0

我有这些模型：

class Relationship < ActiveRecord::Base
  belongs_to :student
  belongs_to :guardian
end

class Student < ActiveRecord::Base
  has_many :relationships
  has_many :guardians, through: :relationships
end

class Guardian < ActiveRecord::Base
  has_many :relationships
  has_many :students, through: :relationships
end

我不希望任何监护人与一名学生多次发生关系。这可以通过验证来防止，但是假设为时已晚，并且在某些情况下会发生这种情况。例如，g监护人在哪里：

g.relationships

[
    [0] #<Relationship:0x0000000bc33650> {
                          :id => 40321,
        :relationship_type_id => 2,
                  :student_id => 41700,
                 :guardian_id => 45820,
                  :created_at => Tue, 23 Apr 2013 17:44:29 UTC +00:00,
                  :updated_at => Tue, 23 Apr 2013 17:44:29 UTC +00:00,
    },
    [1] #<Relationship:0x0000000bc32e80> {
                          :id => 40923,
        :relationship_type_id => 2,
                  :student_id => 41700,
                 :guardian_id => 45820,
                  :created_at => Tue, 23 Apr 2013 18:58:46 UTC +00:00,
                  :updated_at => Tue, 23 Apr 2013 18:58:46 UTC +00:00,
    }
]

如您所见，这两个关系共享相同的 student_id。我想知道是否有办法从数据库中删除重复的关系。我尝试了以下代码行无济于事：

g.relationships.uniq!(&:student_id)
g.update_attributes(relatioships: g.relationships.uniq!(&:student_id))

我怎么解决这个问题。谢谢！

4

1 回答 1

0 
subquery = Relationship.select("student_id, guardian_id").
             group(:student_id, :guardian_id).
             having("count(*) > 1").to_sql
groups = Relationship.joins("JOIN (#{subquery}) sub ON relationships.student_id = sub.student_id AND relationships.guardian_id = sub.guardian_id").
                      group_by{|r| [r.student_id, r.guardian_id] }

groups.values.each do |duplicates|
  duplicates.drop(1).each(&:destroy)
end
于 2013-04-25T17:19:28.827 回答