Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

When use ResponseEntity<T> and @RestController for Spring RESTful applications

People also ask

Why do we use ResponseEntity in a RESTful service?

ResponseEntity represents the whole HTTP response: status code, headers, and body. As a result, we can use it to fully configure the HTTP response. If we want to use it, we have to return it from the endpoint; Spring takes care of the rest.

When should I use controller and RestController?

@Controller is used to mark classes as Spring MVC Controller. @RestController annotation is a special controller used in RESTful Web services, and it's the combination of @Controller and @ResponseBody annotation. It is a specialized version of @Component annotation.

Why does Spring boot need ResponseEntity?

ResponseEntity represents an HTTP response, including headers, body, and status. While @ResponseBody puts the return value into the body of the response, ResponseEntity also allows us to add headers and status code.

What is the use of @RestController in Spring?

Spring RestController annotation is used to create RESTful web services using Spring MVC. Spring RestController takes care of mapping request data to the defined request handler method. Once response body is generated from the handler method, it converts it to JSON or XML response.


ResponseEntity is meant to represent the entire HTTP response. You can control anything that goes into it: status code, headers, and body.

@ResponseBody is a marker for the HTTP response body and @ResponseStatus declares the status code of the HTTP response.

@ResponseStatus isn't very flexible. It marks the entire method so you have to be sure that your handler method will always behave the same way. And you still can't set the headers. You'd need the HttpServletResponse.

Basically, ResponseEntity lets you do more.


To complete the answer from Sotorios Delimanolis.

It's true that ResponseEntity gives you more flexibility but in most cases you won't need it and you'll end up with these ResponseEntity everywhere in your controller thus making it difficult to read and understand.

If you want to handle special cases like errors (Not Found, Conflict, etc.), you can add a HandlerExceptionResolver to your Spring configuration. So in your code, you just throw a specific exception (NotFoundException for instance) and decide what to do in your Handler (setting the HTTP status to 404), making the Controller code more clear.


According to official documentation: Creating REST Controllers with the @RestController annotation

@RestController is a stereotype annotation that combines @ResponseBody and @Controller. More than that, it gives more meaning to your Controller and also may carry additional semantics in future releases of the framework.

It seems that it's best to use @RestController for clarity, but you can also combine it with ResponseEntity for flexibility when needed (According to official tutorial and the code here and my question to confirm that).

For example:

@RestController
public class MyController {

    @GetMapping(path = "/test")
    @ResponseStatus(HttpStatus.OK)
    public User test() {
        User user = new User();
        user.setName("Name 1");

        return user;
    }

}

is the same as:

@RestController
public class MyController {

    @GetMapping(path = "/test")
    public ResponseEntity<User> test() {
        User user = new User();
        user.setName("Name 1");

        HttpHeaders responseHeaders = new HttpHeaders();
        // ...
        return new ResponseEntity<>(user, responseHeaders, HttpStatus.OK);
    }

}

This way, you can define ResponseEntity only when needed.

Update

You can use this:

    return ResponseEntity.ok().headers(responseHeaders).body(user);

A proper REST API should have below components in response

  1. Status Code
  2. Response Body
  3. Location to the resource which was altered(for example, if a resource was created, client would be interested to know the url of that location)

The main purpose of ResponseEntity was to provide the option 3, rest options could be achieved without ResponseEntity.

So if you want to provide the location of resource then using ResponseEntity would be better else it can be avoided.

Consider an example where a API is modified to provide all the options mentioned

// Step 1 - Without any options provided
@RequestMapping(value="/{id}", method=RequestMethod.GET)
public @ResponseBody Spittle spittleById(@PathVariable long id) {
  return spittleRepository.findOne(id);
}

// Step 2- We need to handle exception scenarios, as step 1 only caters happy path.
@ExceptionHandler(SpittleNotFoundException.class)
@ResponseStatus(HttpStatus.NOT_FOUND)
public Error spittleNotFound(SpittleNotFoundException e) {
  long spittleId = e.getSpittleId();
  return new Error(4, "Spittle [" + spittleId + "] not found");
}

// Step 3 - Now we will alter the service method, **if you want to provide location**
@RequestMapping(
    method=RequestMethod.POST
    consumes="application/json")
public ResponseEntity<Spittle> saveSpittle(
    @RequestBody Spittle spittle,
    UriComponentsBuilder ucb) {

  Spittle spittle = spittleRepository.save(spittle);
  HttpHeaders headers = new HttpHeaders();
  URI locationUri =
  ucb.path("/spittles/")
      .path(String.valueOf(spittle.getId()))
      .build()
      .toUri();
  headers.setLocation(locationUri);
  ResponseEntity<Spittle> responseEntity =
      new ResponseEntity<Spittle>(
          spittle, headers, HttpStatus.CREATED)
  return responseEntity;
}

// Step4 - If you are not interested to provide the url location, you can omit ResponseEntity and go with
@RequestMapping(
    method=RequestMethod.POST
    consumes="application/json")
@ResponseStatus(HttpStatus.CREATED)
public Spittle saveSpittle(@RequestBody Spittle spittle) {
  return spittleRepository.save(spittle);
}