Dear List, This may not be rails unique but what is the best practice to validate the uniqueness of has_many associations. In particular the models are class Word end class PhraseComponent belongs_to :word belongs_to :dictionary_entry acts_as_list :scope => :dictionary_entry end class DictionaryEntry has_many :phrase_components has_many :words, :through => :phrase_components end Dictionary entries (words and phrases) are composed of words (which are uniq) and I want to check the same phrase occurs only once in the DictionaryEntry table. Is there a better or more efficient way to do this than create and update a footprint of the word associations on dictionary entries? This means I would have a string field in the dictionary_entries table which is a concatenation of the associated words and would check the uniquness of that. If we generalize the problem, we need to assure uniqueness of lists of arbitrary objects (where a list is via a join table), in that case the only representation that I can create on the list is the concatenation of ids of list items and check the uniqueness of this? I begin to think there must be a more canonical way to handle this. Any help appreciated Viktor I found someone asking the same question but no reply followed. http://groups.google.com/group/rubyonrails-talk/br...
on 2009-01-06 15:03
on 2009-01-07 18:50
Hello, I am still trying to find a solution to this. Please feel free to throw any ideas at me or help me find the appropriate forum. Thank you V
on 2009-01-07 19:45
tron wrote: > class Word > end > > class PhraseComponent > belongs_to :word > belongs_to :dictionary_entry > acts_as_list :scope => :dictionary_entry > end > > class DictionaryEntry > has_many :phrase_components > has_many :words, :through => :phrase_components > end > Well from the database side of thing that's easily solved by adding a unique index across the two foreign keys in the join table. That would definitely prevent duplicates in the join table. CREATE UNIQUE INDEX idx_word_dictionary_entry ON phrase_components (word_id, dictionary_entry_id); From the Rails side you might be able to use something like the following example: class TeacherSchedule < ActiveRecord::Base validates_uniqueness_of :teacher_id, :scope => [:semester_id, :class_id] end Maybe (untested): class PhraseComponent < ActiveRecord::Base validates_uniqueness_of :word_id, :scope => :dictionary_entry validates_uniqueness_of :dictionary_entry_id, :scope => :word_id end Note: This may be an expensive method for validating uniqueness! In any of the cases mentioned you'll still have issues related to "Optimistic Concurrency Control:" http://en.wikipedia.org/wiki/Optimistic_concurrency_control This section of the Rails documentation has a good explanation of these issues and some options on how to solve them: http://www.railsbrain.com/api/rails-2.2.2/doc/inde...
on 2009-01-07 21:58
Hello Robert, You misunderstood my problem I guess. Think of bora as a unique word, with id 1. there is a dictionary entry for bora bora with id 2 then the phrase components representing this association are not unique in word_id X dictionary_id (both are 1,2) only word_id and dictionary_id AND list index (position) meaning there can only be one word on the same position of a list. The real problem is that I want to make sure bora bora is entered only once in the dictionary. For those who like more examples, think of travel itineraries (which are a list of cities, say New York - London - Paris - Budapest) Cities are a table, Itineraries are another and there is a join model (Visits) which acts as a list. How do I make sure an itinerary is created only once? Thanks a million all Vik On Jan 7, 5:45 pm, Robert W. <firstname.lastname@example.org>