If that helps, I'm currently working on a "custom" type of collection
used for sampling objects based on probabilities. Should be ready in a
few days if anyone's interested. For lack of a better word, it's
currently called "SamplingSet". To make a long story short, I'm using
this to generate random data that mimics the statistical properties of
those objects in the real world. For example, if I'm generating last
names for say an instance of class Person for Russia, chances are you
should very rarely get any "St-Jean" as opposed to "Ivanov" so this
collection samples those last names accordingly based on their
probabilities. So all that SamplingSet does is return a random object
that respects the statistical probability of such an occurrence.
I'm working on a few optimizations as I'm having to deal with sets of a
few million items and right now, the 4M mark seems to be hard to break
without paying a high price. But I should be done withing a week or so.
On 2020-01-30 09:19, Cyril Ferlicot wrote:
> In Moose there is a project call "Hashtable". The goal of this project
> is to provide faster collections when they have a big number of
> I think this is something that can be useful for more than just Moose
> so I extracted it to pharo-contributions so that we can use it and
> improve it more.
> https://github.com/pharo-contributions/Hashtable >
> Have a nice day.
Yahoo! Messenger: bstjean
"A standpoint is an intellectual horizon of radius zero". (A. Einstein)