From: William James <w_a_x_man@...>
Date: 2007-01-28T11:10:14+09:00
Subject: Re: Review of code please



On Jan 27, 12:30 pm, "Erik Veenstra" <erikv...@gmail.com> wrote:
> Here's a map-reduce-ish [1] solution.
>
> I don't say it's better, or faster, or smaller (memory-wise).
> It's just, uh, different...
>
> (Sorry, couldn't resist, again... ;])
>
> gegroet,
> Erik V. -http://www.erikveen.dds.nl/
>
>  [1]http://labs.google.com/papers/mapreduce.html
>
> ----------------------------------------------------------------
>
>  WITHOUT THE TOTALS
>
>  File.open("anagrams.txt", "w") do |f|
>    File.readlines("/usr/share/dict/words").collect do |word|
>      word.chomp.downcase
>    end.collect do |word|
>      [word.scan(/./).sort, word]
>    end.inject({}) do |hash, (base, word)|
>      (hash[base] ||= []) << word ; hash
>    end.values.reject do |set|
>      set.length == 1
>    end.collect do |set|
>      set.sort
>    end.sort_by do |set|
>      [-set.length, set]
>    end.collect do |set|
>      set.join("\t")
>    end.each do |line|
>      f.puts(line)
>    end
>  end

Do you know what a one-to-one mapping is?
It's senseless to use "collect"; use "map" instead.

open( 'anagrams.txt', 'w' ) { |file|
  IO.readlines( 'words' ).inject({}){ |hash,word|
  word.strip!.downcase!
  (hash[word.split(//).sort.join] ||=[]) << word; hash }.
  values.reject{|list| list.size < 2}.
  map{|list| list.sort}.sort_by{|list| [ -list.size, list ] }.
  each{|list| file.puts list.join( "\t" ) } }