From: Stefano Crocco <stefano.crocco@...>
Date: 2007-11-29T19:29:44+09:00
Subject: Re: Hash Sorting

Alle giovedì 29 novembre 2007, Nathan Viswa ha scritto:
> Can not understand how the block after sort works! Need help. thanks.
>
>    h = { "a" => 20, "b" => 30, "c" => 10  }
>    puts h.sort                     #=> [["a", 20], ["b", 30], ["c", 10]]
>    puts h.sort {|a,b| a[0]<=>b[0]} # as above
>    puts h.sort {|a,b| a[1]<=>b[1]} #=> [["c", 10], ["a", 20], ["b", 30]]

What Hash#sort does is: 
* first: convert the hash into a nested array:
  [['a', 20], ['b', 30], ['c', 10]]
* second: sort that array according to the block.

In the first call to sort, you don't pass a block to it, so the sorting will 
be performed by calling the <=> of the contents of the array (i.e, on the 
key - value arrays). Array#<=> compares the contents of the two arrays in 
order, returning +1 or -1 as soon as one of the items is different from the 
other. For example, ['a', 20]<=>['c', 10] returns -1 because 'a' 
precedes 'c'. If two entries had the same first element (here this is 
impossible since they come from a hash), the second element would be 
compared, an so on.

Your second call to send gives the same result, because you're explicitly 
telling sort to compare only the first element of the array. As explained 
above, even in the blockless case sort never needs to check the second 
element, so the results are the same.

In the third case, you tell sort to compare the second element of each pair 
(i.e, the value in the original hash). This way, the pair ['c', 10] becomes 
the first, because 10 is the lesser of the three values; ['a', 20] is the 
second because 20 is the middle value and ['b', 30] is the last because 30 is 
the greater value.

I hope this helps

Stefano