RogerBW's Blog

I’ve been doing the Weekly Challenges. The latest involved decoding a string format and partially sorting characters. (Note that this ends today.)

Task 1: Sort String

You are given a shuffle string, $str.

Write a script to return the sorted string.

A string is shuffled by appending word position to each word.

Thus "Challenge3 The1 Weekly2" becomes "The Weekly Challenge".

In Perl this is relatively straightforward: build an output array, use regexps to split word from index, and slot words into the output as they arrive. (Yeah, I could have used an actusal sort with external key, which would have worked if there had been gaps in the index sequence, but this seemed easier.)

sub sortstring($a) {
  my @words = split ' ', $a;
  my @out = ("") x scalar @words;
  foreach my $w (@words) {
    $w =~ /^(.*?)([0-9]+)$/;
    @out[$2 - 1] = $1;
  }
  return join(' ', @out);
}

Most of the other languages work more or less the same way; Lua doesn't have true regexps built in, but does have patterns which do basically the same thing.

But PostScript doesn't have this at all. Sooner or later I'm going to have to write a regexp-like pattern matcher in PostScript. But not right now.

/sortstring {
    0 dict begin
    ( ) strsplit /words exch def
    /out words length array def
    words {

Split the word into an array of characters.

        s2a /ws exch def
        /m -1 def
        /n -1 def

For each character from the end backwards,

        ws reverse {
            /c exch def
            /n n 1 add def

Is it not a digit?

            c c.isdigit not {

Tag that location and exit.

                /m ws length 1 sub n sub def
                exit
            } if
        } forall

Then extract the number…

        ws m 1 add ws length m sub 1 sub getinterval a2s cvi 1 sub /ix exch def

and use it to locate the extracted string.

        out ix ws 0 m 1 add getinterval a2s put
    } forall
    out ( ) strjoin
    end
} bind def

Task 2: Reverse Word

You are given a word, $word and a character, $char.

Write a script to replace the substring up to and including $char with its characters sorted alphabetically. If the $char doesn't exist then don't do anything.

Mostly I didn't need regexps here, and a simple find first character in string is usually quicker. Find the split point, grab the first tranche, sort it, append the second tranche. Ruby:

def reverseword(a, c)
  mm = a.index(c)
  if mm.nil?
    return a
  end
  b = a[0, mm + 1].split("")
  b.sort!()
  b.concat(a[mm + 1, a.size - mm].split("") )
  b.join("")
end

and near-identically in Raku:

sub reverseword($a, $c) {
    with $a.index($c) -> $m {
        my @aa = $a.substr(0, $m+1).comb;
        @aa = sort @aa;
        @aa.push($a.substr($m+1).comb.Slip);
        return @aa.join('');
    } else {
        return $a;
    }
}

Full code on github.