Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)? I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
but that doesn't look pretty at all in my opinion. I want to just
write like this:
newhash = oldhash.map { |k,v| some_operation(v) }
I finally got around to change this behaviour for my own code, but are
all Ruby users supposed to invent this wheel on their own? Wouldn't it
be better if Hash#map behaved like this? Or is there something I am
missing?
class Hash
def hashmap
self.inject({}) do |newhash, (k,v)|
newhash[k] = yield(k, v)
newhash
end
end
end
"I have never let my schooling interfere with my education" - Mark Twain
···
On Wed, Apr 29, 2009 at 8:00 AM, Fredrik <fredjoha@gmail.com> wrote:
Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)? I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
but that doesn't look pretty at all in my opinion. I want to just
write like this:
newhash = oldhash.map { |k,v| some_operation(v) }
I finally got around to change this behaviour for my own code, but are
all Ruby users supposed to invent this wheel on their own? Wouldn't it
be better if Hash#map behaved like this? Or is there something I am
missing?
class Hash
def hashmap
self.inject({}) do |newhash, (k,v)|
newhash[k] = yield(k, v)
newhash
end
end
end
On Wed, Apr 29, 2009 at 2:00 PM, Fredrik <fredjoha@gmail.com> wrote:
Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)? I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
but that doesn't look pretty at all in my opinion. I want to just
write like this:
newhash = oldhash.map { |k,v| some_operation(v) }
I finally got around to change this behaviour for my own code, but are
all Ruby users supposed to invent this wheel on their own? Wouldn't it
be better if Hash#map behaved like this? Or is there something I am
missing?
class Hash
def hashmap
self.inject({}) do |newhash, (k,v)|
newhash[k] = yield(k, v)
newhash
end
end
end
class Hash
def new_map
result = self.map do |k,v|
r = yield v
[k,r]
end
Hash[*result.flatten]
end
end
a = {}
a[2] = 4
a[3] = 6
b=a.new_map do |v|
v+2
end
puts b.class
puts b.inspect
yields:
Hash
{2=>6, 3=>8}
Does this help?
Jayanth
···
On Wed, Apr 29, 2009 at 11:30 AM, Fredrik <fredjoha@gmail.com> wrote:
Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)? I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
but that doesn't look pretty at all in my opinion. I want to just
write like this:
newhash = oldhash.map { |k,v| some_operation(v) }
I finally got around to change this behaviour for my own code, but are
all Ruby users supposed to invent this wheel on their own? Wouldn't it
be better if Hash#map behaved like this? Or is there something I am
missing?
class Hash
def hashmap
self.inject({}) do |newhash, (k,v)|
newhash[k] = yield(k, v)
newhash
end
end
end
Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)? I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
but that doesn't look pretty at all in my opinion. I want to just
write like this:
newhash = oldhash.map { |k,v| some_operation(v) }
class Hash
def hashbackmap
result = {}
self.each do |key, val|
result[key] = yield val
end
result
end
end
h = {
"red" => 10,
"blue" => 20,
"green" => 30
}
p h.hashbackmap {|x| x*2}
--output:--
{"green"=>60, "blue"=>40, "red"=>20}
I HAVE A PATENT!! My licensing fees are cheap: $10 mth.
Is there any good reason why Hash#map does not give back a Hash (Ruby
1.8 but same in 1.9 as far as I know)?
It wouldn't make sense for Hash#map to return a Hash since you are not required to transform it into key-value pairs:
>> {1 => "one", 2 => "two"}.map { |k, v| "#{k} is #{v}" }
=> ["1 is one", "2 is two"]
I often find myself writing
these kind of things:
newhash = oldhash.inject({}) { |h,(k,v)| h[k] = some_operation(v); h }
1.9 adds each_with_object() for this very usage. You don't want the return value of the block to carry forward, but instead the object you started with. That's what each_with_object() does. So your code becomes the following in Ruby 1.9:
Becuase #map comes from form Enumerable (and is an alias for collect)
Well ok, that is a reason but it's not a good motivation why it HAS to
be like that. What use does one really have of getting an Array (with
elements in an undefined order) from a Hash#map ?
Thanks! It's not a very big improvement over inject though. I don't
see the use of Hash#map -> Array so if it was up to me, Hash#map would
give a Hash. If one really wants an Array it seems more reasonable to
use something like
hash.to_a{|k,v| "#{k} is #{v}" }
Maybe it's just me, but when I read "map" I am thinking "map this
block to each element and give back the same structure" (i.e.
Array#map gives an Array and Hash#map gives a Hash).
(I know map is just an alias for collect, but I never understood why
"collect" is a good name for this method.)
···
On 29 Apr, 21:39, James Gray <ja...@grayproductions.net> wrote:
On Apr 29, 2009, at 1:00 AM, Fredrik wrote:
> Is there any good reason why Hash#map does not give back a Hash (Ruby
> 1.8 but same in 1.9 as far as I know)?
It wouldn't make sense for Hash#map to return a Hash since you are not
required to transform it into key-value pairs:
>> {1 => "one", 2 => "two"}.map { |k, v| "#{k} is #{v}" }
=> ["1 is one", "2 is two"]
> I often find myself writing
> these kind of things:
1.9 adds each_with_object() for this very usage. You don't want the
return value of the block to carry forward, but instead the object you
started with. That's what each_with_object() does. So your code
becomes the following in Ruby 1.9: