6. Room Architecture

Written by Subhrajyoti Sen

In the previous section, you learned all the basics behind data storage on Android. You learned how to work with permissions, shared preferences, content providers, JetPack DataStore and SQLite.

SQLite is a fast, lightweight local database natively supported by Android that allows you to store large amounts of data in a structured way. The only downside of SQLite is that its syntax isn’t very intuitive since the way to interact with it can be very different from platform to platform.

Therefore, in this chapter, you’ll learn about one of the most popular libraries that helps you simplify your interaction with SQLite: Room.

Along the way, you’ll also learn:

• How Object Relational Mappers work.
• About Room’s integration with Google’s architecture components
• The basics behind entities, DAOs and Room databases.
• The advantages and disadvantages of Room.
• The app you are going to build in the rest of this section.

Surely, you’re ready to dive in!

Object Relational Mappers

Before using Room properly in your projects, you first need to learn what Room is.

Room is a type of data persistence library commonly known as Object Relational Mapper or ORM. ORMs are tools that allow you to retrieve, delete, save and update the contents of a relational database using the programming language of your choice.

ORMs are implemented as libraries or frameworks that provide an additional layer of abstraction called the data layer, allowing you to better interact with your database using a syntax similar to the object-oriented world.

To better understand how ORMs work, imagine that you have a Movie class with three properties: An id, a name and a release_date.

The diagram above is a class diagram that represents the Movie class. Like most object-oriented languages, each of these properties has a specific data type such as Int, String or Date.

With the help of an ORM, you can easily use the Movie class to create a new table in your database. In an ORM, classes represent a table inside your database and each property represents a column. For example, the ORM will translate the Movie class into a table like this:

Each column would also have the data type that best represents the original data type of the original property. For example, a String would be translated as a varchar and an Integer as an Int.

The way to create new records inside the tables differs from each implementation. For instance, some ORMs automatically create new entries each time a new class instance is created. Other ORMs such as Room use Data Access Objects or DAOs to query your tables.

The following is a simple example of how you would use a DAO in Room to create new Movie records in the previously mentioned table:

movieDao.insert(Movie(1, "Harry Potter", "10-11-05"))
movieDao.insert(Movie(2, "The Simpsons", "03-10-02"))
movieDao.insert(Movie(3, "Avengers", "08-01-10"))

And your table would look like this:

Easy, right?

Note: Room can autogenerate the primary key, in this case, ID. You will learn how to do that in a later chapter.

Now, check out how Room and Google’s android architecture components work and interact with each other.

Room and Android Architecture Components

Over the years, Android developers have adopted different practices in developing the architecture of their apps. Some programmers preferred to use an MVVM architecture with SugarORM, while others used MVP with ObjectBox and Firebase. This led to confusion since there was no recommended or official way of doing things.

Therefore, at the 2018 I/O conference, Google introduced the Android Architecture Components, a set of libraries focused on creating a robust architecture for your apps.

Room is part of the architecture components and it is Google’s ORM meant to replace other libraries such as ObjectBox or SugarORM.

An app built with Room and other architecture components usually relies on a set of components you’re going to see in detail in the following sections.

Database

On a device, the data is stored on a local SQLite database. Room provides an additional layer on top of the usual SQLite APIs that avoids creating a lot of boilerplate code using the SQLiteOpenHelper class.

For instance, suppose you want to create a simple database that stores a question table for a quiz app, like the one you will be building in the next chapter. A traditional implementation using the standard SQLite APIs would look something like this:

private const val SQL_CREATE_ENTRIES =
        "CREATE TABLE question (" +
                "question_id INTEGER PRIMARY KEY," +
                "text TEXT"

private const val SQL_DELETE_ENTRIES = "DROP TABLE IF EXISTS question"

class QuizDbHelper(context: Context) : SQLiteOpenHelper(context, DB_NAME, null, DATABASE_VERSION) {
    companion object {
        const val DATABASE_VERSION = 1
        const val DB_NAME = "question_database.db"
    }
    override fun onCreate(db: SQLiteDatabase) {
        db.execSQL(SQL_CREATE_ENTRIES)
    }
    override fun onUpgrade(db: SQLiteDatabase, oldVersion: Int, newVersion: Int) {
        db.execSQL(SQL_DELETE_ENTRIES)
        onCreate(db)
    }
    override fun onDowngrade(db: SQLiteDatabase, oldVersion: Int, newVersion: Int) {
        onUpgrade(db, oldVersion, newVersion)
    }
}

On the other hand, with Room the code becomes much more concise and easier to understand:

@Database(entities = [(Question::class)], version = 1)
abstract class QuestionDatabase : RoomDatabase() {
  abstract fun questionsDao(): QuestionDao
}

As you can see, Room drastically reduced the amount of boilerplate code. Room handles most of the database interaction for you under the hood. Instead of you writing the boilerplate code, Room uses code generation to generate the boilerplate code at compile time.

Entities

Entities in Room correspond to tables in your database and are usually defined in Kotlin as data classes. Take a look at the following Entity example:

@Entity(tableName = "question") //1
data class Question(
    @PrimaryKey //2
    @ColumnInfo(name = "question_id") //3
    var questionId: Int,
    @ColumnInfo(name = "text")
    val text: String)

Room gives you many annotations that allow you to define how you want your data class to be translated into an SQLite table.

Taking each commented section in turn:

1. The @Entity annotation tells Room that this data class is an Entity. You can use many different parameters to tell Room how this Entity will be translated into an SQLite table. In this case, you are using the tableName parameter to define the name for the table.
2. The @PrimaryKey annotation is mandatory and each Entity must have at least one field annotated as the primary key. You could also use the primaryKeys parameter inside your @Entity annotation to define your primary key.
3. The @ColumnInfo annotation is optional but very useful since it allows specific customization for your column. For example, you can define a custom name or change the data type.

DAOs

DAO stands for Data Access Object. DAOs are interfaces or abstract classes that Room uses to interact with your database using annotated functions. Your DAOs will typically implement one or more CRUD operations (create, read, update and delete operations) on a particular Entity.

The code for a DAO that interacts with the question Entity defined above would look like this:

@Dao //1
interface QuestionDao {
    
  @Insert(onConflict = OnConflictStrategy.REPLACE) //2
  fun insert(question: Question)

  @Query("DELETE FROM question") //3
  fun clearQuestions()

  @Query("SELECT * FROM question ORDER BY question_id") //4
  fun getAllQuestions(): LiveData<List<Question>>

}

Step-by-Step:

1. The @Dao annotation marks this interface as a Data Access Object. DAOs should always be abstract classes or interfaces since, at compile time, Room will internally create the necessary implementation for you according to your provided query methods.

2. The @Insert annotation declares that this method will perform a create operation by performing an INSERT INTO query using the object received as a parameter to create a new record.

3. The @Query annotation executes the query passed as a parameter. In this case, you are performing a delete operation by executing a DELETE FROM query.

4. This is another @Query annotated method that retrieves all the questions from your database by using SELECT query.

Don’t worry if something doesn’t make sense right now. You’ll be learning much more about DAOs in the The DAO Pattern chapter.

Repository

This class acts as a bridge between your data sources and your app. The repository class handles the interaction with your Room database and other backend endpoints such as web services and Open APIs. Repositories make it easy to abstract the data and network layers away from the rest of the app. This abstraction makes it easy to add new data sources without affecting the rest of the app.

ViewModel

Just like the Repository acts as a bridge between your data sources and your app, the ViewModels act as a bridge between your Repository and your user interface. The ViewModel communicates the data coming from your Repository to your Views and has the advantage of surviving configuration changes since it’s lifecycle-aware.

ViewModels are implemented independently of the views (Activities or Fragments). The separation makes it easy to change the view layer without affecting the business logic.

LiveData

LiveData is a data holder class that implements the Observer pattern. This means it can hold information and be observed for changes. Your views such as Fragments or Activities observe LiveData objects returned from your ViewModels and update the relevant widgets as needed. The unique thing about LiveData is that it emits changes only if the observer is in an active state. Hence, if your Activity is in the backstack or the background, LiveData will not emit any changes.

The following diagram illustrates the interaction between the components mentioned above:

You will be learning much more about the above components in the upcoming chapters, but for now, this is all you need to know.

Advantages of using Room

Room uses a local SQLite database to store your data. Therefore, some of the advantages and disadvantages of SQLite apply to Room as well.

Room, however, has some important advantages over the SQLite APIs:

• Compile-time query verification: SQLite queries are usually executed at run-time without any verification. If the query has any errors, an exception is thrown, usually crashing your app. Room performs verifications on the SQL queries at compile-time and notifies you of errors early on.

• Convenience annotations: As you already noticed in the previous examples, there are some awesome annotations for reducing number of repetitive code parts and boilerplate code.

• Data migration: Writing migrations between different versions of your database can be a tedious task. Room provides a set of APIs that streamline this process and also make it very easy to test the migrations.

Frequently asked Room questions

Are ORMs really necessary? Can’t I just use plain old SQLite?

Of course, you can use plain old SQLite! In fact, Android standard libraries include many utilities and classes that help you work directly with SQLite. The only downside with this approach is that you often have to deal with a lot of boilerplate code that can slow down your development.

Are there other ORMs for Android besides Room?

Sure! There many ORMs out there like ObjectBox or SugarORM.

What are the advantages of using Room vs. other ORMs?

The main advantage of using Room vs. ORMs is that Room offers the best integration with other architecture components like ViewModel and LiveData. Since Google develops it, you can be sure this library will be maintained and improved for a very long time to come.

Your app

This chapter has been full of theory and concepts. You’re probably wondering when you are actually going to start writing some code.

Well, the rest of the following chapters are going to be focusing solely on how to apply the previously mentioned concepts to build a fun quiz app called DroidQuiz. This app will allow your users to test their Android knowledge with a set of questions stored in a Room database:

DroidQuiz will help you learn about many important concepts behind Room:

• How to add the appropriate dependencies to your build.gradle file for Room and most of the architecture components such as LiveData.

• How to create a local SQLite database using Room.

• How to use Database Access Objects or DAOs to interact with your database.

• How to use Google’s Android architecture components such as LiveData and ViewModel to interact with your Room database.

• How to create indices and relationships between your tables.

• How to test your database, migrations, and ViewModels, and much more!

As you can see, there is a lot to learn. This section will guide you through every, single step needed to build a final version of the app.

Key points

• Room is an ORM developed by Google as a part of Android Architecture Components to simplify the interaction with your SQLite database and reduce boilerplate code.
• Entities in Room correspond to tables in your database.
• DAO stands for Data Access Object.
• The Repository class handles the interaction with your Room database and other backend endpoints.
• The ViewModel communicates the data coming from your repository to your views and has the advantage of surviving configuration changes.
• LiveData is a data holder class that can hold information and be observed for changes.
• ORMs provide an additional layer of abstraction that allows you to interact with your relational database with an Object-Oriented Language syntax.