obj.update_attribute(:only_one_field, 'Some Value')
obj.update_attributes(field1: 'value', field2: 'value2', field3: 'value3')
Both of these will update an object without having to explicitly tell ActiveRecord to update.
Rails API says:
update_attribute
Updates a single attribute and saves the record without going through the normal validation procedure. This is especially useful for boolean flags on existing records. The regular update_attribute method in Base is replaced with this when the validations module is mixed in, which it is by default.
update_attributes
Updates all the attributes from the passed-in Hash and saves the record. If the object is invalid, the saving will fail and false will be returned.
So if I don't want to have the object validated I should use #update_attribute. What if I have this update on a #before_save, will it stackoverflow?
My question is does #update_attribute also bypass the before save or just the validation.
Also, what is the correct syntax to pass a hash to #update_attributes ... check out my example at the top.
11 Answers
Please refer to update_attribute. On clicking show source you will get following code
# File vendor/rails/activerecord/lib/active_record/base.rb, line 2614
2614: def update_attribute(name, value)
2615: send(name.to_s + '=', value)
2616: save(false)
2617: end
and now refer update_attributes and look at its code you get
# File vendor/rails/activerecord/lib/active_record/base.rb, line 2621
2621: def update_attributes(attributes)
2622: self.attributes = attributes
2623: save
2624: end
the difference between two is update_attribute uses save(false) whereas update_attributes uses save or you can say save(true).
Sorry for the long description but what I want to say is important. save(perform_validation = true), if perform_validation is false it bypasses (skips will be the proper word) all the validations associated with save.
For second question
Also, what is the correct syntax to pass a hash to update_attributes... check out my example at the top.
Your example is correct.
Object.update_attributes(:field1 => "value", :field2 => "value2", :field3 => "value3")
or
Object.update_attributes :field1 => "value", :field2 => "value2", :field3 => "value3"
or if you get all fields data & name in a hash say params[:user] here use just
Object.update_attributes(params[:user])
Tip: update_attribute is being deprecated in Rails 4 via Commit a7f4b0a1. It removes update_attribute in favor of update_column.
update_attribute
This method update single attribute of object without invoking model based validation.
obj = Model.find_by_id(params[:id])
obj.update_attribute :language, “java”
update_attributes
This method update multiple attribute of single object and also pass model based validation.
attributes = {:name => “BalaChandar”, :age => 23}
obj = Model.find_by_id(params[:id])
obj.update_attributes(attributes)
Hope this answer will clear out when to use what method of active record.
Also worth noting is that with update_attribute, the desired attribute to be updated doesn't need to be white listed with attr_accessible to update it as opposed to the mass assignment method update_attributes which will only update attr_accessible specified attributes.
update_attribute simply updates only one attribute of a model, but we can pass multiple attributes in update_attributes method.
Example:
user = User.last
#update_attribute
user.update_attribute(:status, "active")
It pass the validation
#update_attributes
user.update_attributes(first_name: 'update name', status: "active")
it doesn't update if validation fails.
You might be interested in visiting this blog post concerning all the possible ways to assign an attribute or update record (updated to Rails 4) update_attribute, update, update_column, update_columns etc. . For example it differs in aspects such as running validations, touching object's updated_at or triggering callbacks.
As an answer to the OP's question update_attribute does not by pass callbacks.
Great answers. notice that as for ruby 1.9 and above you could (and i think should) use the new hash syntax for update_attributes:
Model.update_attributes(column1: "data", column2: "data")
update_attribute and update_attributes are similar, but
with one big difference: update_attribute does not run validations.
Also:
update_attributeis used to update record with single attribute.Model.update_attribute(:column_name, column_value1)update_attributesis used to update record with multiple attributes.Model.update_attributes(:column_name1 => column_value1, :column_name2 => column_value2, ...)
These two methods are really easy to confuse given their similar names and works. Therefore, update_attribute is being removed in favor of update_column.
Now, in Rails4 you can use Model.update_column(:column_name, column_value) at the place of Model.update_attribute(:column_name, column_value)
Click here to get more info about update_column.
To answer your question, update_attribute skips pre save "validations" but it still runs any other callbacks like after_save etc. So if you really want to "just update the column and skip any AR cruft" then you need to use (apparently)
Model.update_all(...) see
Recently I ran into update_attribute vs. update_attributes and validation issue, so similar names, so different behavior, so confusing.
In order to pass hash to update_attribute and bypass validation you can do:
object = Object.new
object.attributes = {
field1: 'value',
field2: 'value2',
field3: 'value3'
}
object.save!(validate: false)
I think your question is if having an update_attribute in a before_save will lead to and endless loop (of update_attribute calls in before_save callbacks, originally triggered by an update_attribute call)
I'm pretty sure it does bypass the before_save callback since it doesn't actually save the record. You can also save a record without triggering validations by using
Model.save false
