This section describes how Models and tables are constructed via DBFlow. first let's describe how to get a database up and running.
In DBFlow, creating a database is as simple as only a few lines of code. DBFlow supports any number of databases, however individual tables and other related files can only be associated with one database.
@Database(version = 1)
abstract class AppDatabase : DBFlowDatabase()
@Database(version = 1)
public abstract class AppDatabase extends DBFlowDatabase {
}
The name of the database by default is the class name. To change it, read here.
Writing this file generates (by default) a AppDatabaseAppDatabase_Database.java
file, which contains tables, views, and more all tied to a specific database. This class is automatically placed into the main GeneratedDatabaseHolder
, which holds potentially many databases. The name, AppDatabaseAppDatabase_Database.java
, is generated via {DatabaseClassName}{DatabaseFileName}_Database
To learn more about what you can configure in a database, read here
DBFlow currently needs an instance of Context
in order to use it for a few features such as reading from assets, content observing, and generating ContentProvider
.
Initialize in your Application
subclass. You can also initialize it from other Context
but we always grab the Application
Context
(this is done only once).
class ExampleApplication : Application {
override fun onCreate() {
super.onCreate()
FlowManager.init(this)
}
}
public class ExampleApplication extends Application {
@Override
public void onCreate() {
super.onCreate();
FlowManager.init(this);
}
}
By default without passing in a DatabaseConfig
, we construct an AndroidSQLiteOpenHelper
database instance. To learn more about what you can configure in a database, read here, including providing own database instances.
Finally, add the custom Application
definition to the manifest (with the name that you chose for the class):
<application
android:name="{packageName}.ExampleApplication"
...>
</application>
A database within DBFlow is only initialized once you call database<SomeDatabase>()
. If you don't want this behavior or prefer it to happen immediately, modify your FlowConfig
:
override fun onCreate() {
super.onCreate()
FlowManager.init(FlowConfig.builder(this)
.openDatabasesOnInit(true)
.build())
}
@Override
public void onCreate() {
super.onCreate();
FlowManager.init(FlowConfig.builder(this)
.openDatabasesOnInit(true)
.build());
}
If you do not like the built-in DefaultTransactionManager
, or just want to roll your own existing system:
FlowManager.init(FlowConfig.builder(this)
.database(DatabaseConfig.builder(AppDatabase::class)
.transactionManagerCreator { db -> CustomTransactionManager(db))
.build()))
You can define different kinds for each database. To read more on transactions and subclassing BaseTransactionManager
go here
Creating models are as simple as defining the model class, and adding the @Table
annotation. To read more on this, read here.
For now: Models must provide a default, parameterless constructor. Also, all fields must be mutable (currently, we hope to evolve this requirement in the near future). An example:
@Table(database = TestDatabase::class)
class Currency(@PrimaryKey(autoincrement = true) var id: Long = 0,
@Column @Unique var symbol: String? = null,
@Column var shortName: String? = null,
@Column @Unique var name: String = "") // nullability of fields are respected. We will not assign a null value to this field.
or with Java:
@Table(database = TestDatabase.class)
public class Currency {
@PrimaryKey(autoincrement = true)
long id; // package-private recommended, not required
@Column
@Unique
String symbol;
@Column
String shortName;
@Column
@Unique
private String name; // private with getters and setters
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
}
Set up a DAO (Data Access Object) to help you interact with your database. Using dependency injection and service locating, we can build better, highly testable code. While not required to use DBFlow, it is highly recommended utilize this approach.
With kotlin, we can utilize it in a powerful way:
/**
* Create this class in your own database module.
*/
interface DBProvider<out T: DBFlowDatabase> {
val database: T
}
interface CurrencyDAO : DBProvider<AppDatabase> {
/**
* Utilize coroutines package
*/
fun coroutineRetrieveUSD(): Deferred<MutableList<Currency>> =
database.beginTransactionAsync {
(select from Currency::class
where (Currency_Table.symbol eq "$")).queryList(it)
}.defer()
/**
* Utilize RXJava2 package.
* Also can use asMaybe(), or asFlowable() (to register for changes and continue listening)
*/
fun rxRetrieveUSD(): Single<MutableList<Currency>> =
database.beginTransactionAsync {
(select from Currency::class
where (Currency_Table.symbol eq "$"))
.queryList(it)
}.asSingle()
/**
* Utilize Vanilla Transactions.
*/
fun retrieveUSD(): Transaction.Builder<MutableList<Currency>> =
database.beginTransactionAsync {
(select from Currency::class
where (Currency_Table.symbol eq "$"))
.queryList(it)
}
/**
* Utilize Paging Library from paging artifact.
*/
fun pagingRetrieveUSD(): QueryDataSource.Factory<Currency, Where<Currency>> = (select from Currency::class
where (Currency_Table.symbol eq "$"))
.toDataSourceFactory(database)
}
DBFlow uses expressive builders to represent and translate to the SQLite language.
We can represent the query above in SQLite:
SELECT * FROM Currency WHERE symbol='$';
Wherever we perform dependency injection we supply the instance:
fun provideCurrencyDAO(db: AppDatabase) = object : CurrencyDAO {
override val database: AppDatabase = db
}
Then in our ViewModel
, we can inject it via the constructor and utilize it in our queries:
class SampleViewModel(private currencyDAO: CurrencyDAO)
We support many kinds of complex and complicated queries using the builder language. To read more about this, see the wrapper language docs
There is much more you can do in DBFlow. Read through the other docs to get a sense of the library.