[JPA] 리턴타입이 옵셔널(Optional)인 이유는?

JPA를 사용 중 아래 코드를 만났고 의문이 생겼다.

1
final Optional<TodoEntity> original = repository.findById(entity.getId());

왜 꼭 Optional로 리턴타입을 감싸줘야하는걸까?
그냥 리턴타입만 쓰면 안될까?

1
final TodoEntity original = repository.findById(entity.getId());

이런식으로 말이다.
위처럼 옵셔널(Optional)을 빼버리니 바로 IDE가 에러를 알려주었다.

1
Type mismatch: cannot convert from Optional<TodoEntity> to TodoEntity

왜 이런 에러가 발생하는 걸까?
원인은 바로 CrudRepository 인터페이스였다.




갑자기 CrudRepository라고?

Repository 인터페이스를 만들때 JPA를 쓰기 위해서 JpaRepository를 상속받는다.
JpaRepository인터페이스에서 find와 관련된 메서드를 알기위해서는 CrudRepository를 확인하라고 적혀있다.

  • JpaRepository 인터페이스
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
/**
* JPA specific extension of {@link org.springframework.data.repository.Repository}.
*
* @author Oliver Gierke
* @author Christoph Strobl
* @author Mark Paluch
* @author Sander Krabbenborg
* @author Jesse Wouters
* @author Greg Turnquist
*/
@NoRepositoryBean
public interface JpaRepository<T, ID> extends PagingAndSortingRepository<T, ID>, QueryByExampleExecutor<T> {

/*
* (non-Javadoc)
* @see org.springframework.data.repository.CrudRepository#findAll()
*/
@Override
List<T> findAll();

/*
* (non-Javadoc)
* @see org.springframework.data.repository.PagingAndSortingRepository#findAll(org.springframework.data.domain.Sort)
*/
@Override
List<T> findAll(Sort sort);

/*
* (non-Javadoc)
* @see org.springframework.data.repository.CrudRepository#findAll(java.lang.Iterable)
*/
@Override
List<T> findAllById(Iterable<ID> ids);

// (중략)
}




CrudRepository 인터페이스

CrudRepository 인터페이스에서 findById 메서드를 확인해보면 리턴값이 Optional 타입으로 고정되어있다.
그래서 바로 IDE가 에러를 알려주었다. 그렇다면 Optional를 쓰면 좋은 점이 뭘까?

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
/**
* Interface for generic CRUD operations on a repository for a specific type.
*
* @author Oliver Gierke
* @author Eberhard Wolff
* @author Jens Schauder
*/
@NoRepositoryBean
public interface CrudRepository<T, ID> extends Repository<T, ID> {

/**
* Retrieves an entity by its id.
*
* @param id must not be {@literal null}.
* @return the entity with the given id or {@literal Optional#empty()} if none found.
* @throws IllegalArgumentException if {@literal id} is {@literal null}.
*/
Optional<T> findById(ID id);

// (생략)
}




Optional 쓰면 좋은 점?

Optional은 Java 8에 추가되었다. 이전 자바버전에서는 사용 불가능하다.
Optional은 고통스러운 null 처리를 도와주는 Wrapper 클래스이다.

코드 예시를 보자.

1
2
3
4
5
6
7
8
9
10
class Person {
String name;

String getName(){
if(name != null) {
return name;
}
return "name is null";
}
}

또는 try catch문을 이용해서 null을 처리할 수 있도 있고 다양한 NPE 처리 방법이 있다.

그런데 만약 필드가 많다면? 모든 필드를 NULL처리를 해줘야한다.

이때 Optional 클래스와 해당 클래스가 제공하는 orElse 메서드 또는 orElseGet 메서드를 이용하면 쉽게 NPE 처리가 가능하다.

메서드명 특징
orElseThrow(NullPointerException::new) null이라면 함수형 파라미터로 생성한 예외(NullPointerException)를 발생시킨다.
orElse(person) null이면 파라미터(person)를 반환한다. 이때 파라미터(person)은 해당 값이 null이든지 말든지 항상 미리 생성된다.
orElseGet null이면 파라미터(person)를 반환한다. 이때 파라미터(person)은 해당 값이 null 일때만 생성된다. 즉 미리 생성되지 않는다.



참고

Comments