Question

I need to write a validation to make sure that new records created in our application do not have duplicate names. I’m told that validates_uniqueness will not work in this instance because the application already uses it to validate the creation of the original name. This validation will ensure that when the record is updated with a new name, that it too, is original…it’s sort of an anti-scope validation. The query also needs to be case insensitive so I will need to write it using MYSQL’s UPPER() function, which is why I’ll need to use an array query and not a hash.

Here’s a sketch of the pseudo code.

Query the CUSTOMER database.
Check to ensure the UPDATED_RECORD to see if its NAME is unique.
If it’s NOT unique check to see if it’s ID belongs to record that’s being updated.
If it is, that’s OK as it’s the same ID and should be allowed to keep the same name.
If the query finds the same name that has a different ID associated with it, then an error message should be displayed such as “This Name is already in use by XXXX.”

I’m fairly new to Rails and have been having a bear of a time getting this structured correctly.

Here’s some clarification on why I don’t think validates_uniqueness will work here.

I think the issue is having multiple versions of the same record, as long as updated record is a new version of the same record (it has the same record ID), then that’s valid. But if a new record version is updated with a name that’s already used by an existing record (with a different record ID, then that shouldn’t be allowed. For example, having two term version records, {:record_id => 100, :name => “Test”} & {:record_id => 100, :name => “Test”}, should be allowed since those two records are versions of the same term. But having two records, {:record_id => 100, :name => “Test”} & {:record_id => 201, :name => “Test”}, should not be allowed since that would result in two versions sharing the same name but associated with different terms.

Answer 1

I don’t understand why you can’t use validates_uniqueness. You said it’s already used to validate the creation of the original name, which I am guessing is a column different than the one you want to now check, but that doesn’t matter.

class MyModel < ActiveRecord::Base
    validates_uniqueness_of :original_name
    validates_uniqueness_of :new_name, :case_sensitive => false
end

When you create a new record, or update an existing one, these validations will pass. You can also set a uniqueness validation in the migration by adding an index on it. This gives extra security.

EDIT

Regarding your clarification, it might be easier to add a custom validation, as follows:

class MyModel < ActiveRecord::Base
    validate :check_names

    private

    def check_names
        existing = MyModel.find_by_name(self.name)
        unless existing.nil?
            if existing.record_id != self.record_id
                self.errors[:base] << "This name already exists for a different record_id"
            end
        end
    end
end