I'll start with a simple example. You have a Spring boot application that runs a CommandLineRunner
class on initialization.
// MyCommandLineRunner.java
public class MyCommandLineRunner implements CommandLineRunner {
private final Log logger = LogFactory.getLog(getClass());
@Autowired //IntelliJ Warning
private DataSource ds;
@Override
public void run(String... args) throws Exception {
logger.info("DataSource: " + ds.toString());
}
}
// Application.java
@SpringBootApplication
public class Application {
public static void main(String... args) {
SpringApplication.run(Application.class, args);
}
@Bean
public MyCommandLineRunner schedulerRunner() {
return new MyCommandLineRunner();
}
}
Now, like this, this works, everything is OK. However, IntelliJ reports a warning where @Autowired
is located (I marked where in the comment)
Spring team recommends: Always use constructor based dependency injection in your beans. Always use assertions for mandatory dependencies.
Now if I follow this, I have a constructor based dependency injection
@Autowired
public MyCommandLineRunner(DataSource ds) { ... }
This also means that I have to edit Application.java
as well, since the constructor needs an argument. In Application.java
if I try to use the setter injection, I'll get the same warning. If I refactor that as well, I'll end up with some, in my opinion, nasty code.
// MyCommandLineRunner.java
public class MyCommandLineRunner implements CommandLineRunner {
private final Log logger = LogFactory.getLog(getClass());
private DataSource ds;
@Autowired // Note that this line is practically useless now, since we're getting this value as a parameter from Application.java anyway.
public MyCommandLineRunner(DataSource ds) { this.ds = ds; }
@Override
public void run(String... args) throws Exception {
logger.info("DataSource: " + ds.toString());
}
}
// Application.java
@SpringBootApplication
public class Application {
private DataSource ds;
@Autowired
public Application(DataSource ds) { this.ds = ds; }
public static void main(String... args) {
SpringApplication.run(Application.class, args);
}
@Bean
public MyCommandLineRunner schedulerRunner() {
return new MyCommandLineRunner(ds);
}
}
The above code yields the same result, but doesn't report any warnings in IntelliJ. I'm confused, how is the 2nd code better than the first one? Am I following an incorrect logic? Should this be wired differently?
In short, what's the correct way to do this?
note DataSource
is just a pure example, this question applies to anything being autowired.
note 2 Just to say that MyCommandLineRunner.java
can't have another, empty, constructor, since DataSource needs to be autowired/initialized. It will report an error and will not be compiled.
There are several ways to improve it.
You can remove @Autowired
from your MyCommandLineRunner
as you are letting a @Bean
method construct an instance of it. Inject the DataSource
directly into the method as an argument.
Or remove @Autowired
and remove the @Bean
and slap a @Component
annotation on your MyCommandLineRunner
to have it detected and remove factory method.
Inline your MyCommandLineRunner
inside your @Bean
method as a lambda.
MyCommandLineRunner
public class MyCommandLineRunner implements CommandLineRunner {
private final Log logger = LogFactory.getLog(getClass());
private final DataSource ds;
public MyCommandLineRunner(DataSource ds) { this.ds = ds; }
@Override
public void run(String... args) throws Exception {
logger.info("DataSource: " + ds.toString());
}
}
And the application class.
@SpringBootApplication
public class Application {
public static void main(String... args) {
SpringApplication.run(Application.class, args);
}
@Bean
public MyCommandLineRunner schedulerRunner(DataSource ds) {
return new MyCommandLineRunner(ds);
}
}
@Component
@Component
public class MyCommandLineRunner implements CommandLineRunner {
private final Log logger = LogFactory.getLog(getClass());
private final DataSource ds;
public MyCommandLineRunner(DataSource ds) { this.ds = ds; }
@Override
public void run(String... args) throws Exception {
logger.info("DataSource: " + ds.toString());
}
}
And the application class.
@SpringBootApplication
public class Application {
public static void main(String... args) {
SpringApplication.run(Application.class, args);
}
}
CommandLineRunner
@SpringBootApplication
public class Application {
private static final Logger logger = LoggerFactory.getLogger(Application.class)
public static void main(String... args) {
SpringApplication.run(Application.class, args);
}
@Bean
public MyCommandLineRunner schedulerRunner(DataSource ds) {
return (args) -> (logger.info("DataSource: {}", ds);
}
}
All of these are valid ways of constructing your instances. Which one to use, use the one that you feel comfortable with. There are more options (all variations on the ones mentioned here).