如何在嵌套数组/哈希上使用ruby uniq

Question

I am trying to call the uniq method on the follow json so that it would only return unique result base on employee_id

# Json array
a ={
  results: [
   {
    employee: {
      name: "A",
      employee_id: "A-00016",
      title: 1
     }
   },{
    employee: {
      name: "A",
      employee_id: "A-00016",
      title: 2
     }
   },{
    employee: {
      name: "C",
      employee_id: "C-00017",
      title: 3
     }
    }
   ]
  }



# Calling uniq on a
a.uniq { |p| p.values_at(:employee_id) }

However, I am only getting this result

{
  results: [
   {
    employee: {
      name: "A",
      employee_id: "A-00016",
      title: 1
     }
    }
   ]
  }

Instead of what I want

{
  results: [
   {
    employee: {
      name: "A",
      employee_id: "A-00016",
      title: 1
     },{
    employee: {
      name: "C",
      employee_id: "C-00017",
      title: 3
     }
    }
   ]
  }

Am I using the correct method to output the result I want?

Answer 1

With uniq :

input[:results].uniq { |e| e[:employee][:employee_id] }
#⇒ [
#    {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>"1"}},
#    {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>"3"}}]

But I believe there should be some condition applied on what to choose from siblings having the same id . The code below selects the one, having max title value:

input[:results].group_by { |e| e[:employee][:employee_id] }
               .map { |_, v| v.max_by { |e| e[:employee][:title].to_i } }
#⇒ [
#    {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>"2"}},
#    {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>"3"}}]

Answer 2

Here is one way to do this, in order to return the modified input hash , we can use uniq! which will modify the array a[:results] in place. We use dup to duplicate the hash a to preserve it, and then use tap to operate on duplicated hash.

r = a.dup.tap do |h|
  h[:results].uniq! do |h|
    h[:employee][:employee_id]
  end
end

#=> {:results=>
#    [
#     {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>1}},
#     {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}
#    ]
#   }

Answer 3

def selection_criterion(h)
  h[:title].to_i
end

{results: a[:results].group_by {|h| h[:employee][:employee_id]}.
                      values.
                      map {|arr| arr.max_by {|h|  selection_criterion(h[:employee])}}}
   #=> {:results=>
   #     [{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}},
   #      {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}]}        

Define selection_criterion as desired, and possible change max_by to min_by .

The steps are as follows.

 b = a[:results]
   # => [{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>1}},
   #     {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}},
   #     {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}] 
 c = b.group_by { |h| h[:employee][:employee_id] }
   #=> {"A-00016"=>[{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>1}},
   #                {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}}],
   #    "C-00017"=>[{:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}]}
 d = c.values
   #=> [[{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>1}},
   #     {:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}}],
   #    [{:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}]] 
 e = d.map { |arr| arr.max_by { |h| selection_criterion(h[:employee]) } }
   #=> [{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}},
   #    {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}]
 { results: e }
   #=> {:results=>
   #     [{:employee=>{:name=>"A", :employee_id=>"A-00016", :title=>2}},
   #      {:employee=>{:name=>"C", :employee_id=>"C-00017", :title=>3}}]}