also consider clearest. The explicit loops can be improved a little by avoiding scopes and scalars, and the "direct" method can perhaps be sped up some by using indices. But these are dreaded micro-optimizations, and for fringe benefits at best.
Note that timings such as these make sense only with truly large amounts of data, what the above isn't. (I did test with both dimensions ten times as large, with very similar results.)

Answer 2

The second program is mostly correct.
The problem is that you didn't unpack the second level of the array.
foreach my $x (#matrixArray){
print "$x \n";
}
should be something like:
foreach my $x (#matrixArray) {
print join(" ", #{$x}), "\n";
}
# or just:
print join(" ", #{$_}), "\n" for #matrixArray;
Your maths function can be made shorter without losing legibility (it may actually make it more legible) by cutting out unnecessary temporary variables and indexing. For example:
sub maths {
my #array1 = #{ $_[0] };
my #array2 = #{ $_[1] }; # or: ... = #{ (shift) };
my #res = ();
for my $x (#array1) {
my #row = (); # <-- bugfix of original code
for my $y (#array2) {
my $maths = $x * $y;
push #row, $maths;
}
push #res, \#row;
}
return #res;
}