Magic-in-Micronaut-JPA

less than 1 minute read

The best way to predict the future is to create it.

How findByIDAndNameOrderBy xx is resolved in Micronaut

DataIntroductionAdvice

If you check the source of micronaunt

io/micronaut/data/intercept/DataIntroductionAdvice.java

You’ll find the key logic is in findInterceptor, which will map the operation to an interceptor

private @NonNull
DataInterceptor<Object, Object> findInterceptor(
@Nullable String dataSourceName,
@NonNull Class<?> operationsType,
            @NonNull Class<?> interceptorType) {
DataInterceptor interceptor;

The interceptor could be

FindOneInterceptor
public class DefaultFindOneInterceptor
public interface FindAllInterceptor<T, R> extends IterableResultInterceptor<T, R> {
public class DefaultFindAllInterceptor<

DefaultFindOneInterceptor

Within following source code, io/micronaut/data/runtime/intercept/DefaultFindOneInterceptor.java You may find following useful logic:

@Override
public Object intercept(RepositoryMethodKey methodKey, MethodInvocationContext<T, Object> context) {
PreparedQuery<?, ?> preparedQuery = prepareQuery(methodKey, context, null);
return convertOne(
context,
operations.findOne(preparedQuery)
);
}

The operation is one HibernateJpaOperations

While in ApplicantChangesRepository

@Repositorypublic interface ApplicantChangesRepository extends JpaRepository<ApplicantChange, Integer> {

Following method

Optional<ApplicantChange> findFirstOrderByChangeIdDesc();

Due to cache of Hibernate, the first call to applicantChangesRepository.findFirstOrderByChangeIdDesc() will actually load ALL records into cache, then the 2nd call applicantChangesRepository.findFirstOrderByChangeIdDesc() would NO need database query.