PWC224 - Special Notes


Here we are with TASK #1 from The Weekly Challenge #224. Enjoy!

The challenge

You are given two strings, $source and $target.

Write a script to find out if using the characters (only once) from source, a target string can be created.

Example 1

Input: $source = "abc"
       $target = "xyz"
Output: false

Example 2

Input: $source = "scriptinglanguage"
       $target = "perl"
Output: true

Example 3

Input: $source = "aabbcc"
       $target = "abc"
Output: true

The questions

This is a deceptively fun and clever challenge, so my question is: how did you think of it?

More seriously, often times we’ve been shown that case might be disregarded, spaces might be disregarded, etc. so my question would be: can we assume that a basic eq comparison is OK to figure out whether two characters are the same?

Additionally, and more importantly: do we already know that $source and $target contain strings of characters and we don’t have to do any decoding?

The solution

For this challenge, Raku can be the mythic Alexander the Great’s sword while dealing with the Gordian Knot (or Indiana Jones’s gun, for a more modern version of the same idea). Brutally effective:

#!/usr/bin/env raku
use v6;
sub MAIN ($s, $t) { put(([(-)] ($t, $s)».comb».Bag).elems == 0) }

We build two [Bag][]s out of each string, then subtract (in [Bag][] terms) the available characters from the source from the target. If we’re left with anything, then the source did not contain what we needed; otherwise, it did.

Going on to Perl, we don’t have all that many batteries included, so we are forced to use the head. There can be many approaches, like the simple one:

sub special_notes_simple ($source, $target) {
   my %available;
   $available{$_}++ for split m{}mxs, $source;
   for my $char (split m{}mxs, $target) {
      return unless $available{$char};
   return 1;

We first collect all available characters from the source, then use them to cover the target’s needs.

At this point… why sweep through the whole source, when we might not need it? So we might think something different instead:

sub special_notes ($source, $target) {
   my $sl = length($source);
   my $tl = length($target);
   return unless $tl <= $sl;

   my %available;
   my $si = 0; # index in $source
   for my $ti (0 .. $tl - 1) {
      my $tch = substr($target, $ti, 1);
      if ($available{$tch}) {
         my $residual = --$available{$tch};
         delete $available{$tch} unless $residual;
      else {
         while ($si < $sl) {
            my $sch = substr($source, $si++, 1);
            next TARGET if $sch eq $tch;
         return; # no luck...
   return 1;

Is this any better? Surely there’s more code that might fail and that might need attention, so not in general. On the plus side it does not have to necessarily sweep through the whole of the source, so it might give a little edge time-wise (and space-wise, I daresay).

Anyway, for small inputs, I think that nothing beats the brutal perfection of the Raku solution.


Comments? Octodon, , GitHub, Reddit, or drop me a line!