Remove hash from hash of hashes if value is duplicated

问题

I have got such hash of hashes

params = {collection_permissions_attributes: {'0' => {:collection_id => '1'},
                                                   '2' => {:collection_id => '1'},
                                                   '3' => {:collection_id => '4'}}}

How can I remove completely from collection_permissions_attributes item which hash contains duplicated value? So from hash above I would like to have:

params = { collection_permissions_attributes: {'0' => {:collection_id => '1'},
                                                   '3' => {:collection_id => '4'}}}

Thanks in advance!

回答1:

Here's another way, but it requires Ruby 1.9+, as it depends on the order of the hash's elements.

Code

newams = {collection: ams[:collection].sort.reverse.to_h.invert.invert}
  #=> {:collection=>{"3"=>{:collection_id=>"4"}, "0"=>{:collection_id=>"1"}}}

If for some reason it is desired that the hash elements be in key order, that can be achieved by adding a few more operations, as explained at the end of the "explanation" section.

Explanation

ams = {collection: {'0' => {:collection_id => '1'},
                    '2' => {:collection_id => '1'},
                    '3' => {:collection_id => '4'}}}

First, reorder the hash elements in descending key order, so that when two or more elements have the same value, only the element having the smallest key value will be kept. To do this and then reconstitute the hash, three steps are required:

a = ams[:collection].sort
  #=> [["0", {:collection_id=>"1"}], ["2", {:collection_id=>"1"}],
  #    ["3", {:collection_id=>"4"}]] 
b = a.reverse
  #=> [["3", {:collection_id=>"4"}], ["2", {:collection_id=>"1"}],
  #    ["0", {:collection_id=>"1"}]] 
c = b.to_h
  #=> {"3"=>{:collection_id=>"4"}, "2"=>{:collection_id=>"1"},
  #    "0"=>{:collection_id=>"1"}} 

Next, we invert the hash, making the keys values and the values keys. Because hashes cannot have duplicate keys, this eliminates the duplicates:

d = c.invert
  #=> {{:collection_id=>"4"}=>"3", {:collection_id=>"1"}=>"0"}

Lastly, invert back:

e = d.invert
  #=> {"3"=>{:collection_id=>"4"}, "0"=>{:collection_id=>"1"}} 

The hash elements can reordered in key order (should that be desired) like this:

e.to_a.reverse.to_h
  #=> {"0"=>{:collection_id=>"1"}, "3"=>{:collection_id=>"4"}} 

回答2:

I can think of as below :

params = {    :employee =>
              {collection_permissions_attributes: {'0' => {:collection_id => '1'},
                                                   '2' => {:collection_id => '1'},
                                                   '3' => {:collection_id => '4'}}}
        }

hash = Hash[params[:employee][:collection_permissions_attributes].group_by do |k,v|
  v[:collection_id]
end.map { |_,v| v.first }]

params[:employee][:collection_permissions_attributes] = hash

output

{:employee => 
    {:collection_permissions_attributes => {
                                             "0"=>{:collection_id=>"1"}, 
                                             "3"=>{:collection_id=>"4"}
                                           }
    }
}

来源：https://stackoverflow.com/questions/22557456/remove-hash-from-hash-of-hashes-if-value-is-duplicated