Sorting an array of hashes which have alphanumeric values - arrays

Question

I am facing an issue with sorting an array of hashes; hashes having alphanumeric values by which I need to sort.
Refer to this question! My question is an extension of this problem.
I tried the solution in the above question but didn't get the required output.
$arr_ref = [
{ brand => "A.1", PO => "1.a", supplier => "X" },
{ brand => "A.2", PO => "2.a", supplier => "X" },
{ brand => "B.1", PO => "1.b", supplier => "X" },
{ brand => "B.2", PO => "2.b", supplier => "X" },
{ brand => "B.3", PO => "1.c", supplier => "Y" },
]
I need to sort by Brand or PO.
#sort the array reference and place the array back into the standard_set
$arr_ref = [sort by_brand #$arr_ref];
sub by_brand {
$a->{brand} cmp $b->{brand}
}
Complexity is the key; can start with numeric or alphabetic character.
The Brand or PO can be of different size, as well. The delimiter may be a dot or hypen.
Can we solve this depending on the input received in $arr_ref?

Answer 1

You were really close. You just need to remove the [ and ] brackets and dereference the array you want to assign to #$array_ref = ....
use strict;
use warnings;
my $arr_ref = [
{ brand => "B.3", PO => "1.c", supplier => "Y" },
{ brand => "B.2", PO => "2.b", supplier => "X" },
{ brand => "B.1", PO => "1.b", supplier => "X" },
{ brand => "A.2", PO => "2.a", supplier => "X" },
{ brand => "A.1", PO => "1.a", supplier => "X" },
];
my #sorted = sort { $a->{brand} cmp $b->{brand} } #$arr_ref;
use Data::Dump;
dd #sorted;
Output:
(
{ brand => "A.1", PO => "1.a", supplier => "X" },
{ brand => "A.2", PO => "2.a", supplier => "X" },
{ brand => "B.1", PO => "1.b", supplier => "X" },
{ brand => "B.2", PO => "2.b", supplier => "X" },
{ brand => "B.3", PO => "1.c", supplier => "Y" },
)

Answer 2

Try this sort method:
sub by_brand_or_po {
my ($a_ba, $a_bn) = split /\.|-/, $a->{brand};
my ($b_ba, $b_bn) = split /\.|-/, $b->{brand};
my ($a_pa, $a_pn) = split /\.|-/, $a->{PO};
my ($b_pa, $b_pn) = split /\.|-/, $b->{PO};
return( $a_ba cmp $b_ba or $a_bn <=> $b_bn or
$a_pa cmp $b_pa or $a_pn <=> $b_pn );
}
It prioritizes the brand alpha over numeric over PO alpha over PO numeric, and will split over a dot or hyphen.

Answer 3

Sort it using Schwartzian transform and correct comparator for each field.
use strict;
use warnings;
use Data::Dumper;
my $arr_ref = [
{ brand => "A.1", PO => "1.a", supplier => "X" },
{ brand => "A.2", PO => "2.a", supplier => "X" },
{ brand => "B.1", PO => "1.b", supplier => "X" },
{ brand => "B.2", PO => "2.b", supplier => "X" },
{ brand => "B.3", PO => "1.c", supplier => "Y" },
];
my #sorted = map $_->[0], sort {
$a->[1][0] cmp $b->[1][0]
or $a->[1][1] <=> $b->[1][1]
or $a->[1][2] <=> $b->[1][2]
or $a->[1][3] cmp $b->[1][3]
} map [ $_, [ map split( /[.-]/, $_, 2 ), #$_{qw(brand PO)} ] ],
#$arr_ref;
print Dumper( \#sorted );
Try use brand "A.10" to spot the difference.

Answer 4

use Sort::Key::Natural!
use Sort::Key::Natural qw(natkeysort);
my #sorted_by_brand = natkeysort { $_->{brand} } #data;
my #sorted_by_po = natkeysort { $_->{po} } #data;