One way to find what's missing among Phobos ranges and higher order functions is to try to use them. This is a small Rosettacode task, given a file of words, it asks to find the longest pair of anagrams that have not even one letter in the same position:

http://rosettacode.org/wiki/Anagrams/Deranged_anagrams

I don't fully understand the Clojure solution:


(let
  [words    (re-seq #"\w+" (slurp "unixdict.txt"))
   anagrams (filter second (vals (group-by sort words)))
   deranged (remove #(some true? (apply map = %)) anagrams)]
  (prn (last (sort-by #(count (first %)) deranged))))


I have written two D solutions, the first tries to be short and the second to be faster. This is the first solution (look in that page for the fast solution):


void main() {
    import std.stdio, std.file, std.algorithm, std.string,
           std.array;

    string[][dstring] anags;
    foreach (const w; "unixdict.txt".readText.split)
        anags[w.array.sort().release.idup] ~= w;

    anags
    .byValue
    .map!(anags => anags.cartesianProduct(anags)
                   .filter!q{a[].equal!{ a != b }})
    .join
    .minPos!q{ a[0].length > b[0].length }[0]
    .writeln;
}



A better short D version:

void main() {
    import std.stdio, std.file, std.algorithm, std.string,
           std.array;

    "unixdict.txt"
    .readText
    .split
    .hashGroupBy!(w => w.sorted.release)
    .byValue
    .map!(anags => anags
                   .pairwise
                   .filter!q{a[].equal!q{ a != b }})
    .join
    .max!q{ a[0].length }
    .writeln;
}


Where:

hashGroupBy returns an associative array of the equivalence classes, given a key function (http://d.puremagic.com/issues/show_bug.cgi?id=9842 ).

sorted is similar to an array+sort. But it allocates a new array, so the its output should be implicitly castable to const (http://d.puremagic.com/issues/show_bug.cgi?id=5076  https://d.puremagic.com/issues/show_bug.cgi?id=12140 ).

pairwise is a range that yields the pairs, but only the lower triangle. So it's like two nested loops. In this program it's up to twice faster than using cartesianProduct (http://d.puremagic.com/issues/show_bug.cgi?id=6788 ).

max and min accept an optional mapping key function, as in Python (http://d.puremagic.com/issues/show_bug.cgi?id=4705 ).

Bye,
bearophile