Updating (i.e., saving) and destroying records are the next logical steps after creating and retrieving. (There is a mnemonic for this: CRUD, which stands for “create, read, update, destroy.”) The methods save and destroyRecord are available directly on model instances. These methods trigger requests through the adapter to update the data store. They return Promise objects, so you can chain callbacks with .then to do something with the returned data.

As you saw in this chapter, set is the way to change property values on your model records. If you change a value locally in the app, the data will be different from the persisted source. Therefore, after you change a value with set, you should save your data. You can do this with modelRecord.save. Saving a model will tell the store to make a request to the API with a POST or PUT, depending on the state of the record.

Although this was not mentioned above, when retrieving data the store will make get requests for data. When saving data, requests are sent with a type POST when the data does not exist in the persisted source and PUT when the data does exist and is being updated.

When you call createRecord, as noted earlier, you are not saving the data to the database. You are merely making an in-memory object. Calling save is for creation and updating.

Destroying records is the last step. Like the methods for retrieving and updating records, modelRecord.destroyRecord uses a request method – in this case DELETE – to remove a record from the persisted source. The store then removes the record from memory. Like save, destroyRecord is actually two function calls in one, deleteRecord and save, because deleteRecord only deletes the record locally. destroyRecord is more commonly used because it combines these steps into a single action.