I have two classes, User and Product in a 'many-to-many through' association, using the class Prole (for product role).
class User < ActiveRecord::Base
has_many :proles
has_many :products, through: :proles
end
class Product < ActiveRecord::Base
has_many :proles
has_many :users, through: :proles
end
class Prole < ActiveRecord::Base
# has an attribute called 'role'
end
prole has an attribute called role which I'd like to use to qualify the user-product association.
The association works fine, but I can't figure out how to access the role attribute after creating the association. For example, if I do:
user.products << product
how can I access the attribute in the prole object just created?
I guess I could iterate through the prole objects and find the correct one, but I'm hoping there's a cleaner way.
Is this possible? Any hints?
TIA.
I was hoping for something a little more direct, but here's a
POSSIBLE ANSWER:
prole = Prole.find_by user_id: user.id, product_id: product.id
or even better
prole = user.proles.where("product_id = #{product.id}")
After some testing, it looks like the easiest way to grab specifically the Prole object that was just created is by querying by the two foreign keys directly against the Prole model, as suggested in your possible answer.
Prole.find_by(user_id: user.id, product_id: product.id)
If you want it as an association on the user object, you could use the includes approach to do eager loading, but it will still load every prole for the user in question
# specifying proles: {product_id: product.id} in the where clause here
# only limits users retrieved, not proles
user = User.includes(:proles).where(id: user.id)
# eager-loaded prole array
user.proles.find { |prole| prole.product_id == product.id }
See this answer for more info on that. But it looks like your possible answer is the cleanest way.
Related
How would you set up an activerecord/datamapper association for the following scenario:
A user creates a "bookshelf" which has many books(a book object just has an isbn that is used to query an api, and has_many review objects associated with it). Let's say Jack creates a "bookshelf" with a book object. Then, lets say that Jill creates a "bookshelf" with the same book object(it has the same id and the same reviews). The book object has the following code as of now:
class Book < ActiveRecord::Base
has_many :reviews
end
Then, when you view the page for a book (you click the link to it from the "bookshelf" created by Jack) you should see the same book object when you clicked the link to it from Jill's "bookshelf" (e.g. both "bookshelves" have a link to /books/23 because they have the same book object).
I have not been able to figure this out with the has_many association because that requires me to make a new book each time a user adds a book to their "bookshelf." I have trouble understanding the has_and_belongs_to_many relationship, is that what should be used here? I was not able to find any similar questions on SO, so any help is greatly appreciated.
I am using Rails 4 with Ruby 2.1.
Here is a drawing of what I would like to accomplish:
Drawing
Yes, you would have to define many-to-many relationship between a Bookshelf and a Book. There are two ways to achieve this in Rails:
Option 1) Use has_and_belongs_to_many
See guide
According to official documentation has_and_belongs_to_many association:
Specifies a many-to-many relationship with another class. This associates two classes via an intermediate join table. Unless the join table is explicitly specified as an option, it is guessed using the lexical order of the class names. So a join between Developer and Project will give the default join table name of “developers_projects” because “D” precedes “P” alphabetically.
So, your classes should look like this:
class Bookshelf < ActiveRecord::Base
has_and_belongs_to_many :books
end
class Book < ActiveRecord::Base
has_and_belongs_to_many :bookshelves
has_many :reviews
end
Add a join table generation to your migrations:
class CreateBooksBookshelvesJoinTable < ActiveRecord::Migration
def change
create_table :books_bookshelves, id: false do |t|
t.belongs_to :book, index: true
t.belongs_to :bookshelf, index: true
end
end
end
This will create a books_bookshelves table in your database. The table will have no primary key. There would be two foreign keys to your models Book and Bookshelf.
So, if you call self.books in the context of an user's bookshelf, you will get a list of books in the bookshelf. Vice versa, calling self.bookshelves in the context of a book will return a set of bookshelves the book belongs to.
The problem with this approach is that every time you add a new book to the bookshelf a new record is created in the database. If you are okay with that, there is no easier option than using has_and_belongs_to_many association. Otherwise, I recommend you to go with the Option #2.
Option 2) Use has_many :through
Another option is to use has_many, :through association (see guide). You would have to define one more model to do that, but it might come handy in some use cases (see below for an example).
Your classes should look like this:
class Bookshelf < ActiveRecord::Base
has_many :books, through: :books_bookshelves
has_many :books_bookshelves
end
class Book < ActiveRecord::Base
has_many :bookshelves, through: :books_bookshelves
has_many :books_bookshelves
has_many :reviews
end
class BooksBookshelf < ActiveRecord::Base
belongs_to :book
belongs_to :bookshelf
end
Probably the best thing about using has_many :through association is that it allows you to add custom columns to the join table (e.g. add column count to keep track how many books of the same type are there in the bookshelf).
The migration would look pretty much the same as the one we used in Option 1, except for the fact we are adding an unique constraint on the foreign keys (please note that adding the constraint is optional):
class CreateBooksBookshelvesJoinTable < ActiveRecord::Migration
def change
create_table :books_bookshelves, id: false do |t|
t.belongs_to :book, index: true
t.belongs_to :bookshelf, index: true
# add your custom columns here
end
add_index :books_bookshelves, [:book_id, :bookshelf_id], unique: true # to make sure you won't create duplicate records
end
end
By going with this approach, adding a new would be a bit more complicated as you would have to make sure you are not inserting duplicate records in the join table. (However, you may remove the unique constraint from the migration, to achieve exactly the same kind of behavior as you would get with has_and_belongs_to_many.)
I have two classes:
class User
include Mongoid::Document
has_one :preference
attr_accessible :name
field :name, type: String
end
class Preference
include Mongoid::Document
belongs_to :user
attr_accessible :somepref
field :somepref, type: Boolean
end
And I have two factories:
FactoryGirl.define do
factory :user do
preference
name 'John'
end
end
FactoryGirl.define do
factory :preference do
somepref true
end
end
After I create a User both documents are saved in the DB, but Preference document is missing user_id field and so has_one relation doesn't work when I read User from the DB.
I've currently fixed it by adding this piece of code in User factory:
after(:create) do |user|
#user.preference.save! #without this user_id field doesn't get saved
end
Can anyone explain to me why is this happening and is there a better fix?
Mongoid seems to be lacking support here.
When FactoryGirl creates a user, it first has to create the preference for that new user. As the new user does not have an id yet, the preference can't store it either.
In general, when you try create parent & child models in one operation, you need two steps:
create the parent, persist to database so it get's an id.
create the child for the parent and persist it.
Step two would end up in an after(:create) block. Like this:
FactoryGirl.define do
factory :user do
name 'John'
after(:create) do |user|
preference { create(:preference, user: user) }
end
end
end
As stated in this answer:
To ensure that you can always immediately read back the data you just
wrote using Mongoid, you need to set the database session options
consistency: :strong, safe: true
neither of which are the default.
Just getting started with ActiveRecord (in a sinatra app). Trying to port existing queries to AR but getting a little stuck.
if i have a has_one relation for users and profiles (using legacy tables unfortunately)
class User < ActiveRecord::Base
self.table_name = "systemUsers"
self.primary_key = "user_id"
has_one :profile, class_name: 'Profile', foreign_key: 'profile_user_id'
end
class Profile < ActiveRecord::Base
self.table_name = "systemUserProfiles"
self.primary_key = "profile_id"
belongs_to :user, class_name: "User", foreign_key: 'user_id'
end
if i want to query all users with profile using an inner join then get the user_age field from the profiles using one query can i do it?
for example
(just added .first to reduce code but would be looping through all users with profiles)
user = User.all(:joins => :profile).first
user.profile.user_age
gives me the correct data and uses a INNER join for the first query but then issues a second query to get the profile data
it also gives a depreciated warning and suggests i use load, which i tried but won't use an inner join.
similar case with
user = User.joins(:profile).first
user.profile.user_age
i get an INNER join but a query for each user row.
I have tried includes
user = User.includes(:profile).first
user.profile.user_age
This lazy loads the profile and would reduce the number of queries in the loop, however i think it would pull users without a profile too
I also tried with a reference
user = User.includes(:profile).references(:profile).first
user.profile.user_age
This gives me the correct data and reduces the queries to 1 but uses a LEFT JOIN
I probably have not quite grasped it and am trying to achieve something thats not do-able, i figured i might either need to use includes and check for nil profiles inside the loop or use joins and accept the additional query for each row.
Thought i'd check incase i was missing something obvious.
Cheers
Pat.
Profile should always have one user. So, I would do Profile.first.user_age for the first user profile. But going by the user approach like you did,
User.find { |u| u.profile }.profile.user_age
User.find { |u| u.profile } returns the first user with true value.
To query all the user profiles and get their user_ages. Assuming all profiles has user_id and that should be the case.
Profile.pluck(:user_age)
This checks the presence of user_id if you save profiles without user id. This where.not is a new feature in Activerecord, check your version.
Profile.where.not(user_id: nil).pluck(:user_age)
This looks like it should be something pretty easy but I can't seem to get it to work. I have a model with a has_many relationship and I'd like a scope on the parent that allows me to select only certain attributes for each.
An example:
class Bakery < ActiveRecord::Base
has_many :pastries
scope :summary, select([:id, :name, 'some option calling pastries.summary'])
class Pastry < ActiveRecord::Base
belongs_to :bakery
scope :summary, select([:id, :image_url])
I'd like to be able to call something like Bakery.first.summary and get a Bakery model with only the id and name populated and for each pastry in it's pastries array to only have the id and image_url attributes populated.
You could do this, but it won't affect the SQL queries that are made as a result (assuming you're trying to optimise the underlying query?):
class Pastry
...
def summary
{
:id => self.id,
:image_url => self.image_url
}
end
end
class Bakery
...
def summary
pastries.collect {|i| i.summary }
end
end
This would then give you an array of hashes, not model instances.
ActiveRecord doesn't behave how you're expecting with models - it will fetch whatever data it thinks you need. You could look at using the Sequel gem instead, or executing a raw SQL query such as:
Pastry.find_by_sql("SELECT id, name from ...")
But this could give you unexpected behaviour.
Hope your all enjoying your hollydays.
Ive run into a pretty funny problem when trying to insert rows into a really really simple database table.
The basic idea is pretty simple. The user selects one/multiple users in a multiselect which are supposed to be added to a group.
This piece of code will insert a row into the user_group_relationships table, but the users id always
#group = Group.find(params[:group_id])
params[:newMember][:users].each do |uid|
# For debugging purposes.
puts 'Uid:'+uid
#rel = #group.user_group_relationships.build( :user_id => uid.to_i )
#rel.save
end
The user id always gets inserted as null even though it is clearly there. You can see the uid in this example is 5, so it should work.
Uid:5
...
SQL (0.3ms) INSERT INTO
"user_group_relationships"
("created_at", "group_id",
"updated_at", "user_id") VALUES
('2010-12-27 14:03:24.331303', 2,
'2010-12-27 14:03:24.331303', NULL)
Any ideas why this fails?
Looks like user_id is not attr_accessible in the UserGroupRelationship model.
You might also want to check this it may be relevant.
I think #zabba's answer is probably the one you need to look for but i would suggest a couple of extra things here.
Your "Group" and "User" models are connected to each other thru the "UserGroup" model it seems. You would have relationship
class Group < ActiveRecord::Base
has_many :user_group_relationships
has_many :users, :through => :user_group_relationships
end
class UserGroupRelationship < ActiveRecord::Base
belongs_to :group
belongs_to :user
end
In your controller
# Find the group
#group = Group.find(params[:group_id])
# For each user id, find the user and add the user_group_relationship
params[:newMember][:users].each{|u| #group.users << User.find(u) }
Read up Rails Documentation on associations and the methods generated automatically for you when associations are defined. More often than not, working with association is easier than you might think! I discover new APIs in Rails constantly! :) Good Luck!