From: Phrogz <phrogz@...>
Date: 2007-11-12T00:40:00+09:00
Subject: Re: equivalent injecting implementations?

On Nov 10, 1:23 pm, Ezra Zygmuntowicz <ezmob...@gmail.com> wrote:
> injecting is similar to returning in rails. There is a common pattern
> of accumulating a hash via inject that looks like this:
>
> [1,2,3].inject({}) { |memo,n| memo[n] = n ;  memo}
> #=> {1=>1, 3=>3,2=>2}
>
> Note the ugly trailing ;memo, this is so the fll accumulated hash is  
> the final return result.

Note that you can avoid the trailing memo with this:
  [1,2,3].inject({}) { |memo,n| memo.update n=>n }
...but it's a good deal slower:

  a = (1..10_000).to_a
  N = 500

  require 'benchmark'
  Benchmark.bmbm{ |x|
    x.report( 'set' ){
      N.times{
        a.inject({}){ |memo,n| memo[n] = n; memo }
      }
    }
    x.report( 'update' ){
      N.times{
        a.inject({}){ |memo,n| memo.update n=>n }
      }
    }
  }

  Rehearsal ------------------------------------------
  set      5.860000   0.010000   5.870000 (  5.981301)
  update  11.830000   0.020000  11.850000 ( 11.865100)
  -------------------------------- total: 17.720000sec

               user     system      total        real
  set      5.960000   0.040000   6.000000 (  6.011882)
  update  11.850000   0.070000  11.920000 ( 11.925828)