Custom fields with mongoid - ruby-on-rails

I started to use the mongoid gem in my project, and I'm a little confused about how it store and get the information on the database. I have fields of specifics types in my models, but when I get it from the DB it returns a Hash.
Here is my models:
service.rb
class Service
include Mongoid::Document
field :username, type: String
field :strategy, type: Strategy
field :design, type: Design
end
strategy.rb
class Strategy
include Mongoid::Document
field :name, type: String
field :description, type: String
field :resources, type: Resources
field :scalability, type: Scalability
field :localization, type: Localization
field :contact, type: Contact
end
If I initialize a new service #service, and do #service.class it returns Service, the right one, but if I try do #service.strategy.class, it returns Hash, and not Strategy, like I was expecting. I read on the mongoid manual there are the "Custom field serialization", what I think allows me to do what I want. But I was wondering if there are not any other way to do that easily, because I have lots of models to change.

Related

Connecting Rails app to existing Mongodb database

I have a mongodb database with a single collection containing 400+ entries of basic data.
I'm using Rails and the mongoid gem to link the two together however when I query my model in the rails console there are no entries found.
QuizQuestion.first
Yields no results
My model:
class QuizQuestion
include Mongoid::Document
field :question, type: String
field :correctAnswer, type: String
field :wrongAnswers, type: Array, default: []
field :category, type: String
end
I have configured the mongoid.yml configuration file to point to the address of the database.
Does anyone know how to correctly do this or where I'm going wrong?
The reasons why you see no result:
1) database config is incorrect and you are pointing to a different database on the same mongodb instance
2) class name does not match the name for the collection within mongo. Open up a console/terminal and type:
mongo
then type this:
show dbs
This is the name of the dbs you need in the first part
use x
Where x is the db name
show collections
This will list the names of the collections.
Once you have the name of your collections, you can add this to your model:
store_in collection: "name_of_collection_as_in_mongo"
Therefore if the name of your collection was quiz_question as shown in the mongo client you can do this on your model:
class QuizQuestion
include Mongoid::Document
store_in collection: "quiz_question"
field :question, type: String
field :correctAnswer, type: String
field :wrongAnswers, type: Array, default: []
field :category, type: String
end
The reason you are not seeing any records (if you are pointing at the correct db name) is most likely due to mongoid expecting the class name to equal a pluralised collection name so QuizQuestions == quiz_questions within mongo

Rails and mongo embed relations

I'm trying to create some relations on Mongoid but when I try to save the inner object or add it to the user.personal_accounts collection I get the following error
NoMethodError: undefined method `bson_type' for #<Bank:0x71c01a8>
My Object in rails console is correct
#<PersonalAccount _id: 56e87f669c27691be0d3041b, number: "55", active: true, bank: #<Bank _id: 56d74cdb9c27692fb4bd4c6d, code: 123, name: "Bradesco", country: "USA">>
My mappings
class PersonalAccount
include Mongoid::Document
field :number, type: String
field :active, type: Boolean
field :bank, type: Bank
embedded_in :user
end
class User
include Mongoid::Document
field :first_name, type: String
field :last_name, type: String
embeds_many :personal_accounts
end
class Bank
include Mongoid::Document
field :code, type: Integer
field :name, type: String
field :country, type: String
end
The mapping that I was expecting is:
User
PersonalAccounts
Bank
Bank
As I have read that I need to copy the outer bank to each PersonalAccount.
I have already tried the following Link
Versions installed:
bson (4.0.2)
bson_ext (1.5.1)
mongoid (5.0.2)
mongo (2.2.4)
The root of your problem is right here:
field :bank, type: Bank
MongoDB doesn't know how to store a Bank so Mongoid will try to convert it to something that MongoDB will understand while Mongoid is preparing the data for the database, hence the NoMethodError.
Presumably you want Bank to exist as its own collection and then each PersonalAccount would refer to a Bank. That would be a standard belongs_to setup:
class PersonalAccount
#... but no `field :bank`
belongs_to :bank
end
That will add a field :bank_id, :type => BSON::ObjectId to PersonalAccount behind the scenes and hook up accessor (bank) and mutator (bank=) methods for you.
Normally you'd want the other half of the relation in Bank:
class Bank
#...
has_many :personal_accounts
end
but that won't work (as you found out) because PersonalAccount is embedded inside User so Bank can't get at it directly. Keep in mind that embeds_one is just a fancy of wrapping the Mongoid machinery around a Hash field in a document and embeds_many is just a fancy way of wrapping the Mongoid machinery around an array of hashes inside another document; embedded documents don't have an independent existence, they're just a part of their parent.

Generate a ruby class in memory

I have a requirement to convert an ActiveRecord model class into a MongoDB Document class automatically. I am able to do so using a rails generator which will read the attributes of a model and generate the new document.rb.
If a ActiveRecord model class looks like below:
class Project < ActiveRecord::Base
attr_accessible :completed, :end_date, :name, :start_date
end
Then, a generated class confirming to Mongoid's structure will be as below:
class ProjectDocument
field :name, type: String
field :start_date, type: Date
field :end_date, type: Date
field :completed, type: Boolean
field :created_at, type: Time
field :updated_at, type: Time
end
But I don't want to store a different document files, one for each model. I want to be able to generate this document class on the fly, whenever the rails application is started.
Is this possible? Is this approach of generating and using classes from memory advised? I don't have constraints on changes to AR model structure; the document is flexible w.r.t data structure and changed columns will get added automatically.
My first attempt would look something like this:
klass = Project
new_class = Object.const_set(klass.name + "Document", Class.new)
klass.columns.each do |c|
new_class.class_eval do
field c.name.to_sym, type: c.type
end
end
You'll almost certainly have to do something more complicated to set the field type correctly, but this should give you a good starting point.

Mongoid relation with a pre-defined object model?

Here's my use case:
I've got a collection full of sales tax rates that have been imported from CSV files. I created the Mongoid model to mirror the field names (these are not changeable):
class SalesTaxRate
include Mongoid::Document
field :state, type: String
field :zip_code, type: String
field :tax_region_name, type: String
field :tax_region_code, type: String
field :combined_rate, type: Float
end
Next, I'm building a model for use in my app. Let's say I want to create something called a Location:
class Location
include Mongoid::Document
field :name, type: String
field :street, type: String
field :city, type: String
field :state, type: String
field :zip_code, type: String
end
I'd like to be able to get a location's sales tax rate simply by calling something like this:
home = new Location(...)
home.sales_tax_rate
I'll never be setting the rate via home, just looking it up.
What's the "right" way to do this? I can think of two approaches -- the simple way seems to be just to define a method that does the lookup, as so:
class Location
...
def sales_tax_rate
SalesTaxRate.where(zip_code: self.zip_code).first.combined_rate
end
And this works. But I'm wondering whether I should be using a belongs_to association and, if so, why and how best to do that.
Still learning the ropes here, so apologies if this is a novice/silly question. Many thanks in advance!
If you have an index on zip_code in model SalesTaxRate what you are doing is essentially the same as what belongs_to will do. Just have a nil check in your code to ensure that it doesn't fail:
SalesTaxRate.where(zip_code: self.zip_code).first.try(:combined_rate)
# or
rate = SalesTaxRate.where(zip_code: self.zip_code).first
rate.nil? ? nil : rate.combined_rate
If you still want to go belongs_to route, you can define zip_code to be the identity in your SalesTaxRate. But you should take care of few things if you do that: First, all the zip codes in imported data need to be unique. Second, your location model can not have any zip code which is not available in SalesTaxRate otherwise you will face issues.

Should i store my facebook_ids as Integer or as String in mongo?

I have the following (using mongoid):
class FacebookUser
include Mongoid::Document
include Mongoid::Timestamps
field :uid, type: String
field :friends, type: Array
end
I am using mongoid and am wondering whether to store facebook_ids and facebook_ids_of_friends(Array) as BigDecimal, Integer or String. I will be using the uid for queries so am somewhat concerned about speed.
OR:
class FacebookUser
include Mongoid::Document
include Mongoid::Timestamps
field :uid, type: Integer
field :friends, type: Array #???How do I get this to store ints instead of strings
end
To avoid casting things repeatedly, I believe that the first option is better but wanted to get another opinion? In addition, how would I store an Array with Integer if I go with option 2?
Use a string. The only reason to use numeric types is if you will be doing calculations on the data. It is nonsensical to try to add two user id's so they should not be stored as integers.
Try this
class FacebookUser
identity :type => String
end
due to line identity :type => String your id of model will become String type and you can save facebook user id(uid) directly in there instead of creating a new field. The queries like finding users will be really easy and fast
For e.g.
find one user
FacebookUser.find(facebook user id)

Resources