Re^2: Spoiled by Perl (sorting sugar)

[LanX]

coincidentally I procrastinated this Sunday implementing syntactic sugar for exactly this in Perl5.

There are some edge-cases tho, which keep we wondering ...

In Raku:

• how do you decide its a cmp or <=> ?
• what if you want to apply another ordering? Like Czech sorting?
• how do you sort by multiple criteria? Method chaining?

> This will even use a schwartzian transform if the optimizer decides it is worth it.

an optimizer detecting a good case for a Guttman-Rosler Transform transform would be more impressive tho. ;-)

[holli]

how do you decide its a cmp or <=> ?

The dispatcher does that by finding the appropriate multi-sub that implements the actual comparison.

# cmp. Generic, "smart" three-way comparator.

multi sub infix:<cmp>(Any,       Any)
multi sub infix:<cmp>(Real:D,    Real:D)
multi sub infix:<cmp>(Str:D,     Str:D)
multi sub infix:<cmp>(Version:D, Version:D)

# Compares strings with string semantics, numbers with number semantic
+s, Pair objects first by key and then by value etc.
# if $a eqv $b, then $a cmp $b always returns Order::Same. Otherwise O
+rder::Less or Order::More.

what if you want to apply another ordering? Like Czech sorting?

Rakus Str class is fully Unicode aware and by default sorts characters on the numeric value of the codepoint.

dd 'a ä b'.comb.sort;
(" ", " ", "a", "b", "ä").Seq;

If you wanted to implement your own behaviour you could just add another multi

subset CzechString of String;
multi sub infix:<cmp>(CzechString $a, CzechString $b) {
  # sort logic here
}

Or you can simply pass your comparator to sort

sub czech-sort(String $a, String $b) {
  # sort logic here
}

@czech-words.sort( &czech-sort );

how do you sort by multiple criteria? Method chaining?

You can adress that in the comparator like in Perl, or make the comparator return a list of things.

dd ([1,'z'], [2,'a'], ['1', 'a']).sort( *.[0,1] );
(["1", "a"], [1, "z"], [2, "a"]).Seq

dd ([1,'z'], [2,'a'], ['1', 'a']).sort( { $^a[0] <=> $^b[0] || $^a[1] 
+cmp $^b[1] });

[LanX]

thanks! :)

> The dispatcher does that by finding the appropriate multi-sub that implements the actual comparison.

ehm ... if I have an array with numbers and strings it will implicitly decide based on the type of $a and $b which comparison to apply?

> Rakus Str class is fully Unicode aware and by default sorts characters on the numeric value of the codepoint.

I'm not sure if that's sufficient to handle different standards.

E.g. the telephone book in Germany has another sorting than the dictionaries.

update

clearer rewording

[raiph]

> if I have an array with numbers and strings it will implicitly decide based on the type of $a and $b which comparison to apply?

It will decide for you if you don't explicitly direct it to do otherwise. By default, given a number and a string to compare, it will coerce the number to a string. But if both elements being compared are numbers it will compare them numerically. If you want to sort all numbers as strings, you must say so, eg by passing `~*` as the comparator.

> I'm not sure if \sorting by codepoint is\ sufficient to handle different standards. E.g. the telephone book in Germany has another sorting than the dictionaries.

Use https://docs.raku.org/routine/collate instead of `sort`.