Say I have the following example of associations in a Rails app:
I'm considering combining the *Posting models under STI. One problem with STI is the potential for many attributes that are only related to one subclass (i.e., a lot of denormalized nil values). This is especially worrisome when your subclasses and going to evolve and grow in the future. I've read a few related posts (such as this), however, as you can see in my example, the potential subclass-specific fields will not necessarily be just attributes, but rather, a lot of belongs_to associations.
My question is, how could I restructure this to use STI from a Posting model for all the common attributes/methods (of which there will be quite a few in my actual app), but keep the unique subclass-specific attributes and belongs_to associations from piling up in the Posting model? Also, the ability to access #board.postings and work with those standard methods is important.
For example, I've thought about moving the type-specific attributes to another model:
class CarPosting < Posting
has_one: car_posting_detail
end
class CarPostingDetail < ActiveRecord::Base
belongs_to :car_posting
belongs_to :car_make
belongs_to :car_model
end
Although, this starts to create a lot of joins, I'm not sure I have the has_one/belongs_to declarations in the right direction, and you have to start chaining calls (e.g., #posting.car_posting_detail.car_make).
Are there other design patterns you have seen for accomplishing this?
You basically have to 2 options for accomplishing inheritance.
First, you can use rails STI as you suggested. The downside is that end up with nil attribute for the child classes that do not use all of the fields. Your idea to reduce this by adding type-specific attributes to another model is a great way to reduce this. However, you should keep the implementation as DRY as possible by defining a has_one :detail for the Posting. Then you can simply assign specific detail types in the Posting childs. For example, CarPosting's detail would be CarPostingDetail. This is convenient because then all Posting children will have their details accessed identically, but will still have different details. So the query now looks like #posting.detail.car_make. To take this one step further, you can define a custom helper method in your Posting model to grab each attribute in the current Posting's detail and create an accessor for it. Now the entire detail layer is transparent and you can simply access those attributes by saying #posting.car_make.
Second, you can use an abstract class. This is essentially the reverse of STI. You create an abstract model class which can never be instantiated. Thus, you cannot define any relationships in the Posting class because it has no table. Each child of the abstract Posting class has its own separate table. The main advantage of doing this would be the ability to define methods for all of your Posting types without copy and pasting them into every model. So this options is better if there are some overlapping functionality across the models, but very little data overlap.
You could use polymorphic associations for this.
Post model belongs_to :postable, :polymorphic => true
car, event and all the other "postable" classes would have this relationship
has_many :posts, as: :postable
Post would hold the postable_id and postable_type
More info here
http://guides.rubyonrails.org/association_basics.html#polymorphic-associations
Related
I have an application where a User can create many Links, and each Link can store different type of data, depending on what type of Link it is. For example, a TelephoneLinkData stores a telephone number, an EmailLinkData stores an email address, a subject and a body. Each Link also has some fields in common, such as a reference to the user and a name.
I've tried to map this into ActiveRecord as cleanly as I can. Currently, I have a polymorphic relationship from Link to its data-type:
class Link < ApplicationRecord
belongs_to :user
belongs_to :link_data, polymorphic: true
...
class EmailLinkData < ApplicationRecord
has_one :link, as: :link_data
accepts_nested_attributes_for :links
...
Technically, I think this would be described as a reverse polymorphic relationship as instead of a class having possibly different parent classes, what I'm trying to model is a class having multiple possible different child classes. This works fine, and I'm able to create Links through the various *LinkData controllers, but what I'd really want to do is have the Link act as the primary source of interaction for the user, so that the user manages their links through the /links path. For example, I would like the API to allow a User to create a link by posting to /links with the data for the LinkData nested in the link_data field
I've looked around for other ways to model this relationship, and the most common other suggestion seems to be Single-Table Inheritance, but the majority of my columns will differ between LinkData classes, so that feels like the wrong abstraction.
Is there a more idiomatic way to model this data structure?
As is always the case, the best choice depends on the business or application needs, so it's difficult to provide a recommendation without knowing more about what you're trying to do.
It sounds like you prefer the MTI approach, essentially using actual foreign keys and an XOR constraint to the Link table instead of a type column. That's a totally reasonable (although not as common) alternative to a polymorphic association.
However, I think there was a bit of a misunderstanding in your question.
Technically, I think this would be described as a reverse polymorphic relationship as instead of a class having possibly different parent classes...
A polymorphic association in Ruby/Rails doesn't have anything to do with class inheritance (e.g. parents and children). You might be thinking of Single table inheritance. A polymorphic association allows one class (e.g. a Link) to be associated a record in any other table (e.g. the various classes of LinkData) via two fields, a association_id and association_type. These associated classes need not be related to each other. For example, a common use case might be the acts_as_commentable gem, that allows you to add a comment to any other object, and the comment would have a polymorphic association with the other classes.
In the second part of your question you mention that you'd like the User to interact with Link's via a single controller.
I would like the API to allow a User to create a link by posting to /links with the data for the LinkData nested in the link_data field
There's nothing stopping you from implementing this using the initially proposed data model. ActiveRecord may not handle this completely for you out of the box, but you can imagine implementing a link_data= method on the Link class that would create the appropriate associated object.
I'd say the pros/cons of using a polymorphic association would be...
Pros:
easy to setup and use
easy to make required (validate presence of / not null)
easy to associate with a new class
Cons:
no referential / database integrity
have to migrate data if you change a class name
And using the MTI approach is basically the opposite. A bit harder to setup and use, harder to add a new association/table, harder to ensure exactly one association exists... but the long term data quality benefits are significant.
I was able to get things to work the way I wanted to using multiple table inheritance, based largely on this chapter: https://danchak99.wordpress.com/enterprise-rails/chapter-10-multiple-table-inheritance/
I have an abstract class called Work and another class I called Fanfic. In the Fanfic model, I had that it belongs_to :user and has_many :characters along with other associations. I decided that it would be easier when I add more classes that they will be subclasses of Work instead of having all of the types of works as totally separate classes.
Now I'm wondering if I could write the associations ALL of the works would have in the Work model, with all of the sub-classes keeping all of those associations as well.
Sorry if it's a bit confusing. Let me know if you need clarification or if you need to see my code, or if you need any other information.
Not only should you define associations in Work (super) class (and they will be usable in subclasses), you should also use works table for all the subclasses of Work.
It's typical case of Single Table Inheritance.
To enable it, you just need to add type column to works table (add migration for it).
When you query for Tales, for instance, Rails creates the following SQL query:
SELECT * FROM WORKS WHERE type='Tale';
There is a great (and consice) writeup about this concepts in Rails documentation: http://api.rubyonrails.org/classes/ActiveRecord/Inheritance.html
I have a Citation class and I want each Citation to have many Resources. A resource can be a LocalFile, Link (and maybe something like a dropbox file). Resources will share most functionality and columns. I want to use single table inheritance in the following manner.
class Resource < ActiveRecord::Base
belongs_to :citation
acts_as_list :scope => :citation
end
class LocalFile < Resource
end
class Link < Resource
end
class Citation < ActiveRecord::Base
has_many :resources, class_name: "Resource"
end
I want to be able to get a collection of resources with code something like
cites = mycite.citations
and have each object in cites have the appropriate type (LocalFile, Link etc..). Will this work? Strangely despite searching extensively I haven't been able to find an answer to this simple question.
Also, will the methods added by acts_as_list work appropriately when called on objects on the various subtypes, i.e., give move them up or down in the list on all resources for a single citation
Alternatively, is there a better way to do this? I want to have the Citation class have a collection of resources of multiple subtypes. Conceptually, each resource type is capable of representing the same content just accessed in different ways.
STI should be considered when dealing with model classes that share much of the same functionality and data fields, but you as the developer may want more granular control over extending or adding to each class individually. Rather than duplicate the code over and over for multiple tables (and not being DRY) or forego the flexibility of adding idiosyncratic functionality or methods, STI permits you to use keep your data in a single table while writing specialized functionality.
to implement STI:-
Create the base class
Along with your needed attributes, add a :type attribute as a string
to the base class
Create any necessary descendant classes
Drawbacks of STI
STI isn’t always the best design choice for your schema. If sub-classes that you intend to use for STI have many different data fields, then including them all in the same table would result in a lot of null values and make it difficult to scale over time. In this case, you may end up with so much code in your model sub-classes that the shared functionality between sub-classes is minimal and warrants separate tables. Also, as I mentioned above, STI shouldn’t be used just because two object types have similar attributes - both airplanes and bicycles have wheels, but it probalby doesn’t make sense to group them into the same table, given that intuitively, they’re different objects that will have vastly different functionality and data fields in an application.
HOPE IT HELPS :)
I have User model and Organization model. The only difference is that organization has_many users, all other properties are same. I don't want to put it in one table/model. How can I remove tons of code duplicating in this models? Should I use Concerns? I think, it will be not normal if User model will looks like :
class User < ActiveRecord::Base
include user_concern
end
So, how can I extend user model in other model? And how to generate this model with rails g with all User's fields inside?
beware STI
I would keep with concerns rather than using STI. STI often causes more problem that it solves (type mismatches, form urls, etc), plus inheritance won't make sense, here : an user is not a kind of company, and a company is not a kind of user.
That's a naming problem
I think your problem is a naming one. Obviously, your concern should not be "UserConcern". The question is : what kind of methods do you group in that concern ?
Are your methods about social relation between users ? Then, you need a Socializable concern. Are they about subscribing to mailing list ? Then you need a Subscribable concern.
It's ok to have several ones, even if they have a single method in it, because developers won't wonder "what the hell does this do ?" if all concerns are correctly named.
What to duplicate anyway
You should also probably let class level method calls out concerns.
If it's ok for scopes to be embedded in concerns (after all, they resolve in method definitions), it feels less natural to me to put relations in there.
It's ok to duplicate #has_many :foos, we do it all the time in separate models, and it's already difficult enough to get an idea of table schema from a model without hiding more information.
You could use single table inheritance (STI) for this. To get it to work, your model needs a type-field of type string, in which ActiveRecord stores the actual type of your record. Then, you just extend your base model for this.
migration
add_column :users, :type, :string
models
class User < ActiveRecord::Base and class Organisation < User.
ActiveRecord will now fill your type-field with the model-name, and store both in your users table (since this is the one the organisation model is inheriting from).
Have a look at the according section on http://api.rubyonrails.org/classes/ActiveRecord/Base.html .
However, in your case, I'd create a base model, e.g. Address, and then extend User and Organisation from it, to maintain semantically correct models.
I have just started investigating using more advanced models in Rails. One that I use regularly, with great success, is model where a many-to-many cross-reference relationship is accessed by a class that itself is a sub-class of the base class in the many-to-many relationship.
This way the cross-reference class can act as a stand-in for the base class.
A good example is where a navigation hierarchy node (NavigationNode) is cross-referenced to a user role. At the cross-reference point a class (RoleNavigationNode) can inherit from NavigationNode and still have intimate knowledge of the user role.
My question is (in the case above) can RoleNavigationNode inherit from NavigationNode and yet access the cross-reference table rather than the one that NavigationNode accesses -- this of course using ActiveRecord.
I have not investigated polymorphic association, which may be more appropriate.
Thanks in advance...,
tried set_table_name on the subclass?
Also, look into setting #abstract_class in model classes.
Lastly, what you need may simply be a Mixin that you include in both models.
Anyway, what you're trying to do sounds rather un-ActiveRecord-ish. You might want to post a clearer example of what you're trying to achieve, maybe we'll be able to come up with something simpler.
This works in Rails 3:
class Common < ActiveRecord::Base
#abstract_class = true
def common
"Foobar!"
end
end
class Model < Common
end
class AnotherModel < Common
end
Without setting abstract_class, Rails will look for a table named commons in your database.