From: Tammo Freese <tammo.freese@...>
Date: 2001-03-06T23:50:14+09:00
Subject: [ruby-talk:12139] Re: (long) Re: hash.invert loses data if  equal values exist - is this the right behaviour?

At 23:10 06.03.01 +0900, you wrote:
>Tammo Freese <tammo.freese@offis.de> writes:
>
>> By inverting a hash that has same values, I expected to see an Exception.
>> Curious which it would be, I tested it and found a surprising result:
>> 
>> D:\software>ruby -ve 'p [hash = Hash[1, 3, 2, 3], hash.invert]'
>> ruby 1.6.2 (2000-12-18) [i386-cygwin]
>> [{1=>3, 2=>3}, {3=>2}]
>> 
>> No exception. Is this correct? 
>
>Any reason why there should be an exception?
>
>  h = { 1 => 2 }	# => {1=>2}
>  h[1] = 3 		# => 3
>  h                     # => {1=>3}
>
>It isn't exceptional to overwrite a value in a hash with another.

Yes, it isn't. But the problem here is that it is not easy to predict
which hash value overrides the others, because it seems to depend
on the value and not on the order of appending it to the hash.

Can you tell me how to figure out the result of the following code
easily without looking at variable a and without thinking about
the hashing strategy?

puts "First:"

#
a = {}
a[7] = 100
a[10] = 100
a[12] = 100

p a.invert

puts "Second:"

a = {}
a[11] = 100
a[27] = 100
a[41] = 100

p a.invert

puts "Third: " 

a = {}
a[1] = 100
a[5] = 100
a[14] = 100

p a.invert
#

Or is the morale that I should not invert hashes with equal values
because the results are too difficult to predict?



-
Tammo Freese
freese@acm.org