> For the complete documentation index, see [llms.txt](https://certil-remy.gitbook.io/learn/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://certil-remy.gitbook.io/learn/oop-ruby/untitled-22.md). # Has Many Object ## Objectives * Describe the "has many" relationship between Ruby objects. * Build classes that produce objects with a "belongs-to" and "has-many" relationship. * Explain why we need to associate objects in this way. ## Introduction We know that the programs we write are meant to model real-world environments. This is because the programs we write are designed to carry out real-world jobs and solve real-world problems. Whether you're creating an app that connects users around the world in some kind of social network or writing a program for a major university that manages their course offerings and students, your code will need to be able to realistically map the relationships between different entities. We already know about the "belongs-to" relationship. Let's say we have a `Song` class that produces individual song objects. Each song belongs to the artist that wrote it. We can build that relationship by creating an `attr_accessor` in the `Song` class for `artist`: ``` class Song attr_accessor :artist, :name, :genre def initialize(name, genre) @name = name @genre = genre end end ``` If we also have an `Artist` class that looks like this: ``` class Artist attr_accessor :name def initialize(name) @name = name end end ``` We can set an individual instance of `Song` equal to an instance of the `Artist` class like this: ``` kiki = Song.new("In My Feelings", "hip-hop") drake = Artist.new("Drake") kiki.artist = drake kiki.artist.name # => "Drake" ``` The benefit here is that in setting the `artist=` method equal to a real instance of the `Artist` class, instead of equal to a simple string, we are associating our song to a robust object that has its own attributes and behaviors. For example, in the code above, we are calling the `#name` method on the artist of `kiki`. With method chaining like this, we can do even more with our code. The inverse of the "belongs-to" relationship is the "has-many" relationship. If a song belongs to an artist, then an artist should be able to *have many* songs. This makes sense in the real-world––most musical artists have authored and performed many more than one song. Let's take a closer look. ## The "has-many" Relationship How can we represent an object's "having many" of something? Well, having many of something means you own a collection of that thing. Ruby offers us a great way to store collections of data in list form: arrays. We would like to be able to call: And have returned to us a list, or array, of the songs that Drake has written. A given artist should start, or be initialized, with a songs collection that is empty. Later, we will write a method that adds songs to that collection. ### Initializing with an Empty Collection ``` class Artist attr_accessor :name def initialize(name) @name = name @songs = [] end end ``` Here we set an instance variable, `@songs`, equal to an empty array. Recall that we use instance variables to store the attributes of a given instance of a class. This instance variable is set equal to an empty array because our artist doesn't have any songs yet. Let's write the method that will allow us to add some. ### Adding items to the collection Whose responsibility is it to add a new song to a given artist's collection? Well, at what point in time does an artist add another song to his or her repertoire? When that artist writes a new song. Consequently, it isn't the song's responsibility to add itself to the artist's collection of songs, it is the artist's responsibility to add a new song to their collection. That's why we'll write the method that adds songs to an artist's collection in the `Artist` class: ``` class Artist attr_accessor :name def initialize(name) @name = name @songs = [] end def add_song(song) @songs << song end end ``` Now we can execute the following code: ``` drake = Artist.new("Drake") drake.add_song("In My Feelings") drake.add_song("Hotline Bling") ``` Now we need a method that will allow a given artist to show us all of the songs in their collection. Let's do it. ### Exposing the Collection Let's write an instance method, `#songs`, that we can call on an individual artist to return the list of songs that the artist has. ``` class Artist attr_accessor :name def initialize(name) @name = name @songs = [] end def add_song(song) @songs << song end def songs @songs end end ``` The `#songs` method simply return the `@songs` array, which contains the list of songs that the artist has many of. Let's try it out: ``` drake.songs # => ["In My Feelings", "Hotline Bling"] ``` ### Relating Objects with "belongs to" and "has many" "Wow, those look like interesting songs," you might be thinking. "I wonder what kind of music Drake makes." Well, let's ask `drake` to tell us the genres of the songs he has many of. Oh no! We can't do that because `drake`'s songs are simply a list of strings. We can't ask a plain old string what genre it has—it will have no idea what we are talking about. This is the limitation of one-sided relationships. Just like associating a given song to a string that contains an artist's name instead of to a real `Artist` instance had its drawbacks, so too does associating a given artist to a list of strings. With this setup, we are limited to references to a given artist's songs by their name alone. We cannot associate any further information to an artist's songs or enact any further behavior on an artist's songs. Let's fix this now. Instead of calling the `#add_song` method with an argument of a string, let's call that method with an argument of a real song object: ``` kiki = Song.new("In My Feelings", "hip-hop") hotline = Song.new("Hotline Bling", "pop") drake.add_song(kiki) drake.add_song(hotline) drake.songs # =>[#, #] ``` Great, now our artist has many songs that are real, tangible `Song` instances, not just strings. We can do several useful things with this collection of real song objects, such as iterate over them and collect their genres: ``` drake.songs.collect do |song| song.genre end # => ["hip-hop", "pop"] ``` #### Object Reciprocity Now that we can ask our given artist for his songs, let's make sure that we can ask an individual song for its artist: Although we do have an `attr_accessor` for `artist` in our `Song` class, this particular song doesn't seem to know that it belongs to Drake. That is because our `#add_song` method only accomplished associating the song object to the artist object. Our artist knows it has a collection of songs and knows how to add songs to that collection. But, we didn't tell the song that we added to the artist that it belonged to that artist. Let's fix that now. Telling a song that it belongs to an artist should happen when that song is added to the artist's `@songs` collection. Consequently, we will write the code that accomplishes this inside our `#add_song` method: ``` class Artist attr_accessor :name def initialize(name) @name = name @songs = [] end def add_song(song) @songs << song song.artist = self end def songs @songs end end ``` Let's take a closer look at the code in our `#add_song` method: ``` def add_song(song) @songs << song song.artist = self end ``` Here, we use the `self` keyword to refer to the artist on which we are calling this method. We call the `#artist=` method on the song that is being passed in as an argument and set it equal to `self`––the artist. Let's try calling `#add_song` again: Now, we should be able to ask `kiki` for its artist: ``` kiki.artist.name # => "Drake" ``` We did it! Not only does an artist have many songs, but a song belongs to an artist and we built a method that enacts those associations at the appropriate time. ## Maintaining a Single Source of Truth The `#add_song` method works, but look at it again and see if you might be able to find a flaw in this setup: ``` def add_song(song) @songs << song song.artist = self end ``` With this implementation, we're maintaining this relationship on both the `Song` instance and the `Artist` instance. We've done this so that an artist knows which songs it has, and a song knows the artist it belongs to. However, keeping this information maintained on both sides of the relationship means there are ***two sources of truth***. What happens if we *don't* consistently use the `add_song` method? What if, instead, somewhere along the lines we did something like this: ``` lil_nas_x = Artist.new("Lil Nas X") old_town_road = Song.new("Old Town Road","hip-hop") old_town_road.artist = lil_nas_x old_town_road.artist.name #=> "Lil Nas X" lil_nas_x.songs #=> [] ``` Now, the `Song` instance `old_town_road` is associated with an artist, but `lil_nas_x` ***does not know about*** `old_town_road`. We have multiple sources of truth about artists and their songs, and they're not aligned. A better way to approach this would be to figure out how to maintain our "has-many" / "belongs-to" relationship *on only one side of the relationship*. Think about it this way - imagine we have many artists, each with their own songs. Rather than have each artist keep track of their own songs, if we had access to a list of *all of the songs by all artists*, we could just query that list by asking for all songs that belong to a given artist. This may become clearer if we make some updates to `Song` and `Artist`. Say, for instance, in `Song`, we set up a class variable, `@@all`, set to an empty Array, and a getter method, `.all`. This way, when a song is initialized, we can push the instance into the `@@all` and be able to use `Song.all` to retrieve all `Song` instances: ``` class Song attr_accessor :artist, :name, :genre @@all = [] def initialize(name, genre) @name = name @genre = genre save end def save @@all << self end def self.all @@all end end ``` Okay, so now that we can get all songs, we should be able to do things like this: ``` lil_nas_x = Artist.new("Lil Nas X") rick = Artist.new("Rick Astley") old_town_road = Song.new("Old Town Road","hip-hop") never_gonna_give_you_up = Song.new("Never Gonna Give You Up","pop") old_town_road.artist = lil_nas_x never_gonna_give_you_up.artist = rick Song.all.first.name #=> "Old Town Road" Song.all.first.genre #=> "hip-hop" Song.all.first.artist #=> # Song.all.first.artist.name #=> "Lil Nas X" Song.all.last.name #=> "Never Gonna Give You Up" Song.all.last.genre #=> "pop" Song.all.last.artist #=> # Song.all.last.artist.name #=> "Rick Astley" ``` Now that we've got a way to get all songs, if we want to find all the songs that belong to a particular artist, we can just *select* the appropriate songs: ``` Song.all.select {|song| song.artist == lil_nas_x} #=> [#>] Song.all.select {|song| song.artist == rick} #=> [#>] ``` So, with `Song.all`, if we have an `Artist` instance like `lil_nas_x` or `rick`, we can retrieve all the songs associated with that artist. We can incorporate this directly into our `Artist` class, replacing the implementation of the `#songs` method so that it *selects* instead of returning the `@songs` instance variable: ``` class Artist ... def songs Song.all.select {|song| song.artist == self} end end ``` This is an instance method, so we can use `self` to represent the `Artist` instance this method is called on. This changes the rest of the class - if we can just get the necessary information selecting from `Song.all`, we no longer need the `@songs` instance variable in our `Artist` class. We can also update `#add_song` accordingly: ``` class Artist attr_accessor :name def initialize(name) @name = name end def add_song(song) song.artist = self end def songs Song.all.select {|song| song.artist == self} end end ``` With this implementation, we're able to achieve a "has-many" / "belongs-to" relationship while maintaining a single source of truth! Not only that, we were able to simplify the `Artist` class without losing any functionality! Now, let's go back to the original example in this section. With our new setup, the issue of maintaining both sides of the relationship is solved: ``` lil_nas_x = Artist.new("Lil Nas X") old_town_road = Song.new("Old Town Road","hip-hop") old_town_road.artist = lil_nas_x old_town_road.artist.name #=> "Lil Nas X" lil_nas_x.songs #=> [#>] ``` And `add_song` functions just as it did before: ``` rick = Artist.new("Rick Astley") never_gonna_give_you_up = Song.new("Never Gonna Give You Up","pop") rick.add_song(never_gonna_give_you_up) rick.songs #=> [#>] never_gonna_give_you_up.artist #=> # ``` ## Extending the Association and Cleaning up our Code The code we have so far is pretty good. The best thing about it though is that it accommodates future change. We've built solid associations between our `Artist` and `Song` class via our "has-many"/"belongs-to" code. With this foundation we can make our code even better in the following ways: ### The `#add_song_by_name` Method As it currently stands, we have to *first* create a song and *then* add it to a given artist's collection of songs. We are lazy programmers, if we could combine these two steps, that would make us happy. Furthermore, if you think about our domain model, i.e. the program we are writing to model the real-world environment of an artist and their songs, the current need to create a song and then add it to an artist doesn't really make sense. A song doesn't exist *before* an artist creates it. Instead, let's build a method `#add_song_by_name`, that takes in an argument of a name and genre and both creates the new song *and* adds that song to the artist's collection. ``` class Artist ... def add_song_by_name(name, genre) song = Song.new(name, genre) add_song(song) end ``` Here we use the logic of our original `#add_song` method, which adds a song to an artist's collection and tells that song that it belongs to that particular artist. But, we also create a new song using the name and genre from the arguments. This is not only neater and more elegant––now we don't have to create a new song on a separate line *every time* we want to add one to an artist––but it also makes more sense. ### The `#artist_name` Method Since we've already set up these great associations between instances of the `Song` and `Artist` class, we can use them to build other helpful methods. Currently, to access the name of a given song's artist, we have to chain our methods like this: ``` kiki.artist.name # => "Drake" ``` We can imagine knowing the name of an artist that a particular song belongs to would be helpful and probably used in mulitple situations. Rather than having to chain multiple methods, wouldn't it be nice if we have one simple and descriptive method that could return the name of a given song's artist? Let's build one! ``` class Song ... def artist_name self.artist.name end ``` Now we can call: ``` kiki.artist_name # => "Drake" ``` Much better. Notice that we used the `self` keyword inside the `#artist_name` method to refer to the instance of `Song` on which the method is being called. Then, we call `#artist` on that song instance. This would return the `Artist` instance associated with the song. Chaining a call to `#name` after that is equivalent to saying: call `#name` on the return value of `self.artist`, i.e. call `#name` on the artist of this song. ## Conclusion Using the foundation of a "has-many" / "belongs-to" associations, we can create many different useful methods. We can write methods like `add_song_by_name` that handle initializing and associating instances. We can also create methods like `artist_name`, that can simplify retrieving information from an associated instance. Establishing both "has-many" and "belongs-to" associations between two objects allows us to ask a song who its artist is, and ask an artist what their songs are. We've established a bi-directional relationship! Can you think of any other real world relationships where these associations could be applied? View [Has Many Object](https://learn.co/lessons/ruby-objects-has-many-readme) on Learn.co and start learning to code for free. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://certil-remy.gitbook.io/learn/oop-ruby/untitled-22.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.