Re^14: Is there any difference between prototype % and @ ?

[LanX]

> And so far; neither you nor I have come up with a use case.

trivial example

%salary=(
     boss      => 1_000,
     secretary => 300,
     admin     => 600,
     janitor   => 150,
    );

print
  map {"$_ earns $salary{$_} \n" }
  sort { $salary{$a} <=> $salary{$b} }
  grep { $salary{$_} > 500 }
  keys %salary;

translating it to hmap, hsort and hgrep is left as an excercise.

[BrowserUk]

sub hpairs (\%) {
    my $href = shift;
    map[ each %$href ], 1 .. keys %$href;
}

my %salary = (
    boss      => 1_000,
    secretary => 300,
    admin     => 600,
    janitor   => 150,
);

print for map { 
    "$_->[0] earns $_->[1]" 
} sort { 
    $a->[1] <=> $b->[1] 
} grep { 
    $_->[1] > 500 
} hpairs %salary;

__END__
admin earns 600
boss earns 1000

Or prettier:

sub hpairs (\%) {
    use enum 'KEY', 'VAL';
    my $href = shift;
    map[ each %$href ], 1 .. keys %$href;
}

my %salary = (
    boss      => 1_000,
    secretary => 300,
    admin     => 600,
    janitor   => 150,
);

print for map {
    "$_->[KEY] earns $_->[VAL]"
} sort {
    $a->[VAL] <=> $b->[VAL]
} grep {
    $_->[VAL] > 500
} hpairs %salary;

[LanX]

nice, but
1. doesn't even try to use the hmap or hgrep you defended for so long
2. you needed to introduce a completely new function hpairs
3. still needs a unhpairs to flatten LoA before assigning after chain to a new hash
4. my hmap could already be used instead of your hpairs (and unhpairs) in this approach ...

Let's count, you're going to introduce 5 new functions where I only need 3 to cover all discussed cases...

You really hate symmetry, don't you?

[BrowserUk]

1. doesn't even try to use the hmap or hgrep you defended so long

   I made no attempt to "defend". There is nothing to "defend".

   I offered an alternative, With reasoning. (I'm not forcing anyone to agree; nor use my ideas.)

2. you needed to introduce completely new hpairs

   I use (or construct) the right tool for the job.

3. still needs a unhpairs to flatten LoA before assigning to a new hash after chain

   Nothing new required, map serves the purpose perfectly.

4. I could already use my hmap in your solution, so no need to create hpairs

   Look as hard as I might; I can't find your definition of hmap?

5. you really hate symmetry, don't you?

   Nope! I "hate" very little. (And is that less "personal" that my recognising your ability?)

   But I do not seek to artificially impose symmetry where it does not naturally belong.