From: David Alan Black <dblack@...>
Date: 2000-12-18T00:11:34+09:00
Subject: [ruby-talk:7508] Re: flatten_by(n)

On Sun, 17 Dec 2000, David Alan Black wrote:

> And, what's more....
> 
>   def flatten_once
>     res = []
>     for e in self     # a little faster than each
>       res.push(*e)
>     end
>     res
>   end


Correction:

  def flatten_once
    res = []
    for e in self
      res += e.to_a   # push won't handle *[] (no args)
    end
    res
  end

> 10000 iterations:
> 
>              user     system      total        real
> mine:    1.210000   0.040000   1.250000 (  1.268839)
> yours:   3.040000   0.000000   3.040000 (  3.042355)


It gets a little less dramatic, but still worthwhile:

      user     system      total        real
  1.740000   0.010000   1.750000 (  1.750059)
  3.110000   0.000000   3.110000 (  3.109317)

(using [1,[2,[3],[]], [5],[],[4,[[6]]]])

Can anyone improve on -- or break -- this version?  


David

-- 
David Alan Black
home: dblack@candle.superlink.net
work: blackdav@shu.edu
Web:  http://pirate.shu.edu/~blackdav