merging two arrays of hashes wisely - ruby-on-rails

Question

I am trying to combine two arrays of hashes arr1 and arr2:
arr1 = [{"id"=>1, "a"=>1, "c"=>2}, {"id"=>2, "a"=>1}]
arr2 = [{"id"=>1, "a"=>10, "b"=>20}, {"id"=>3, "b"=>2}]
And I want the result to include all elements in both arrays, but the ones that have the same value for the "id" key, should be merged so that if a key exists in both hashes, it should be selected from arr2, otherwise, it just picks the value from any hash that the key exists in. So the combination of the example above would be:
combined = [
{"id"=>1, "a"=>10, "b"=>20, "c"=>2}, # "id"=>1 exists in both, so they are merged
{"id"=>2, "a"=>1},
{"id"=>3, "b"=>2}
]
The code below works, but I am new to Ruby and I am sure there is a better way to do this. Can you provide a more ruby-ic way?
combined = []
# merge items that exist in both and add to combined
arr1.each do |a1|
temp = arr2.select {|a2| a2["id"] == a1["id"]}[0]
if temp.present?
combined << temp.reverse_merge(a1)
end
end
# Add items that exist in arr1 but not in arr2
arr1.each do |a1|
if arr2.pluck("id").exclude? a1["id"]
combined << a1
end
end
# Add items that exist in arr2 but not in arr1
arr2.each do |a2|
if arr1.pluck("id").exclude? a2["id"]
combined << a2
end
end

Answer 1

I assume that no two elements (hashes) of arr1, g and h, have the property that g["id"] == h["id"].
In this case one could write:
(arr1 + arr2).each_with_object(Hash.new { |h,k| h[k] = {} }) { |g,h|
h[g["id"]].update(g) }.values
#=> [{"id"=>1, "a"=>10, "c"=>2, "b"=>20}, {"id"=>2, "a"=>1},
# {"id"=>3, "b"=>2}]
Note that:
(arr1 + arr2).each_with_object(Hash.new { |h,k| h[k] = {} }) { |g,h|
h[g["id"]].update(g) }
#=> {1=>{"id"=>1, "a"=>10, "c"=>2, "b"=>20}, 2=>{"id"=>2, "a"=>1},
# 3=>{"id"=>3, "b"=>2}}
If a hash is defined:
h = Hash.new { |h,k| h[k] = {} }
then, possibly after keys have been added to h, if h does not have a key k, h[k] = {} is executed and the empty hash is returned. See the form of Hash::new that takes a block. See also Hash#update (aka Hash#merge!).
One may alternatively write:
(arr1 + arr2).each_with_object({}) { |g,h| (h[g["id"]] ||= {}).update(g) }.values
#=> {1=>{"id"=>1, "a"=>10, "c"=>2, "b"=>20}, 2=>{"id"=>2, "a"=>1},
# 3=>{"id"=>3, "b"=>2}}
Another way is to use Emumerable#group_by, where the grouping is on the value of the key "id":
(arr1 + arr2).group_by { |h| h["id"] }.values.map { |a| a.reduce(&:merge) }
#=> [{"id"=>1, "a"=>10, "c"=>2, "b"=>20}, {"id"=>2, "a"=>1}, {"id"=>3, "b"=>2}]