From: Robert Dober <robert.dober@...>
Date: 2009-07-29T04:20:02+09:00
Subject: Re: Indexing hash with longer strings faster?

On Tue, Jul 28, 2009 at 9:16 PM, Robert Dober<robert.dober@gmail.com> wrote:
> 2009/7/28 Mladen Jablanović <jablan@gmail.com>:
> Is this 1.8?
> Given your results it is not completely unthinkable that the hash
> values for "00".."99" have more collisions and if collision handling
> is with linked lists that might slow down.
> A priori I would not believe in that but why not check?
>
> I get in 1.9 the expected
>  [*"00".."99"].map( &:hash ).uniq.size
> => 100
>
> can you try this in 1.8
>  ...map{ |x| x.hash }.uniq.size
>
Sorry forgot two things
you did not produce 00, 01, thus
[*0..99].map( &:to_s ).map( a:hash )
and secondly,
a negative result does not say anything about collisions, only a
positive (result < 100) would. In that case we would need to look into
the source, ( I do not think profiling might show internal hash
behavior ).
Cheers
Robert
> ?
>
> HTH
> Robert
> --
> Toutes les grandes personnes ont d’abord été des enfants, mais peu
> d’entre elles s’en souviennent.
>
> All adults have been children first, but not many remember.
>
> [Antoine de Saint-Exupéry]
>



-- 
Toutes les grandes personnes ont d’abord été des enfants, mais peu
d’entre elles s’en souviennent.

All adults have been children first, but not many remember.

[Antoine de Saint-Exupéry]