English | 简体中文
Edit Page

Connect to Databases

To use Ktorm, you need to connect to your databases first. Ktorm provides a Database class to manage your database connections. The Database class is an abstraction of real databases. To create an instance of it, you can call the connect function on its companion object, providing your connection arguments or an existing DataSource object.

Connect with a URL

The code connecting to a MySQL database with a URL, user name and password:

1
2
3
4
5
6
val database = Database.connect(
url = "jdbc:mysql://localhost:3306/ktorm",
driver = "com.mysql.jdbc.Driver",
user = "root",
password = "***"
)

Easy to know what we do in the connect function. Just like any JDBC boilerplate code, Ktorm loads the MySQL database driver by Class.forName method first, then calls DriverManager.getConnection with your arguments to obtain a connection.

Of course, Ktorm doesn’t call DriverManager.getConnection in the beginning. Instead, we obtain connections only when it’s really needed (such as executing a SQL), then close them after they are not useful anymore. Therefore, Database objects created by this way won’t reuse any connections, creating connections frequently can lead to huge performance costs. It’s highly recommended to use connection pools in your production environment.

Connect with a Pool

Ktorm doesn’t limit you, you can use any connection pool you like, such as DBCP, C3P0 or Druid. The connect function provides an overloaded version which accepts a DataSource parameter, you just need to create a DataSource object and call that function with it:

1
2
val dataSource = SingleConnectionDataSource() // Any DataSource implementation is OK. 
val database = Database.connect(dataSource)

Now, Ktorm will obtain connections from the DataSource when necessary, then return them to the pool after they are not useful. This avoids the performance costs of frequent connection creation.

Connection pools are applicative and effective in most cases, we highly recommend you manage your connections in this way.

Connect Manually

If you want to manage connections’ lifecycle manually by yourself without using any connection pools, how to do that with Ktorm? For example, in some special business cases, there is only one connection needed in our whole App’s lifecycle. The connection is created when the App starts and closed when the process exits. The connect function provides another flexible overloaded version which accepts a parameter of type () -> Connection, a function that returns a Connection. The code below shows how to use it:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
// Create a connection when the App starts. 
val conn = DriverManager.getConnection("jdbc:mysql://localhost:3306/ktorm")

Runtime.getRuntime().addShutdownHook(
thread(start = false) {
// Close the connection when the process exits.
conn.close()
}
)

val database = Database.connect {
object : Connection by conn {
override fun close() {
// Override the close function and do nothing, keep the connection open.
}
}
}

Here, we call the connect function with a closure in which we should generally create a connection. However, the Connection is an interface, this allows us to return a proxy object to Ktorm instead of a real connection. The proxy overrides the close function as a no-op. In this way, Ktorm will always get the same connection object by calling the closure, and the connection is never closed in the whole App’s lifecycle.

Use Multiple Databases

The Database.connect function returns a new-created Database object. We should define a variable to save the returned value so as to use it to perform our database operations later. Sometimes, we also need to use many databases in one App, so it’s needed to create many Database instances and choose one while performing our database specific operations.

The code below connects to two different databases using Database.connect and shows how to switch between them:

1
2
3
4
5
6
7
8
val mysql = Database.connect("jdbc:mysql://localhost:3306/ktorm")
val h2 = Database.connect("jdbc:h2:mem:ktorm;DB_CLOSE_DELAY=-1")

// Obtain employees from MySQL database.
mysql.sequenceOf(Employees).toList()

// Obtain employees from H2 database.
h2.sequenceOf(Employees).toList()

Logging

During the running process, Ktorm outputs some useful internal information to the logs, such as generated SQLs, execution parameters, returned results, etc. If you want to see this information and monitor the running process, you may need to configure the logging output.

In order not to depend on any third-party logging frameworks, Ktorm adds a simple abstraction layer for logging, in which there are only two core classes:

  • LogLevel: similar to most of the logging frameworks, it’s an enum class that contains five logging levels: TRACE, DEBUG, INFO, WARN, ERROR.
  • Logger: an interface that provides some methods used to output logs.

Ktorm provides many implementations for the Logger interface:

Class NameDescription
ConsoleLoggerOutput logs to the console
JdkLoggerAdapterDelegate logs to java.util.logging
Slf4jLoggerAdapterDelegate logs to slf4j framework
CommonsLoggerAdapterDelegate logs to Apache commons logging lib
AndroidLoggerAdapterDelegate logs to android.util.Log

By default, Ktorm auto detects the logging framework we are using from the classpath while creating Database instances, and delegates the logs to it. If you want to output logs using a specific logging framework, you can choose an adapter implementation above and set it to the logger parameter. The code below uses a ConsoleLogger, and print logs whose level is greater or equal INFO to the console.

1
2
3
4
5
6
7
val database = Database.connect(
url = "jdbc:mysql://localhost:3306/ktorm",
driver = "com.mysql.jdbc.Driver",
user = "root",
password = "***",
logger = ConsoleLogger(threshold = LogLevel.INFO)
)

Ktorm prints different logs at different levels:

  • Generated SQLs and their execution arguments are printed at DEBUG level, so if you want to see the SQLs, you should configure your logging framework to enable the DEBUG level.
  • Detailed data of every returned entity object are printed at TRACE level, if you want to see them, you should configure your framework to enable TRACE.
  • Besides, start-up messages are printed at INFO level, warnings are printed at WARN level, and exceptions are printed at ERROR level. These levels should be enabled by default.