Java 8 Stream on JpaSpecificationExecutor with @QueryHint fetchSize support [DATAJPA-1438]

[spring-projects-issues]

László Csatlós opened DATAJPA-1438 and commented

Actual method findAll(Specification<T> spec) on interface JpaSpecificationExecutor<T> can return only with List or Page but not with Stream<T>.

This would be useful for huge resultSets combined with @QueryHint(name="org.hibernate.fetchSize", value="50" ) to truly benefit the power of streaming concept and not to exhaust all of the JVM memory. Hibernate already has the capability to do this with ScrollableResults:

Query query = session.createQuery(query);
query.setReadOnly(true);
query.setFetchSize(100);
ScrollableResults results = query.scroll(ScrollMode.FORWARD_ONLY);

If ambiguous method name is a problem (mentioned at DATAJPA-906), then a
Stream<T> streamAll(Specification<T> spec, Long fetchSize)
signiture would also perfect

---

2 votes, 3 watchers

[gregturn]

Simply put, we don't plan on adding other permutations of the JpaRepository based on alternative return types, like Stream.

If you wish to have an alternative base repository, Spring Data supports this concept. If you look at https://docs.spring.io/spring-data/jpa/docs/current/reference/html/#repositories.definition, it shows how to define your own base repository.

An alternative is to you the fluent API we added as shown in #2274.

[gregturn]

BONUS: Spring Data JPA, in fact, supports ScrollableResult outputs through its PersistenceProvider.