According to the has_many documentation, the "shovel" method collection<<(object, …)
Adds one or more objects to the collection by setting their foreign keys to the collection's primary key. Note that this operation instantly fires update SQL without waiting for the save or update call on the parent object, unless the parent object is a new record.
If you want to construct a new-record without saving it to the database just yet, use collection.build.
Returns one or more new objects of the collection type that have been instantiated with attributes and linked to this object through a foreign key, but have not yet been saved.
Using Club and Member as example models:
club = Club.find(params[:id])
club.members.build(member_attributes) # member is not saved
club.save # saves club and members
BUT WHAT IF
I would like to use << to create associations and not firing SQL at all. How can I override << to behave like in scenario when the parent object is a new record?
I have a User model where there is a department field which has a default value of "Engineering".
This field was introduced 2 months after our website went live,and since its mongo, there was no migration.
When I try to get the object using where, correct value is returned
User.find_by(:name => "John).department
However, if I try to pluck values, it returns nil and not the default value.
User.limit(2).pluck(:department)
returns
[nil,"Finance"]
I researched a bit and came across this blog post http://ahmadsherif.com/blog/2013/01/29/mongoid-default-fields-can-give-you-hard-time/
I think I am facing the same issue. Is there any work around for this? I chose to go with pluck because it's not memory intensive and saves time.
Basically the behaviour here can be explained by the difference in how MongoDB deals with default values vs a traditional relational database like Postgres.
In the SQL world you set defaults via the database schema and the DB will fill a NULL field everytime you insert a row.
Since MongoDB is schemaless document fields have a default value of nil, which cannot be changed* since there is no schema where we could define defaults on the database level. Instead defaults are implemented on the application level. For Mongoid this means when you initialize a new model instance it will fill in default values if they are nil.
In ActiveRecord terms it would look like this:
class Thing < ActiveRecord::Base
after_initialize :set_default_foo, if: -> { self.foo.nil? }
private
def set_default_foo
self.foo = "bar"
end
end
However if you have existing documents and add a new field or defaults to an existing field Mongoid does not update the existing documents for you!
So how does this explain these two cases?
User.find_by(:name => "John").department
User.limit(2).pluck(:department)
In the first case you a pulling a document out of storage and using it to inialize a model instance. When the model instance is initialized a callback is run which sets the default values.
When .pluck is called on the hand Mongoid pulls the values directly from the store without initializing any model instances. Thus for any "legacy" documents it will return a nil value.
To remedy this you you need to set the default value for any document with a nil.
User.where(department: nil).update_all(department: 'engineering')
I have a procedure which receives two models, one which already exists, and another one which holds new attributes which I want to merge in the first one.
Since other parts of the program are holding the same reference to the new model, I can't just operate on the existing one. Therefor I do the following:
def merge(new_model, existing_model)
new_model.attributes = existing_model.attributes.merge(new_model.attributes)
new_model.id = existing_model.id
end
Now the new_model is being saved which gives me the uniqueness erorr (even though it's technically the same model). I also tried using the reload method, but that yields the same result.
Background:
The method above is run in a before_add callback on an association. I want to be able to call update on a model (with nested associations) without having to specify IDs of the nested models. This update is supposed to merge some associations, which is why I try to do the whole merge thing above.
You can't set the id of a model and then save the record expecting the id to be set since the id is the primary key of the database. So you are actually creating a whole new record and, thus, the uniqueness validation error. So you'll need to think of some other design to accomplish what you are wanting. It may help to know that what you are trying to do sounds similar to a deep_dup, except that ActiveRecord doesn't define this method (but Hash does).
#user = User.new
#user.id returns nil but i need to know it before i save. Is it possible ?
YES you can!
I had the same question and investigated the docs.
The ability to solve this question is very related to your database type in fact.
Oracle and Postgresql do have useful functions to easily solve this.
For MySQL(oracle) or SkySQL(open-source) it seems more complicated (but still possible). I would recommend you avoid using these (MySQL/SkySQL) databases if you need advanced database tools.
First you must try to avoid this situation as much as possible in your application design, as it is dangerous to play with IDs before they get saved.
There may be situation where you don't have any other choice:
For instance when two tables are referencing themselves and for security reason you don't allow DELETE or UPDATE on these tables.
When this is the case, you can use the (PostgreSQL, Oracle) database nextval function to generate the next ID number without actually inserting a new record.
Use it in conjunction with the find_by_sql rails method.
To do this with postgreSQL and Rails for instance, choose one of your rails models and add a class method (not an instance method!).
This is possible with the "self" word at the beginning of the method name.
self tells Ruby that this method is usable only by the class, not by its instance variables (the objects created with 'new').
My Rails model:
class MyToy < ActiveRecord::Base
...
def self.my_next_id_sequence
self.find_by_sql "SELECT nextval('my_toys_id_seq') AS my_next_id"
end
end
When you generate a table with a Rails migration, by default Rails automatically creates a column called id and sets it as the primary key's table. To ensure that you don't get any "duplicate primary key error", Rails automatically creates a sequence inside the database and applies it to the id column. For each new record (row) you insert in your table, the database will calculate by itself what will be the next id for your new record.
Rails names this sequence automatically with the table name append with "_id_seq".
The PostgreSQL nextval function must be applied to this sequence as explained here.
Now about find_by_sql, as explained here, it will create an array containing new objects instances of your class. Each of those objects will contain all the columns the SQL statement generates. Those columns will appear in each new object instance under the form of attributes. Even if those attributes don't exist in your class model !
As you wisely realized, our nextval function will only return a single value.
So find_by_sql will create an array containing a single object instance with a single attribute.
To make it easy to read the value of this very attribute, we will name the resulting SQL column with "my_next_id", so our attribute will have the same name.
So that's it. We can use our new method:
my_resulting_array = MyToy.my_next_id_sequence
my_toy_object = my_resulting_array[0]
my_next_id_value = my_toy_object.my_next_id
And use it to solve our dead lock situation :
my_dog = DogModel.create(:name => 'Dogy', :toy_id => my_next_id_value)
a_dog_toy = MyToy.new(:my_dog_id => my_dog.id)
a_dog_toy.id = my_next_id_value
a_dog_toy.save
Be aware that if you don't use your my_next_id_value this id number will be lost forever. (I mean, it won't be used by any record in the future).
The database doesn't wait on you to use it. If somewhere at any time, your application needs to insert a new record in your my_table_example (maybe at the same time as we are playing with my_next_id_sequence), the database will always assign an id number to this new record immediately following the one you generated with my_next_id_sequence, considering that your my_next_id_value is reserved.
This may lead to situations where the records in your my_table_example don't appear to be sorted by the time they were created.
No, you can't get the ID before saving. The ID number comes from the database but the database won't assign the ID until you call save. All this is assuming that you're using ActiveRecord of course.
I had a similar situation. I called the sequence using find_by_sql on my model which returns the model array. I got the id from the first object of the arry. something like below.
Class User < ActiveRecord::Base
set_primary_key 'user_id'
alias user_id= id=
def self.get_sequence_id
self.find_by_sql "select TEST_USER_ID_SEQ.nextval as contact_id from dual"
end
end
and on the class on which you reference the user model,
#users = User.get_sequence_id
user = users[0]
Normally the ID is filled from a database sequence automatically.
In rails you can use the after_create event, which gives you access to the object just after it has been saved (and thus it has the ID). This would cover most cases.
When using Oracle i had the case where I wanted to create the ID ourselves (and not use a sequence), and in this post i provide the details how i did that. In short the code:
# a small patch as proposed by the author of OracleEnhancedAdapter: http://blog.rayapps.com/2008/05/13/activerecord-oracle-enhanced-adapter/#comment-240
# if a ActiveRecord model has a sequence with name "autogenerated", the id will not be filled in from any sequence
ActiveRecord::ConnectionAdapters::OracleEnhancedAdapter.class_eval do
alias_method :orig_next_sequence_value, :next_sequence_value
def next_sequence_value(sequence_name)
if sequence_name == 'autogenerated'
# we assume id must have gotten a good value before insert!
id
else
orig_next_sequence_value(sequence_name)
end
end
end
while this solution is specific to Oracle-enhanced, i am assuming the other databases will have a similar method that you could redefine.
So, while it is definitely not advised and you want to be absolutely sure why you would not want to use an id generated by a sequence, if it is needed it is most definitely possible.
It is why I love ruby and Ruby on Rails! :)
In Oracle you can get your current sequence value with this query:
SELECT last_number FROM user_sequences where sequence_name='your_sequence_name';
So in your model class, you can put something like this:
class MyModel < ActiveRecord::Base
self.sequence_name = 'your_sequence_name'
def self.my_next_id_sequence
get_data = self.find_by_sql "SELECT last_number FROM user_sequences where sequence_name='your_sequence_name'"
get_data[0].last_number
end
end
And finally, in controller you can get this value with this:
my_sequence_number = MyModel.my_next_id_sequence
So, there is no need to get your next value by using NEXTVAL and you won't lose you ID.
What you could do is User.max(id). which will return the highest ID in the database, you could then add 1. This is not reliable, although might meet your needs.
Since Rails 5 you can simply call next_sequence_value
Note: For Oracle when self.sequence_name is set, requesting next sequence value creates side effect by incrementing sequence value
I am having a Property model that should contain multiple values (just Strings). With Rails/ActiveRecord it seems that I have to create a new model (and a new table) for those values (like PropertyValue). As every one of those values just stores one String the PropertyValue only need one attribute (like value).
I don't like that idea cause to access one of those values I now have to call property.values[0].value and that looks a bit ugly.
Is there a nicer solution?
Try serialize method
class Property < ...
serialize :value, ::Array
end
value array will be stored as string in properties table and you can access it as normal array: property.value[3].
More details in docs.