Project

General

Profile

Actions

Misc #16436

open

hash missing #last method, make it not so consistent (it has #first)

Added by zw963 (Wei Zheng) about 5 years ago. Updated almost 4 years ago.

Status:
Open
Assignee:
-
[ruby-core:96339]

Description

(pry):main(0)> h = {x:1, y:2, z:3}
{
    :x => 1,
    :y => 2,
    :z => 3
}
(pry):main(0)> h.first
[
    :x,
    1
]
(pry):main(0)> h.first.first
:x
(pry):main(0)> h.first.last
1

last method not exist.

(pry):main(0)> h.last

Exception: NoMethodError: undefined method `last' for {:x=>1, :y=>2, :z=>3}:Hash

We have to use #to_a to make last working.

(pry):main(0)> h.to_a.last
[
    :z,
    3
]

(pry):main(0)> h.to_a.last.first
:z
(pry):main(0)> h.to_a.last.last
3

Updated by zverok (Victor Shepelev) about 5 years ago

Hash doesn't have #first either :)

Hash.instance_method(:first)
# => #<UnboundMethod: Hash(Enumerable)#first(*)>
#                          ^^^^^^^^^^

The first method existence is just because of hash being a Enumerable, not because "first pair(s) of hash" seen as having some frequently used meaning.

(As a side note, Ruby probably could've had BidirectionalEnumerable for collections which can be enumerated both directions without side-effects; this one would be a good place to have #last method.)

Updated by sawa (Tsuyoshi Sawada) about 5 years ago

What is this issue? Is it just a note to everyone, or are you claiming this to be a bug, or is it a feature request?

Updated by shevegen (Robert A. Heiler) about 5 years ago

sawa wrote:

What is this issue? Is it just a note to everyone, or are you
claiming this to be a bug, or is it a feature request?

I think it could only possibly be a feature request since it
is not a bug. But you are uite right that it is difficult to
infer from the issue description alone.

To the topic itself: I can see some points, in particular with
hashes being ordered, where you could use .first and .last
similar as how could be done to Array. I once wondered whether
we should also have .second, .third and so forth, but one
problem I had with that is that this becomes less and less
likely to use; for example, if you have a large array of
array, how often would you use .eight (or [7]) for that
matter? Probably not that often. I do however have found
.second and .third in use now and then.

Of course we can use [] instead anyway, but the code layout
was sometimes strange. Such as in:

first_element  = array[0]
last_element   = array[-1]
second_element = array[1]

versus

first_element  = array.first
last_element   = array.last
second_element = array[1] # <- this one looks a bit weird if used with .first and .last

However had, I think I have not created an issue request. I assume
there may have been reasons why we do not have .second or
similar by default.

Note that the name BidirectionalEnumerable is quite weird too. This
is how we may end up with things such as HashWithIndifferentAccess.
That's an even stranger name ... ;)

Sorry for digressing. May have to see when zw963 can reply and clarify.

Updated by zw963 (Wei Zheng) about 5 years ago

sawa (Tsuyoshi Sawada) wrote:

What is this issue? Is it just a note to everyone, or are you claiming this to be a bug, or is it a feature request?

Sorry, this issue not a bug, i knew Hash#first come from Enumerable, the reason i issue it is: i love ruby language, use it days and days,
when encountering some inconsistencies things, i expect to give some feedback, and what i thought.

So, for this issue, the question is:

why Array include Enumerable module, and implement Array#first Array#last method, if we should do same things for Hash?
make this language more elegant?

Actions #5

Updated by zw963 (Wei Zheng) almost 5 years ago

  • Subject changed from hash missing #last method, make it not so consistent (it has first) to hash missing #last method, make it not so consistent (it has #first)

Updated by fun_with_eval (Fabian Stillhart) almost 4 years ago

This seems to be a duplicate of my old feature request: https://bugs.ruby-lang.org/issues/12165

Actions

Also available in: Atom PDF

Like0
Like0Like0Like0Like0Like0Like0