I have a hash that uses array as its key. When I change the array, the hash can no longer get the corresponding key and value:
1.9.3p194 :016 > a = [1, 2]
=> [1, 2]
1.9.3p194 :017 > b = { a => 1 }
=> {[1, 2]=>1}
1.9.3p194 :018 > b[a]
=> 1
1.9.3p194 :019 > a.delete_at(1)
=> 2
1.9.3p194 :020 > a
=> [1]
1.9.3p194 :021 > b
=> {[1]=>1}
1.9.3p194 :022 > b[a]
=> nil
1.9.3p194 :023 > b.keys.include? a
=> true
What am I doing wrong?
Update:
OK. Use a.clone is absolutely one way to deal with this problem.
What if I want to change "a" but still use "a" to retrieve the corresponding value (since "a" is still one of the keys) ?
The #rehash method will recalculate the hash, so after the key changes do:
b.rehash
TL;DR: consider Hash#compare_by_indentity
You need to decide if you want the hash to work by array value or array identity.
By default arrays .hash and .eql? by value, which is why changing the value confuses ruby. Consider this variant of your example:
pry(main)> a = [1, 2]
pry(main)> a1 = [1]
pry(main)> a.hash
=> 4266217476190334055
pry(main)> a1.hash
=> -2618378812721208248
pry(main)> h = {a => '12', a1 => '1'}
=> {[1, 2]=>"12", [1]=>"1"}
pry(main)> h[a]
=> "12"
pry(main)> a.delete_at(1)
pry(main)> a
=> [1]
pry(main)> a == a1
=> true
pry(main)> a.hash
=> -2618378812721208248
pry(main)> h[a]
=> "1"
See what happened there?
As you discovered, it fails to match on the a key because the .hash value under which it stored it is outdated [BTW, you can't even rely on that! A mutation might result in same hash (rare) or different hash that lands in the same bucket (not so rare).]
But instead of failing by returning nil, it matched on the a1 key.
See, h[a] doesn't care at all about the identity of a vs a1 (the traitor!). It compared the current value you supply — [1] with the value of a1 being [1] and found a match.
That's why using .rehash is just band-aid. It will recompute the .hash values for all keys and move them to the correct buckets, but it's error-prone, and may also cause trouble:
pry(main)> h.rehash
=> {[1]=>"1"}
pry(main)> h
=> {[1]=>"1"}
Oh oh. The two entries collapsed into one, since they now have the same value (and which wins is hard to predict).
Solutions
One sane approach is embracing lookup by value, which requires the value to never change. .freeze your keys. Or use .clone/.dup when building the hash, and feel free to mutate the original arrays — but accept that h[a] will lookup the current value of a against the values preserved from build time.
The other, which you seem to want, is deciding you care about identity — lookup by a should find a whatever its current value, and it shouldn't matter if many keys had or now have the same value.
How?
Object hashes by identity. (Arrays don't because types that .== by value tend to also override .hash and .eql? to be by value.) So one option is: don't use arrays as keys, use some custom class (which may hold an array inside).
But what if you want it to behave directly like a hash of arrays? You could subclass Hash, or Array but it's a lot of work to make everything work consistently. Luckily, Ruby has a builtin way: h.compare_by_identity switches a hash to work by identity (with no way to undo, AFAICT). If you do this before you insert anything, you can even have distinct keys with equal values, with no confusion:
[39] pry(main)> x = [1]
=> [1]
[40] pry(main)> y = [1]
=> [1]
[41] pry(main)> h = Hash.new.compare_by_identity
=> {}
[42] pry(main)> h[x] = 'x'
=> "x"
[44] pry(main)> h[y] = 'y'
=> "y"
[45] pry(main)> h
=> {[1]=>"x", [1]=>"y"}
[46] pry(main)> x.push(7)
=> [1, 7]
[47] pry(main)> y.push(7)
=> [1, 7]
[48] pry(main)> h
=> {[1, 7]=>"x", [1, 7]=>"y"}
[49] pry(main)> h[x]
=> "x"
[50] pry(main)> h[y]
=> "y"
Beware that such hashes are counter-intuitive if you try to put there e.g. strings, because we're really used to strings hashing by value.
Hashes use their key objects' hash codes (a.hash) to group them. Hash codes often depend on the state of the object; in this case, the hash code of a changes when an element has been removed from the array. Since the key has already been inserted into the hash, a is filed under its original hash code.
This means you can't retrieve the value for a in b, even though it looks alright when you print the hash.
You should use a.clone as key
irb --> a = [1, 2]
==> [1, 2]
irb --> b = { a.clone => 1 }
==> {[1, 2]=>1}
irb --> b[a]
==> 1
irb --> a.delete_at(1)
==> 2
irb --> a
==> [1]
irb --> b
==> {[1, 2]=>1} # STILL UNCHANGED
irb --> b[a]
==> nil # Trivial, since a has changed
irb --> b.keys.include? a
==> false # Trivial, since a has changed
Using a.clone will make sure that the key is unchanged even when we change a later on.
As you have already said, the trouble is that the hash key is the exact same object you later modify, meaning that the key changes during program execution.
To avoid this, make a copy of the array to use as a hash key:
a = [1, 2]
b = { a.clone => 1 }
Now you can continue to work with a and leave your hash keys intact.
In Ruby, given an array in one of the following forms...
[apple, 1, banana, 2]
[[apple, 1], [banana, 2]]
...what is the best way to convert this into a hash in the form of...
{apple => 1, banana => 2}
Simply use Hash[*array_variable.flatten]
For example:
a1 = ['apple', 1, 'banana', 2]
h1 = Hash[*a1.flatten(1)]
puts "h1: #{h1.inspect}"
a2 = [['apple', 1], ['banana', 2]]
h2 = Hash[*a2.flatten(1)]
puts "h2: #{h2.inspect}"
Using Array#flatten(1) limits the recursion so Array keys and values work as expected.
NOTE: For a concise and efficient solution, please see Marc-André Lafortune's answer below.
This answer was originally offered as an alternative to approaches using flatten, which were the most highly upvoted at the time of writing. I should have clarified that I didn't intend to present this example as a best practice or an efficient approach. Original answer follows.
Warning! Solutions using flatten will not preserve Array keys or values!
Building on #John Topley's popular answer, let's try:
a3 = [ ['apple', 1], ['banana', 2], [['orange','seedless'], 3] ]
h3 = Hash[*a3.flatten]
This throws an error:
ArgumentError: odd number of arguments for Hash
from (irb):10:in `[]'
from (irb):10
The constructor was expecting an Array of even length (e.g. ['k1','v1,'k2','v2']). What's worse is that a different Array which flattened to an even length would just silently give us a Hash with incorrect values.
If you want to use Array keys or values, you can use map:
h3 = Hash[a3.map {|key, value| [key, value]}]
puts "h3: #{h3.inspect}"
This preserves the Array key:
h3: {["orange", "seedless"]=>3, "apple"=>1, "banana"=>2}
The best way is to use Array#to_h:
[ [:apple,1],[:banana,2] ].to_h #=> {apple: 1, banana: 2}
Note that to_h also accepts a block:
[:apple, :banana].to_h { |fruit| [fruit, "I like #{fruit}s"] }
# => {apple: "I like apples", banana: "I like bananas"}
Note: to_h accepts a block in Ruby 2.6.0+; for early rubies you can use my backports gem and require 'backports/2.6.0/enumerable/to_h'
to_h without a block was introduced in Ruby 2.1.0.
Before Ruby 2.1, one could use the less legible Hash[]:
array = [ [:apple,1],[:banana,2] ]
Hash[ array ] #= > {:apple => 1, :banana => 2}
Finally, be wary of any solutions using flatten, this could create problems with values that are arrays themselves.
Update
Ruby 2.1.0 is released today. And I comes with Array#to_h (release notes and ruby-doc), which solves the issue of converting an Array to a Hash.
Ruby docs example:
[[:foo, :bar], [1, 2]].to_h # => {:foo => :bar, 1 => 2}
Edit: Saw the responses posted while I was writing, Hash[a.flatten] seems the way to go.
Must have missed that bit in the documentation when I was thinking through the response. Thought the solutions that I've written can be used as alternatives if required.
The second form is simpler:
a = [[:apple, 1], [:banana, 2]]
h = a.inject({}) { |r, i| r[i.first] = i.last; r }
a = array, h = hash, r = return-value hash (the one we accumulate in), i = item in the array
The neatest way that I can think of doing the first form is something like this:
a = [:apple, 1, :banana, 2]
h = {}
a.each_slice(2) { |i| h[i.first] = i.last }
You can also simply convert a 2D array into hash using:
1.9.3p362 :005 > a= [[1,2],[3,4]]
=> [[1, 2], [3, 4]]
1.9.3p362 :006 > h = Hash[a]
=> {1=>2, 3=>4}
Summary & TL;DR:
This answer hopes to be a comprehensive wrap-up of information from other answers.
The very short version, given the data from the question plus a couple extras:
flat_array = [ apple, 1, banana, 2 ] # count=4
nested_array = [ [apple, 1], [banana, 2] ] # count=2 of count=2 k,v arrays
incomplete_f = [ apple, 1, banana ] # count=3 - missing last value
incomplete_n = [ [apple, 1], [banana ] ] # count=2 of either k or k,v arrays
# there's one option for flat_array:
h1 = Hash[*flat_array] # => {apple=>1, banana=>2}
# two options for nested_array:
h2a = nested_array.to_h # since ruby 2.1.0 => {apple=>1, banana=>2}
h2b = Hash[nested_array] # => {apple=>1, banana=>2}
# ok if *only* the last value is missing:
h3 = Hash[incomplete_f.each_slice(2).to_a] # => {apple=>1, banana=>nil}
# always ok for k without v in nested array:
h4 = Hash[incomplete_n] # or .to_h => {apple=>1, banana=>nil}
# as one might expect:
h1 == h2a # => true
h1 == h2b # => true
h1 == h3 # => false
h3 == h4 # => true
Discussion and details follow.
Setup: variables
In order to show the data we'll be using up front, I'll create some variables to represent various possibilities for the data. They fit into the following categories:
Based on what was directly in the question, as a1 and a2:
(Note: I presume that apple and banana were meant to represent variables. As others have done, I'll be using strings from here on so that input and results can match.)
a1 = [ 'apple', 1 , 'banana', 2 ] # flat input
a2 = [ ['apple', 1], ['banana', 2] ] # key/value paired input
Multi-value keys and/or values, as a3:
In some other answers, another possibility was presented (which I expand on here) – keys and/or values may be arrays on their own:
a3 = [ [ 'apple', 1 ],
[ 'banana', 2 ],
[ ['orange','seedless'], 3 ],
[ 'pear', [4, 5] ],
]
Unbalanced array, as a4:
For good measure, I thought I'd add one for a case where we might have an incomplete input:
a4 = [ [ 'apple', 1],
[ 'banana', 2],
[ ['orange','seedless'], 3],
[ 'durian' ], # a spiky fruit pricks us: no value!
]
Now, to work:
Starting with an initially-flat array, a1:
Some have suggested using #to_h (which showed up in Ruby 2.1.0, and can be backported to earlier versions). For an initially-flat array, this doesn't work:
a1.to_h # => TypeError: wrong element type String at 0 (expected array)
Using Hash::[] combined with the splat operator does:
Hash[*a1] # => {"apple"=>1, "banana"=>2}
So that's the solution for the simple case represented by a1.
With an array of key/value pair arrays, a2:
With an array of [key,value] type arrays, there are two ways to go.
First, Hash::[] still works (as it did with *a1):
Hash[a2] # => {"apple"=>1, "banana"=>2}
And then also #to_h works now:
a2.to_h # => {"apple"=>1, "banana"=>2}
So, two easy answers for the simple nested array case.
This remains true even with sub-arrays as keys or values, as with a3:
Hash[a3] # => {"apple"=>1, "banana"=>2, ["orange", "seedless"]=>3, "pear"=>[4, 5]}
a3.to_h # => {"apple"=>1, "banana"=>2, ["orange", "seedless"]=>3, "pear"=>[4, 5]}
But durians have spikes (anomalous structures give problems):
If we've gotten input data that's not balanced, we'll run into problems with #to_h:
a4.to_h # => ArgumentError: wrong array length at 3 (expected 2, was 1)
But Hash::[] still works, just setting nil as the value for durian (and any other array element in a4 that's just a 1-value array):
Hash[a4] # => {"apple"=>1, "banana"=>2, ["orange", "seedless"]=>3, "durian"=>nil}
Flattening - using new variables a5 and a6
A few other answers mentioned flatten, with or without a 1 argument, so let's create some new variables:
a5 = a4.flatten
# => ["apple", 1, "banana", 2, "orange", "seedless" , 3, "durian"]
a6 = a4.flatten(1)
# => ["apple", 1, "banana", 2, ["orange", "seedless"], 3, "durian"]
I chose to use a4 as the base data because of the balance problem we had, which showed up with a4.to_h. I figure calling flatten might be one approach someone might use to try to solve that, which might look like the following.
flatten without arguments (a5):
Hash[*a5] # => {"apple"=>1, "banana"=>2, "orange"=>"seedless", 3=>"durian"}
# (This is the same as calling `Hash[*a4.flatten]`.)
At a naïve glance, this appears to work – but it got us off on the wrong foot with the seedless oranges, thus also making 3 a key and durian a value.
And this, as with a1, just doesn't work:
a5.to_h # => TypeError: wrong element type String at 0 (expected array)
So a4.flatten isn't useful to us, we'd just want to use Hash[a4]
The flatten(1) case (a6):
But what about only partially flattening? It's worth noting that calling Hash::[] using splat on the partially-flattened array (a6) is not the same as calling Hash[a4]:
Hash[*a6] # => ArgumentError: odd number of arguments for Hash
Pre-flattened array, still nested (alternate way of getting a6):
But what if this was how we'd gotten the array in the first place?
(That is, comparably to a1, it was our input data - just this time some of the data can be arrays or other objects.) We've seen that Hash[*a6] doesn't work, but what if we still wanted to get the behavior where the last element (important! see below) acted as a key for a nil value?
In such a situation, there's still a way to do this, using Enumerable#each_slice to get ourselves back to key/value pairs as elements in the outer array:
a7 = a6.each_slice(2).to_a
# => [["apple", 1], ["banana", 2], [["orange", "seedless"], 3], ["durian"]]
Note that this ends up getting us a new array that isn't "identical" to a4, but does have the same values:
a4.equal?(a7) # => false
a4 == a7 # => true
And thus we can again use Hash::[]:
Hash[a7] # => {"apple"=>1, "banana"=>2, ["orange", "seedless"]=>3, "durian"=>nil}
# or Hash[a6.each_slice(2).to_a]
But there's a problem!
It's important to note that the each_slice(2) solution only gets things back to sanity if the last key was the one missing a value. If we later added an extra key/value pair:
a4_plus = a4.dup # just to have a new-but-related variable name
a4_plus.push(['lychee', 4])
# => [["apple", 1],
# ["banana", 2],
# [["orange", "seedless"], 3], # multi-value key
# ["durian"], # missing value
# ["lychee", 4]] # new well-formed item
a6_plus = a4_plus.flatten(1)
# => ["apple", 1, "banana", 2, ["orange", "seedless"], 3, "durian", "lychee", 4]
a7_plus = a6_plus.each_slice(2).to_a
# => [["apple", 1],
# ["banana", 2],
# [["orange", "seedless"], 3], # so far so good
# ["durian", "lychee"], # oops! key became value!
# [4]] # and we still have a key without a value
a4_plus == a7_plus # => false, unlike a4 == a7
And the two hashes we'd get from this are different in important ways:
ap Hash[a4_plus] # prints:
{
"apple" => 1,
"banana" => 2,
[ "orange", "seedless" ] => 3,
"durian" => nil, # correct
"lychee" => 4 # correct
}
ap Hash[a7_plus] # prints:
{
"apple" => 1,
"banana" => 2,
[ "orange", "seedless" ] => 3,
"durian" => "lychee", # incorrect
4 => nil # incorrect
}
(Note: I'm using awesome_print's ap just to make it easier to show the structure here; there's no conceptual requirement for this.)
So the each_slice solution to an unbalanced flat input only works if the unbalanced bit is at the very end.
Take-aways:
Whenever possible, set up input to these things as [key, value] pairs (a sub-array for each item in the outer array).
When you can indeed do that, either #to_h or Hash::[] will both work.
If you're unable to, Hash::[] combined with the splat (*) will work, so long as inputs are balanced.
With an unbalanced and flat array as input, the only way this will work at all reasonably is if the last value item is the only one that's missing.
Side-note: I'm posting this answer because I feel there's value to be added – some of the existing answers have incorrect information, and none (that I read) gave as complete an answer as I'm endeavoring to do here. I hope that it's helpful. I nevertheless give thanks to those who came before me, several of whom provided inspiration for portions of this answer.
Appending to the answer but using anonymous arrays and annotating:
Hash[*("a,b,c,d".split(',').zip([1,2,3,4]).flatten)]
Taking that answer apart, starting from the inside:
"a,b,c,d" is actually a string.
split on commas into an array.
zip that together with the following array.
[1,2,3,4] is an actual array.
The intermediate result is:
[[a,1],[b,2],[c,3],[d,4]]
flatten then transforms that to:
["a",1,"b",2,"c",3,"d",4]
and then:
*["a",1,"b",2,"c",3,"d",4] unrolls that into
"a",1,"b",2,"c",3,"d",4
which we can use as the arguments to the Hash[] method:
Hash[*("a,b,c,d".split(',').zip([1,2,3,4]).flatten)]
which yields:
{"a"=>1, "b"=>2, "c"=>3, "d"=>4}
if you have array that looks like this -
data = [["foo",1,2,3,4],["bar",1,2],["foobar",1,"*",3,5,:foo]]
and you want the first elements of each array to become the keys for the hash and the rest of the elements becoming value arrays, then you can do something like this -
data_hash = Hash[data.map { |key| [key.shift, key] }]
#=>{"foo"=>[1, 2, 3, 4], "bar"=>[1, 2], "foobar"=>[1, "*", 3, 5, :foo]}
Not sure if it's the best way, but this works:
a = ["apple", 1, "banana", 2]
m1 = {}
for x in (a.length / 2).times
m1[a[x*2]] = a[x*2 + 1]
end
b = [["apple", 1], ["banana", 2]]
m2 = {}
for x,y in b
m2[x] = y
end
For performance and memory allocation concerns please check my answer to Rails mapping array of hashes onto single hash where I bench-marked several solutions.
reduce / inject can be the fastest or the slowest solution depending on which method you use it which.
If the numeric values are seq indexes, then we could have simpler ways...
Here's my code submission, My Ruby is a bit rusty
input = ["cat", 1, "dog", 2, "wombat", 3]
hash = Hash.new
input.each_with_index {|item, index|
if (index%2 == 0) hash[item] = input[index+1]
}
hash #=> {"cat"=>1, "wombat"=>3, "dog"=>2}