I'm currently working on a small school project that utilizes Ruby on Rails and I'm having some trouble getting my self-referential associations working correctly.
Context
The intended functionality of my web app is for users to post houses/apartments for other users to search through and rent. Since I'm having issues with a specific association, I'm working with a completely stripped down version that only has two models, User and Lease.
What I'm Trying to Accomplish
Ideally, when a person first registers on the site, a User object is created to hold their information such as email and password. A User can then either post a listing or search through listings.
Once a post has been created and another user decides to rent the posted house, a Lease object is created, which holds the ID of the posting User as well as the ID of the renting user, aliased as "landlord_id" and "tenant_id" respectively.
A User should now be identified as either a User, Landlord or a Tenant (or both Landlord and Tenant) based on whether there are any Lease objects with their ID as either a Landlord or a Tenant. This identification will be used to determine whether the User can access other areas of the site.
userFoo.leases
This should give me a list of all Lease objects with which the User's ID is associated, regardless of whether it's as a Landlord or Tenant.
userFoo.tenants
This should give me a list of any User object whose ID is associated with the ID of userFoo as a Tenant through Lease, and the inverse if I ask for landlords.
The Code
User Class
class User < ApplicationRecord
has_many :tenants, class_name: "Lease", foreign_key: "landlord_id"
has_many :landlords, class_name: "Lease", foreign_key: "tenant_id"
end
Lease Class
class Lease < ApplicationRecord
belongs_to :landlord, class_name: "User"
belongs_to :tenant, class_name: "User"
end
Users Table Migration
class CreateUsers < ActiveRecord::Migration[6.0]
def change
create_table :users do |t|
t.string :name
t.string :email
t.string :password_digest
t.timestamps
end
end
end
Leases Table Migration
class CreateLeases < ActiveRecord::Migration[6.0]
def change
create_table :leases do |t|
t.references :landlord, null: false, foreign_key: {to_table: :users}
t.references :tenant, null: false, foreign_key: {to_table: :users}
t.timestamps
end
end
end
Database Schema
ActiveRecord::Schema.define(version: 2020_10_18_005954) do
create_table "leases", force: :cascade do |t|
t.integer "landlord_id", null: false
t.integer "tenant_id", null: false
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
t.index ["landlord_id"], name: "index_leases_on_landlord_id"
t.index ["tenant_id"], name: "index_leases_on_tenant_id"
end
create_table "users", force: :cascade do |t|
t.string "name"
t.string "email"
t.string "password_digest"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
add_foreign_key "leases", "users", column: "landlord_id"
add_foreign_key "leases", "users", column: "tenant_id"
end
What's Wrong?
userFoo.leases
Normally a User would have_many leases by having their ID associated with a lease as "user_id." However, since I'm using "tenant_id" and "landlord_id", this command fails because it can't find "user_id" in the Leases table.
userFoo.tenants
This command gives me a list of all Lease objects where userFoo's ID is associated as "landlord_id" instead of all User objects associated with userFoo's ID as tenants. To retrieve a tenant as is, I have to use the command:
userFoo.tenants.first.tenant
Conclusion
I am having a bit of a hard time understanding these deeper, more complex associations, and I've spent some time trying to find a detailed reference on has_many that covers all the arguments, but all I can really find are small blog posts that reference the "Employees" and "Managers" example on guides.rubyonrails.com . I think one problem is that I'm not sure I'm correctly reflecting my model associations in my table schema.
I'm more than happy to teach myself if someone can point me in the right direction. I'm also open to alternative solutions but only if I can't get the functionality I want out of this setup, because my instructor specifically asked me to try it this way
Thanks in advance for any help! It's much appreciated.
as per your requirement you can try like this:
# app/models/user.rb
class User < ApplicationRecord
has_many :owned_properties, class_name: "Property", foreign_key: "landlord_id"
has_many :rented_properties, class_name: "Property", foreign_key: "tenant_id"
end
Here I have declared two associations with same table but different foreign keys.
# app/models/property.rb
class Property < ApplicationRecord
belongs_to :landlord, class_name: "User"
belongs_to :tenant, class_name: "User"
end
Here I have taken one table by using this user can post one property where landlord is the owner of a house and later you can add tenant who is taking rent to one property.
# db/migrations/20201018054951_create_users.rb
class CreateUsers < ActiveRecord::Migration[6.0]
def change
create_table :users do |t|
t.string :name, null: false
t.string :email, null: false, index: true
t.string :password_digest, null: false
t.timestamps
end
end
end
Above is your users table migration.
# db/migrations/20201018055351_create_properties.rb
class CreateProperties < ActiveRecord::Migration[6.0]
def change
create_table :properties do |t|
t.references :landlord, foreign_key: {to_table: :users}, null: false
t.references :tenant, foreign_key: {to_table: :users}
t.timestamps
end
end
end
Above is your properties table migration.
# db/schema.rb
ActiveRecord::Schema.define(version: 2020_10_18_055351) do
create_table "properties", force: :cascade do |t|
t.bigint "landlord_id", null: false
t.bigint "tenant_id"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
t.index ["landlord_id"], name: "index_properties_on_landlord_id"
t.index ["tenant_id"], name: "index_properties_on_tenant_id"
end
create_table "users", force: :cascade do |t|
t.string "name", null: false
t.string "email", null: false
t.string "password_digest", null: false
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
t.index ["email"], name: "index_users_on_email"
end
add_foreign_key "properties", "users", column: "landlord_id"
add_foreign_key "properties", "users", column: "tenant_id"
end
If you want to fetch all the owned properties of a user, use user.owned_properties.
If you want to fetch all rented properties of a user, use user.rented_properties.
^^ Here both the cases you'll get objects of Property class.
If you want to get landlord of a property, use property.landlord.
If you want to get tenant of a property, use property.tenant.
^^ Here both the cases you'll get objects of User class.
If you want you can add other attributes like: name, price, etc to properties table.
I think, this will help you. Thanks :) Happy Coding :)
Related
I've been researching friendship models using roles, custom associations, etc. But I haven't been able to connect my project to the concepts in a clear way.
I want a "User" to be able to create an event I'm calling a "Gather". A User can also attend a Gather created by other Users. By attending a Gather, the "User" can also be a "Gatherer".
The list of Gatherers will technically be considered friends of the "creator". This is how far I've gotten:
Models:
User
Gather
Gatherer (?)
User
class User < ApplicationRecord
has_many :gathers_as_creator,
foreign_key: :creator_id,
class_name: :Gather
has_many :gathers_as_gatherer,
foreign_key: :gatherer_id,
class_name: :Gather
end
Gather
class Gather < ApplicationRecord
belongs_to :creator, class_name: :User
belongs_to :gatherer, class_name: :User
end
My question is, do I need to a join table, such as Gatherer, to allow multiple attendees and then later pull a friend list for the user/creator ?
Gatherer
belongs_to :gather_attendee, class_name: "User"
belongs_to :attended_gather, class_name: "Gather"
Here's what I think that schema would look like:
create_table "gatherers", force: :cascade do |t|
t.bigint "attended_gather_id"
t.bigint "gather_attendee_id"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
t.index ["attended_gather_id"], name: "index_gatherers_on_attended_gather_id"
t.index ["gather_attendee_id"], name: "index_gatherers_on_gather_attendee_id"
end
Help, my head is spinning trying to understand the connections and how to proceed.
Previous planning:
Schema:
create_table "activities", force: :cascade do |t|
t.string "a_type"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
create_table "gatherers", force: :cascade do |t|
t.bigint "attended_gather_id"
t.bigint "gather_attendee_id"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
t.index ["attended_gather_id"], name: "index_gatherers_on_attended_gather_id"
t.index ["gather_attendee_id"], name: "index_gatherers_on_gather_attendee_id"
end
create_table "gathers", force: :cascade do |t|
t.integer "creator_id"
t.integer "activity_id"
t.text "gather_point"
t.boolean "active"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
create_table "interest_gathers", force: :cascade do |t|
t.string "gather_id"
t.string "interest_id"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
create_table "interests", force: :cascade do |t|
t.string "i_type"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
create_table "users", force: :cascade do |t|
t.string "username"
t.string "img"
t.string "first_name"
t.string "last_name"
t.string "state"
t.string "city"
t.string "bio"
t.datetime "created_at", precision: 6, null: false
t.datetime "updated_at", precision: 6, null: false
end
add_foreign_key "gatherers", "gathers", column: "attended_gather_id"
add_foreign_key "gatherers", "users", column: "gather_attendee_id"
end
class User < ActiveRecord::Base
has_many :gatherers, foreign_key: gather_attendee_id
has_many :attended_gathers, through: :gatherers
has_many :created_gathers, foreign_key: :creator_id, class_name: "Gather"
end
class Gather < ActiveRecord::Base
has_many :gatherers, foreign_key: :attended_gather_id
has_many :attendees, through: :gatherers, source: :gather_attendee
belongs_to :creator, class_name: "User"
end
class Gatherer < ActiveRecord::Base
belongs_to :gather_attendee, class_name: "User"
belongs_to :attended_gather, class_name: "Gather"
end
The naming here is not great. When naming your models choose nouns as models represent the actual things in your buisness logic - choosing verbs/adverbs makes the names of your assocations very confusing.
class User < ApplicationRecord
has_many :gatherings_as_creator,
class_name: 'Gathering',
foreign_key: :creator_id
has_many :attendences
has_many :gatherings,
through: :attendences
end
# think of this kind of like a ticket to an event
# rails g model Attendence user:references gathering:references
class Attendence < ApplicationRecord
belongs_to :user
belongs_to :gathering
end
# this is the proper noun form of gather
class Gathering < ApplicationRecord
belongs_to :creator,
class_name: 'User'
has_many :attendences
has_many :attendees,
though: :attendences,
class_name: 'User'
end
My question is, do I need to a join table, such as Gatherer, to allow multiple attendees and then later pull a friend list for the user/creator ?
Yes. You always need a join table to create many to many assocations. Gatherer is a pretty confusing name for it though as that's a person who gathers things.
If you want to get users attending Gatherings created by a given user you can do it through:
User.joins(attendences: :groups)
.where(groups: { creator_id: user.id })
You're on the right track.
If I understand what you're looking for correctly, you want a Gather to have many Users and a User to have many Gathers (for the attending piece). So you need a join table like this (this is similar to your gatherers table, but is in a more conventional Rails style):
create_join_table :gathers, :users do |t|
t.index [:gather_id, :user_id]
t.index [:user_id, :gather_id]
end
And then you'd want your User model to be like this:
class User < ApplicationRecord
has_many :gathers_as_creator, foreign_key: :creator_id, class_name: "Gather"
has_and_belongs_to_many :gathers
end
class Gather < ApplicationRecord
belongs_to :creator, class_name: "User"
has_and_belongs_to_many :users
end
(You can change the name of that :users association if you really want, by specifying extra options -- I just like to keep to the Rails defaults as much as I can.)
That should be the bulk of what you need. If you want to pull all the friends of a creator for a specific gather, you would just do gather.users. If you want to pull all of the friends of a creator across all their gathers, that will be:
creator = User.find(1)
friends = User.joins(:gathers).where(gathers: { creator: creator }).all
I added a couple of foreign keys to my models and to my tables and it has since broken my use of Active Admin. I'm wondering if anyone knows a work around or a fix to this issue.
schmea.rb
create_table "students", primary_key: "student_id", id: :string, force:
:cascade, options: "ENGINE=InnoDB DEFAULT CHARSET=utf8" do |t|
t.string "last_name"
t.string "first_name"
t.string "home_address"
t.string "home_city"
t.string "home_state"
t.string "home_zip"
t.string "school_year_address"
t.string "school_year_city"
t.string "school_year_zip"
t.string "room_number"
t.string "home_phone"
t.string "cell_phone"
t.boolean "new_student"
t.boolean "returning_student"
t.string "athletic_team"
t.bigint "user_id"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
t.index ["user_id"], name: "fk_rails_148c9e88f4"
end
add_foreign_key "emergency_contacts", "students", primary_key: "student_id"
add_foreign_key "students", "users"
add_foreign_key "vehicles", "students", primary_key: "student_id"
student.rb Students Model
class Student < ApplicationRecord
self.primary_key = :student_id
belongs_to :user
has_one :emergency_contact
has_one :vehicle
end
I'm getting the error, has anyone found a fix for this?
undefined method `emergency_contact_id_eq' for Ransack::Search<class: Student, base: Grouping <combinator: and>>:Ransack::Search
You have defined the assocation wrong.
class Student < ApplicationRecord
self.primary_key = :student_id
belongs_to :user
belongs_to :emergency_contact, class_name: 'User'
has_one :vehicle
end
belongs_to places the foreign key on this table and is exactly what you want. When joining you want to have the id on this table instead of having to look for records where student_id matches this record.
You also need to make sure to add a foreign key column and the correct foreign key constraint:
class AddEmergencyContactIdToStudents < ActiveRecord::Migration[5.0]
def change
add_reference :students, :emergency_contact, foreign_key: false
add_foreign_key :students, :users, column: :emergency_contact_id,
end
end
I would also strongly advise against using non standard primary keys. Prefixing the PK with student_ gives you nothing but headaches and will confuse other developers.
I am developing a portfolio for my website, I decided to add skills to each portfolio item.
class PortfolioSkill < ApplicationRecord
belongs_to :portfolio
belongs_to :skill
end
class Portfolio < ApplicationRecord
has_many :portfolio_skills
has_many :skills, through: :portfolio_skills
def all_tags=(names)
self.skills = names.split(",").map do |name|
Skill.where(name: name.strip).first_or_create!
end
end
def all_tags
self.skills.map(&:name).join(", ")
end
def remove_skill_tags
PortfolioSkill.where(portfolio_id: id).destroy_all
end
end
create_table "portfolio_skills", force: :cascade do |t|
t.integer "portfolio_id"
t.integer "skill_id"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
t.index ["portfolio_id"], name: "index_portfolio_skills_on_portfolio_id"
t.index ["skill_id"], name: "index_portfolio_skills_on_skill_id"
end
create_table "portfolios", force: :cascade do |t|
t.string "name"
t.string "client"
t.date "completed"
t.text "about"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
t.string "long_landscape"
t.string "cover"
t.integer "category_id"
t.index ["category_id"], name: "index_portfolios_on_category_id"
end
When I click destroy on the index page I get the
SQLite3::ConstraintException: FOREIGN KEY constraint failed: DELETE FROM "portfolios" WHERE "portfolios"."id" = ?
error. All the associations look right. I used this same pattern for my tags on other models and it worked with no issues. Any help would be great.
You are deleting from portfolios table, but table portfolio_skills has a column referencing it as foreign key. Hence the error.
Trying to delete a parent without checking and deleting its associated children can lead to data inconsistency. This exception is in place to prevent that.
Rails dependent destroy will take care of removing associated children rows while removing a parent.
Try using a dependent destroy:-
class Portfolio < ApplicationRecord
has_many :portfolio_skills, :dependent => :destroy
...
end
I am trying to create a many to many relationship between my groups table and my keywords table.
When I am in my controller I cant do Keyword.groups or Group.keywords as I get a no method error. I have checked Group.methods and I only have these related methods
"before_add_for_groups_keywords",
"before_add_for_groups_keywords?",
"before_add_for_groups_keywords=",
"after_add_for_groups_keywords",
"after_add_for_groups_keywords?",
"after_add_for_groups_keywords=",
"before_remove_for_groups_keywords",
"before_remove_for_groups_keywords?",
"before_remove_for_groups_keywords=",
"after_remove_for_groups_keywords",
"after_remove_for_groups_keywords?",
"after_remove_for_groups_keywords=",
"before_add_for_keywords",
"before_add_for_keywords?",
"before_add_for_keywords=",
"after_add_for_keywords",
"after_add_for_keywords?",
"after_add_for_keywords=",
"before_remove_for_keywords",
"before_remove_for_keywords?",
"before_remove_for_keywords=",
"after_remove_for_keywords",
"after_remove_for_keywords?",
"after_remove_for_keywords=",
Where as Keyword.methods gives me these
"before_add_for_keywords_groups",
"before_add_for_keywords_groups?",
"before_add_for_keywords_groups=",
"after_add_for_keywords_groups",
"after_add_for_keywords_groups?",
"after_add_for_keywords_groups=",
"before_remove_for_keywords_groups",
"before_remove_for_keywords_groups?",
"before_remove_for_keywords_groups=",
"after_remove_for_keywords_groups",
"after_remove_for_keywords_groups?",
"after_remove_for_keywords_groups=",
"before_add_for_groups",
"before_add_for_groups?",
"before_add_for_groups=",
"after_add_for_groups",
"after_add_for_groups?",
"after_add_for_groups=",
"before_remove_for_groups",
"before_remove_for_groups?",
"before_remove_for_groups=",
"after_remove_for_groups",
"after_remove_for_groups?",
"after_remove_for_groups=",
My models
has_and_belongs_to_many :keywords
has_and_belongs_to_many :groups
My db schema is the following
create_table "groups", force: :cascade do |t|
t.integer "member_id"
t.string "name"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
end
add_index "groups", ["member_id"], name: "index_groups_on_member_id", using: :btree
create_table "groups_keywords", id: false, force: :cascade do |t|
t.integer "group_id"
t.integer "keyword_id"
end
add_index "groups_keywords", ["group_id", "keyword_id"], name: "index_groups_keywords_on_group_id_and_keyword_id", unique: true, using: :btree
create_table "keywords", force: :cascade do |t|
t.string "keyword"
t.string "keyword_hash"
t.datetime "checked_at"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
end
Associations are defined for instances.
Keyword.groups indeed does not exist (as a class method).
However, Keyword#groups, as in Keyword.first.groups, does work. The reason for that is that Keyword.first returns an instance of Keyword class, which Keyword class itself is not.
A better way to implement a many to many relationship is with the has_many :through idiom:
class Keyword
has_many :keyword_groups
has_many :groups, :through => :keyword_groups
class Group
has_many :keyword_groups
has_many :keywords, :through => :keyword_groups
class KeywordGroup
belongs_to :keyword
belongs_to :group
This will give you flexibility in the future if you need to extend your join model.
When you do Keyword.group(1), it doesn't do what you think it is doing. It just finds all the keywords and groups them by 1.
SELECT * "keywords".* FROM "keywords" GROUP BY 1
Indeed, you can't actually call these kind of method on Class level, associations are accessed through instance level. And as you have defined them has_and_belongs_to_many, you will have to be plural about them, not single.
keyword = Keyword.first
keyword.groups # Not 'group', and on instance level
schema.rb:
ActiveRecord::Schema.define(version: 20150324012404) do
create_table "groups", force: :cascade do |t|
t.string "title"
t.integer "teacher_id"
t.datetime "created_at", null: false
t.datetime "updated_at", null: false
end
create_table "groups_students", id: false, force: :cascade do |t|
t.integer "group_id"
t.integer "student_id"
end
add_index "groups_students", ["group_id"], name: "index_groups_students_on_group_id"
add_index "groups_students", ["student_id"], name: "index_groups_students_on_student_id"
create_table "users", force: :cascade do |t|
t.string "email", default: "", null: false
t.string "encrypted_password", default: "", null: false
t.string "reset_password_token"
t.datetime "reset_password_sent_at"
t.datetime "remember_created_at"
t.integer "sign_in_count", default: 0, null: false
t.datetime "current_sign_in_at"
t.datetime "last_sign_in_at"
t.string "current_sign_in_ip"
t.string "last_sign_in_ip"
t.datetime "created_at"
t.datetime "updated_at"
t.boolean "admin", default: false
t.string "type"
t.integer "group_id"
end
add_index "users", ["email"], name: "index_users_on_email", unique: true
add_index "users", ["reset_password_token"], name: "index_users_on_reset_password_token", unique: true
end
group.rb:
class Group < ActiveRecord::Base
belongs_to :teacher
has_and_belongs_to_many :students
end
student.rb:
class Student < User
has_and_belongs_to_many :groups
end
I could have set a simple belongs_to and a has_many relationship between the student and group models, but I want students to be able to belong to more than one group, so I set up a HABTM association and corresponding join table.
I think I that right?
The question is, how do I, in the console, set a Student to belong to more than one group?
I have setup a User with 'type: Student' and I have two Groups. So...
In the console I do:
student = Student.first
Then, I want to set 'student' to belong to both Groups, but I don't know how to do this.
To set it to belong to one group I can do:
student.update_attributes(group_id: 1)
But how do make it belong to both groups? It would have two group_id's wouldn't it? I don't know how to set this.
If you need to see any of the other files, it's the 'handcode' branch here:
https://github.com/Yorkshireman/sebcoles/tree/handcode
The answers others have already provided are correct. But if you're working with id's you can also do something like this
student = Student.first
student.group_ids = 1,2,3,4
You don't need to set group_id for the User, the association is handled by the join table and the HABTM statement. You should remove group_id from the users table in the schema.
From memory you should be able to do something like this:
student = Student.first
groups = Group.all
student.groups << groups
student.save
See the active record guide on HABTM associations - specfically 4.4.1.3
Instead of habtm, just use the normal through and your life becomes easy. Make sure an id is generated for the association table (remove id:false)
create_table "group_students", force: :cascade do |t|
t.integer :group_id, nil:false
t.integer :student_id, nil:false
end
class Group < ActiveRecord::Base
has_many :group_students, dependent: :destroy, inverse_of :group
has_many :students, through :group_students
end
class Student < ActiveRecord::Base
has_many :group_students, dependent: :destroy, inverse_of :student
has_many :groups, through: :group_students
end
class GroupStudent < ActiveRecord::Base
belongs_to :group,
belongs_to :student
validates_presence_of :group, :student
end
Group.last.students << Student.last
or..
Student.last.groups << Group.last
Student.last.groups = [Group.find(1), Group.find(2)]
etc....
Ok, so it took me 3 days of all kinds of pain to work this out.
There was nothing wrong with my original code, except that I needed to remove the group_id from the user table.
roo's answer was correct, except that using 'group' as a variable name in the console confused Rails. This had led me to believe there was something wrong with my code, but there wasn't. You learn the hard way.
So, Students can be pushed into Groups like this:
To push a student into one group:
student = Student.first
OR
student = Student.find(1)
(or whatever number the id is)
group1 = Group.first
OR
group1 = Group.find(1)
student.groups << group1
To push into multiple groups (which was the original goal of this whole debacle:
student = Student.first
OR
student = Student.find(1)
allclasses = Group.all
student.groups << allclasses
To view your handywork:
student.groups
Works beautifully. The only problem I can see with my code is that it's possible to push the same student into a group twice, resulting in two duplicates of that student in one group. If anyone knows how to prevent this happening, I'm all ears.