Dealing with a legacy system that has non-Rails conventional naming. Note that all tables and attributes are UPPERCASE, which in the sqlserver-adapter means that ID is NOT the same as id.
I had thought that alias_attribute :new, :OLD allowed you to specify a name that you could use in ActiveRecord/ActiveRelation queries. From what I'm seeing below (tested in rails console), that is not the case.
The end goal is making the legacy system "act" in a Rails-conventional methodology by making each model have an ID attribute, etc...
Model definition:
# app/models/organization.rb
class Organization < ActiveRecord::Base
self.table_name = "ORGANIZATION"
self.primary_key = "ORGANIZATION_ID"
alias_attribute :id, :ORGANIZATION_ID
end
Does not work:
Organization.select(:id) => invalid column name 'id'
Organization.select(:ID) => invalid column name 'ID'
Organization.select("ID") => invalid column name 'ID'
Does work:
Organization.select(:organization_id) => <finds record>
Organization.select(:ORGANIZATION_ID) => <finds record>
Organization.select("organization_id") => <finds record>
Organization.select("ORGANIZATION_ID") => <finds record>
I just used alias_attribute the other day. Honestly, I kept poking at it until I got mine to work.
I was able to achieve the functionality that you are looking for by reversing the params passed to alias_attribute.
Example:
alias_attribute :active?, :active
alias_attribute :alert, :message
alias_attribute :delete_at, :expire_at
Where the first param is what the column name is, and the second is what you are aliasing.
I understand that seems backwards from documentation, however this is how I got this to work.
I solved this situation by configuring the DB adapter (MS SQL Server) to treat the schema as all lowercase. Thus, even though the actual DB table names can be "USERS", "PEOPLE", etc..., I can reference them as "users", "people" in my application.
See: https://github.com/rails-sqlserver/activerecord-sqlserver-adapter#force-schema-to-lowercase
Here's the code I added to resolve this:
# config/initializers/sql_server_adapter.rb
ActiveRecord::ConnectionAdapters::SQLServerAdapter.lowercase_schema_reflection = true
Related
I have an ActiveAdmin model registered as such:
ActiveAdmin.register MyModel do
permit_params :name, :synonyms
filter :name
index do
selectable_column
column :name
actions
end
end
I noticed that when I update an object from the ActiveAdmin interface, the synonyms are getting cleared out. Synonyms are an array field defined as such in the PostgreSQL database:
synonyms text[] DEFAULT '{}'::text[]
I tried putting the following in app/admin/my_model.rb:
controller do
def update
permitted_params[:my_model][:synonyms] = JSON.parse permitted_params[:my_model][:synonyms]
super
end
end
and I also tried it with params instead of permitted_params but that doesn't work either. Not sure why ActiveAdmin would be discarding this field. Am I doing something incorrectly? The project I'm working with uses ActiveAdmin 1.0.0.pre4 (I realize this is a dated version).
Any advice would be much appreciated. Thanks in advance!
Notes: Seems this might be a common issue?
I'm not sure if this is applicable to string arrays and I don't know about that specific version of Activeadmin, but I encountered this issue in a slightly different context.
I had a model where the array data types were decimals & integers. I had to explicitly state the type of form input to be used when editing / updating the record or nothing was passed back from the field with the update parameters. Activeadmin chose a number input by default, but it needed to be processed as a string.
form do |f|
f.inputs do
f.input :ages, as :string, :input_html => {:maxlength => '100'}
end
end
I had to set maxlength manually because for some reason it was automatically being set very short. In the model, the string then has to be converted into an actual array before being saved.
def ages= items
if items.is_a? String
items = items.split(' ')
items.each do |i|
i.to_i
end
end
super items
end
Short version: I'm building a new Rails 4 application that uses (read-only) some tables from a database used by a legacy Rails 2 application, which is still in use. The old application models/tables were very confusingly named, however (especially in the context of the new application), so I want to use different names for the models/tables using self.table_name. This all works perfectly until I tried to add in a polymorphic relationship. Rails ignores my defined table_name and does a query on the type using the new model name, which of course is different so it doesn't work. Is there any way to change this?
Long version: There are three models in this equation, and here they are:
class Exporter < MysqlBase
has_many :lic_exporter_addresses, :as => :place
self.table_name = 'excons'
self.primary_key = 'id'
end
class LicBusiness < MysqlBase
has_one :physical_address, -> { where(category: 'Physical') }, :class_name => 'LicExporterAddress', :as => :place
has_one :mailing_address, -> { where(category: 'Mailing') }, :class_name => 'LicExporterAddress', :as => :place
has_many :lic_exporter_addresses, :as => :place
self.table_name = 'businesses'
self.primary_key = 'id'
end
class LicExporterAddress < MysqlBase
belongs_to :place, polymorphic: true
self.table_name = 'addresses'
self.primary_key = 'id'
end
We have a ton of different kinds of businesses, so the Business model is the most problematic. I really don't want to have that in the new app because it would be very confusing as to what a "business" actually is. With the current code if I go into the rails console and try to get lic_exporter_addresses for a LicBusiness or Exporter, it does:
SELECT `addresses`.* FROM `addresses` WHERE `addresses`.`place_id` = '00044c693f6848f9b0978f873cf9999a' AND `addresses`.`place_type` = 'LicBusiness'
when what I need is place_type = 'Business'.
Is there any way to tell Rails what place_type to look for? I did see this question and the second answer looked promising, except that I'm already sort of doing that with Physical and Mailing addresses so I can't figure out how that'd work with both options at the same time... Thanks for any info or ideas.
In Rails 4.2, it looks like the exact string used for the query is defined as owner.class.base_class.name, where owner is the model declaring the association. So I don't think it's directly supported. But there are a few ways I can think of to hack around this. I think the most promising might be, in LicBusiness:
has_many :lic_exporter_addresses, ->{where place_type: "Business"}, foreign_key: "place_id"
That is, don't define the association as polymorphic, but define the type scope yourself. This will NOT correctly define place_type in the lic_exporter_addresses table if you ever use lic_exporter_address.place = some_lic_business_instance. However you said this table was read-only, so this may in fact not be an issue for you. If it is, there may be ways to override the behavior to get what you need.
Two other ideas both make me very nervous and I think they are probably quite dangerous for unintended side-effects. They are to override LicBusiness.base_class (this might actually be ok if you do not now and never will have STI set up on LicBusiness, but I'm still nervous), or to override LicBusiness.name (I'm pretty sure this would have unintended side-effects).
I just started trying my hands on Ruby on rails . i have created a mode states and want that every state name must remain unique so i used
uniqueness: true
as per http://guides.rubyonrails.org/active_record_validations.html .
As the above document says that the validation is invoked automatically when object.save is called. but When i try to save the objects with same state_name value , no exception is thrown and the record is saved. Can one please help where i am doing it wrong.
Model code
class State < ActiveRecord::Base
acts_as_paranoid
validates_presence_of :state_name ,uniqueness: true
end
Controller code
def create
#stateName = params[:stateName];
#state = State.new();
#state.state_name=#stateName;
if(#state.save())
resp = { :message => "success" }
else
resp = { :message => "fail" }
end
respond_to do |format|
format.json { render :json => resp }
end
end
Thanks in advance for helping!
If you want uniqueness check, change
validates_presence_of :state_name ,uniqueness: true
to
validates :state_name, uniqueness: true
If you want uniqueness and presence check both, use
validates :state_name, uniqueness: true, presence: true
An alternate syntax is shown below, however, syntax shown above should be preferred
validates_uniqueness_of :fname
validates_presence_of :fname
Also, as per documentation, please note following with respect to usage of uniqueness: true
It does not create a uniqueness constraint in the database, so it may
happen that two different database connections create two records with
the same value for a column that you intend to be unique. To avoid
that, you must create a unique index on both columns in your database.
The validation happens by performing an SQL query into the model's
table, searching for an existing record with the same value in that
attribute.
What this means is that it is possible that if multiple users are trying to save records concurrently, there is a possibility records with duplicate state_name can get created as each save is happening on different thread
Rails4 is getting depreciation warning when I am upgrading from rails 3.2 to rails 4.0. I have this query.
Child.find(:all, :include => :children_users, :conditions => "state = 'active' AND owner_id = #{self.id} AND children_users.user_id = #{other_user.id}")
I am getting deprecation warning as follow :-
DEPRECATION WARNING: It looks like you are eager loading table(s) (one of: splits, accounts) that are referenced in a string SQL snippet. For example:
Post.includes(:comments).where("comments.title = 'foo'")
Currently, Active Record recognizes the table in the string, and knows to JOIN the comments table to the query, rather than loading comments in a separate query. However, doing this without writing a full-blown SQL parser is inherently flawed. Since we don't want to write an SQL parser, we are removing this functionality. From now on, you must explicitly tell Active Record when you are referencing a table from a string:
Post.includes(:comments).where("comments.title = 'foo'").references(:comments)
If you don't rely on implicit join references you can disable the feature entirely by setting `config.active_record.disable_implicit_join_references = true`. (called from splits_under_100_percent at /Users/newimac/RailsApp/bank/app/models/user.rb:274)
To solve this problem, I have try like this
Child.includes(:children_users).where(state: active, owner_id: self.id, children_users.user_id = other_user.id).load
or
Child.where{(state: active, owner_id: self.id, children_users.user_id = other_user.id).includes(:children_users)}
But none of them work.
children_users.user_id = other_user.id wrong.
The correct one is: "children_users.user_id" => other_user.id
Thank You for #Zakwan. Finally, this query works.
Child.includes(:children_users).where(state: 'active', owner_id: self.id, "children_users.user_id" => other_user.id).load
Apologies for posting what seems to be an oft-asked question but I have toiled for two days without getting closer.
I am running Rails 4.0.1 (with Devise, CanCan, among others) and have user and role models with a HABTM users_roles table (as created by the since-removed rolify).
class Role < ActiveRecord::Base
has_and_belongs_to_many :users, :join_table => :users_roles
belongs_to :resource, :polymorphic => true
attr_reader :user_tokens
def user_tokens=(ids)
self.user_ids = ids.split(",")
end
end
My submitted form data is:
"role"=>{"name"=>"disabled", "resource_id"=>"", "resource_type"=>"Phone", "user_tokens"=>["", "3"]}
(Not sure where the empty string comes from; it's not in the option list, but hey, welcome to Rails).
My problem is with getting the user_tokens data passed into the appropriate variable in the controller.
There are lots of postings suggesting the correct format (e.g., how to permit an array with strong parameters)
If I specify the permitted parameters thus:
params.require(:role).permit(:name, :resource_id, :resource_type, :user_tokens => [] )
I get a '500 Internal server error' with no other details.
If I specify the permit as
params.require(:role).permit(:name, :resource_id, :resource_type, user_tokens: [] )
I don't get any errors (No unpermitted parameters in the dev log) but the user_tokens array is not passed through in the #role.
If I run this scenario through a console I get:
params = ActionController::Parameters.new(:role => {:resource_id => "", :resource_type => "Phone", :user_tokens => ["", "3"]})
params.require(:role).permit(:name, :resource_id, :resource_type, user_tokens: [] )
=> {"resource_id"=>"", "resource_type"=>"Phone", "user_tokens"=>["", "3"]}
params.require(:role).permit(:user_tokens).permitted?
Unpermitted parameters: resource_id, resource_type, user_tokens
=> true
I'm stuck as to where I should try looking next.
Even though I was sure I had already tried it, changing the model to be
def user_tokens=(ids)
self.user_ids = ids
end
fixed it. I've always struggled to understand this bit of notation except when I've spent 4 days trying why my many:many models don't work.
I guess it'd be easier if I was doing this full time rather than just for particular sysadmin projects.