Re: Array.to_hash

Nitay wrote:

> I like your from_assoc idea, and support the notion that it should be 
> added.
> What was the status of it?

It's not added in 1.9 yet. I'm not even sure if I posted a RCR or RFE 
for this. If so I didn't get much feedback.

> I also think a specific one for array (or Enumerable mixin maybe?) is 
> useful
> because it is very clear cut and understandable since it is specific to one
> class.

Hm, I think that checking that the passed object is indeed in the form 
of [[x, y], ...] would be a good idea. It would probably make sense to 
check both the inner and outer arrays for being direct or indirect 
instances of Array. As far as I understand Array#lassoc and #rassoc 
already have those checks.

> As far as I know, there are not many ruby from_X methods, whereas
> there are many to_Y ones, which I believe makes it more immediately 
> apparent what the method will do.

That is true, but Array#assoc_to_hash seems a bit odd to me. It's 
definitely not Array#to_hash because it doesn't work with all arrays. I 
think .to_hash would work like this:

   ["foo", "bar"].to_hash # => { 0 => "foo", 1 => "bar" }

But I don't think that such a .to_hash is necessary. I think 
Hash.from_assoc is more important because it is commonly needed. And 
most of the time it is implemented as Array[*assoc.flatten] which has 
problems because flatten is recursive.

-- 
http://flgr.0x42.net/