From: SpringFlowers AutumnMoon <summercoolness@...>
Date: 2007-09-18T22:21:56+09:00
Subject: Re: how to use tuple as hash key

Rick Denatale wrote:
> On 9/18/07, SpringFlowers AutumnMoon <summercoolness@gmail.com> wrote:

> I think you might be tripping over the difference between a variable
> and an object here.  I know that lots of posters to this forum do.
> Variables aren't objects per se. A variable references an object.
> Variable assignment copies the reference NOT the object.

actually, my posting of "The meaning of a = b in object oriented 
languages"
reminded me of what I learned in Java... that the variables are not 
objects
themselves.  the variable merely references an object, like a pointer to 
an
object.

the main confusion I had with hash (and i think now clear) is that Hash 
first
used hash() on the key's value to find a hash slot.  and then at the 
slot, the
mapping is done not for the key's value to the final value.
the mapping merely stores the pointer to the key object, and the pointer
to the value object.

so if the key object changes, then the hash() will most likely hash to 
another
hash slot, and then found that there is no mapping there, and therefore 
store
another pointer to key, pointer to value pair there.

thus the result:

irb(main):084:0> a = [1,2]
=> [1, 2]
irb(main):085:0> h = {}
=> {}
irb(main):086:0> h[a] = "ha"
=> "ha"
irb(main):087:0> a[1] = 99
=> 99
irb(main):088:0> a
=> [1, 99]
irb(main):089:0> h[a] = "hee"
=> "hee"
irb(main):090:0> h
=> {[1, 99]=>"ha", [1, 99]=>"hee"}
irb(main):091:0> h[[1,2]]
=> nil
irb(main):092:0> h[[1,99]]
=> "hee"
irb(main):093:0> h.rehash
=> {[1, 99]=>"hee"}
irb(main):094:0>


Example 2:

irb(main):094:0> h = {}
=> {}
irb(main):095:0> b = "haha"
=> "haha"
irb(main):096:0> h["a"] = b
=> "haha"
irb(main):097:0> h
=> {"a"=>"haha"}
irb(main):098:0> b[3] = "i"
=> "i"
irb(main):099:0> h
=> {"a"=>"hahi"}
irb(main):100:0> b = "another"
=> "another"
irb(main):101:0> h
=> {"a"=>"hahi"}
irb(main):102:0>

-- 
Posted via http://www.ruby-forum.com/.