Introduction

If you’ve been using Hibernate long enough, then you surely must have seen this WARN log message when doing pagination while join-fetching multiple entities.

HHH000104: firstResult/maxResults specified with collection fetch; applying in memory!

In this article, I’m going to show two ways you can to fix this issue.

Domain Model

Considering we have the following entities:

The Post entity has a bidirectional @OneToMany association with the PostComment child entity.

Notice that both entities use the Fluent-style API. For more details about building entities using a Fluent-style API with JPA and Hibernate, check out this article.

Now, let’s assume we create 50 Post entities, each one with several PostComment child entities.

LocalDateTime timestamp = LocalDateTime
.of(
    2018, 10, 9, 12, 0, 0, 0
);

LongStream.rangeClosed(1, 50)
.forEach(postId -> {
    Post post = new Post()
    .setId(postId)
    .setTitle(
        String.format("High-Performance Java Persistence - Chapter %d",
        postId)
    )
    .setCreatedOn(
        Timestamp.valueOf(timestamp.plusMinutes(postId))
    );

    LongStream.rangeClosed(1, COMMENT_COUNT)
    .forEach(commentOffset -> {
        long commentId = ((postId - 1) * COMMENT_COUNT) + commentOffset;

        post.addComment(
            new PostComment()
            .setId(commentId)
            .setReview(
                String.format("Comment nr. %d - A must read!", commentId)
            )
            .setCreatedOn(
                Timestamp.valueOf(timestamp.plusMinutes(commentId))
            )
        );

    });

    entityManager.persist(post);
});

The problem

We want to fetch all Post entities whose titles match a given pattern. However, we want to fetch the associated PostComment entities as well.

As I explained in this article, you might be tempted to use a JPA pagination query to fetch the Post entities while also join fetching the PostComment entities as illustrated by the following JPQL query:

List<Post> posts  = entityManager.createQuery("""
    select p
    from Post p
    left join fetch p.comments
    where p.title like :titlePattern
    order by p.createdOn
    """, Post.class)
.setParameter(
    "titlePattern", 
    "High-Performance Java Persistence %"
)
.setMaxResults(5)
.getResultList();

We want to fetch the Post entities along with their comments and limit the result set to a maximum number of entries.

When running the JPQL query above, Hibernate limits the number of Post entries, but it issues the aforementioned warning while executing the following SQL query:

-- HHH000104: firstResult/maxResults specified with collection fetch; 
-- applying in memory!

SELECT 
    p.id AS id1_0_0_,
    pc.id AS id1_1_1_,
    p.created_on AS created_2_0_0_,
    p.title AS title3_0_0_,
    pc.created_on AS created_2_1_1_,
    pc.post_id AS post_id4_1_1_,
    pc.review AS review3_1_1_,
    pc.post_id AS post_id4_1_0__,
    pc.id AS id1_1_0__
FROM 
     post p
LEFT OUTER JOIN 
     post_comment pc ON p.id=pc.post_id
WHERE 
     p.title LIKE 'High-Performance Java Persistence %'
ORDER BY 
     p.created_on

Notice that the SQL query uses no pagination whatsoever. As mentioned by the HHH000104 warning message, the pagination is done in memory, which is bad.

The reason why Hibernate does the pagination in memory is that it cannot just truncate the result set using SQL-level pagination. If it did that, then the result set would be truncated in the middle of the PostComments rows, therefore returning a Post entity with just a subset of comments.

Because Hibernate favors consistency, it fetches the entire result set and does the pagination in memory. However, that can be suboptimal, so what can we do about it?

Fixing the issue with two SQL queries that can fetch entities in read-write mode

The easiest way to fix this issue is to execute two queries:

. The first query will fetch the Post entity identifiers matching the provided filtering criteria.
. The second query will use the previously extracted Post entity identifiers to fetch the Post and the PostComment entities.

This approach is very easy to implement and looks as follows:

List<Long> postIds = entityManager.createQuery("""
    select p.id
    from Post p
    where p.title like :titlePattern
    order by p.createdOn
    """, Long.class)
.setParameter(
    "titlePattern", 
    "High-Performance Java Persistence %"
)
.setMaxResults(5)
.getResultList();

List<Post> posts = entityManager.createQuery("""
    select distinct p
    from Post p
    left join fetch p.comments
    where p.id in (:postIds)
    order by p.createdOn
    """, Post.class)
.setParameter("postIds", postIds)
.setHint(
    QueryHints.HINT_PASS_DISTINCT_THROUGH, 
    false
)
.getResultList();

assertEquals(5, posts.size());

Post post1 = posts.get(0);

List<PostComment> comments = post1.getComments();

for (int i = 0; i < COMMENT_COUNT - 1; i++) {
    PostComment postComment1 = comments.get(i);

    assertEquals(
        String.format(
            "Comment nr. %d - A must read!",
            i + 1
        ),
        postComment1.getReview()
    );
}

Notice the hibernate.query.passDistinctThrough JPA query hint we used to instruct Hibernate to prevent passing the JPQL DISTINCT keyword to the underlying SQL query. For more details about this query hint, check out this article.

The second query also requires the ORDER BY clause, as, without it, the ordering of the Post records will not be guaranteed.

When executing the two JPQL queries above, Hibernate generates the following SQL queries:

Query:["
    SELECT 
        p.id AS col_0_0_
    FROM 
        post p
    WHERE 
        p.title LIKE ?
    ORDER BY 
        p.created_on
    LIMIT 
        ?
"], 
Params:[(
    'High-Performance Java Persistence %', 5
)]

Query:["
    SELECT 
        p.id AS id1_0_0_,
        pc.id AS id1_1_1_,
        p.created_on AS created_2_0_0_,
        p.title AS title3_0_0_,
        pc.created_on AS created_2_1_1_,
        pc.post_id AS post_id4_1_1_,
        pc.review AS review3_1_1_,
        pc.post_id AS post_id4_1_0__,
        pc.id AS id1_1_0__
    FROM 
        post p
    LEFT OUTER JOIN 
        post_comment pc ON p.id = pc.post_id
    WHERE 
        p.id IN (?, ?, ?, ?, ?)
    ORDER BY 
        p.created_on
"], 
Params:[(
    1, 2, 3, 4, 5
)]

That’s the easiest way to fix the issue causing the HHH000104 warning message.

Fixing the issue with one SQL query that can only fetch entities in read-only mode

As I already explained, Window Functions are the answer to many query-related problems.

So, we just need to calculate the DENSE_RANK over the result set of post and post_comments that match our filtering criteria and restrict the output for the first N post entries only.

For that, we need to define the following @NamedNativeQuery along with its associated @SqlResultSetMapping:

@NamedNativeQuery(
    name = "PostWithCommentByRank",
    query = """
        SELECT *
        FROM (
            SELECT 
                *, 
                DENSE_RANK() OVER (
                    ORDER BY "p.created_on", "p.id"
                ) rank
            FROM (
                SELECT 
                    p.id AS "p.id", p.created_on AS "p.created_on",
                    p.title AS "p.title", pc.post_id AS "pc.post_id",
                    pc.id as "pc.id", pc.created_on AS "pc.created_on",
                    pc.review AS "pc.review"
                FROM  post p
                LEFT JOIN post_comment pc ON p.id = pc.post_id
                WHERE p.title LIKE :titlePattern
                ORDER BY p.created_on
            ) p_pc
        ) p_pc_r
        WHERE p_pc_r.rank <= :rank
        """,
    resultSetMapping = "PostWithCommentByRankMapping"
)
@SqlResultSetMapping(
    name = "PostWithCommentByRankMapping",
    entities = {
        @EntityResult(
            entityClass = Post.class,
            fields = {
                @FieldResult(name = "id", column = "p.id"),
                @FieldResult(name = "createdOn", column = "p.created_on"),
                @FieldResult(name = "title", column = "p.title"),
            }
        ),
        @EntityResult(
            entityClass = PostComment.class,
            fields = {
                @FieldResult(name = "id", column = "pc.id"),
                @FieldResult(name = "createdOn", column = "pc.created_on"),
                @FieldResult(name = "review", column = "pc.review"),
                @FieldResult(name = "post", column = "pc.post_id"),
            }
        )
    }
)

The @NamedNativeQuery fetches all Post entities matching the provided title along with their associated PostComment child entities. The DENSE_RANK Window Function is used to assign the rank for each Post and PostComment joined record so that we can later filter just the amount of Post records we are interested in fetching.

The SqlResultSetMapping provides the mapping between the SQL-level column aliases and the JPA entity properties that need to be populated.

For more details about the best way to use the JPA SqlResultSetMapping annotation, you should read this article.

Now, we can execute the PostWithCommentByRank @NamedNativeQuery:

List<Post> posts = entityManager
.createNamedQuery("PostWithCommentByRank")
.setParameter(
    "titlePattern", 
    "High-Performance Java Persistence %"
)
.setParameter(
    "rank", 
    5
)
.unwrap(NativeQuery.class)
.setResultTransformer(
    new DistinctPostResultTransformer(entityManager)
)
.getResultList();

assertEquals(5, posts.size());

Post post1 = posts.get(0);

List<PostComment> comments = post1.getComments();

for (int i = 0; i < COMMENT_COUNT - 1; i++) {
    PostComment postComment1 = comments.get(i);

    assertEquals(
        String.format(
            "Comment nr. %d - A must read!",
            i + 1
        ),
        postComment1.getReview()
    );
}

We used the READONLY JPA query hint to instruct Hibernate to discard the underlying entity detached state. For more details about this optimization, check out this article.

Now, by default, a native SQL query like the PostWithCommentByRank one would fetch the Post and the PostComment in the same JDBC row, so we will end up with an Object[] containing both entities.

However, we want to transform the tabular Object[] array into a tree of parent-child entities, and for this reason, we need to use the Hibernate ResultTransformer For more details about the ResultTransformer, check out this article.

The DistinctPostResultTransformer looks as follows:

public class DistinctPostResultTransformer 
        extends BasicTransformerAdapter {

    private final EntityManager entityManager;

    public DistinctPostResultTransformer(
            EntityManager entityManager) {
        this.entityManager = entityManager;
    }

    @Override
    public List transformList(
            List list) {
            
        Map<Serializable, Identifiable> identifiableMap = 
            new LinkedHashMap<>(list.size());
            
        for (Object entityArray : list) {
            if (Object[].class.isAssignableFrom(entityArray.getClass())) {
                Post post = null;
                PostComment comment = null;

                Object[] tuples = (Object[]) entityArray;

                for (Object tuple : tuples) {
                    if(tuple instanceof Identifiable) {
                        entityManager.detach(tuple);

                        if (tuple instanceof Post) {
                            post = (Post) tuple;
                        } 
                        else if (tuple instanceof PostComment) {
                            comment = (PostComment) tuple;
                        } 
                        else {
                            throw new UnsupportedOperationException(
                                "Tuple " + tuple.getClass() + " is not supported!"
                            );
                        }
                    }
                }

                if (post != null) {
                    if (!identifiableMap.containsKey(post.getId())) {
                        identifiableMap.put(post.getId(), post);
                        post.setComments(new ArrayList<>());
                    }
                    if (comment != null) {
                        post.addComment(comment);
                    }
                }
            }
        }
        return new ArrayList<>(identifiableMap.values());
    }
}

The DistinctPostResultTransformer must detach the entities being fetched because we are overwriting the child collection and we don’t want that to be propagated as an entity state transition:

post.setComments(new ArrayList<>());

Now, not only that we can fetch both the Post and its PostComments with a single query, but we can even later modify these entities and merge them back in a subsequent read-write transaction:

List<Post> posts = doInJPA(entityManager -> {
    return entityManager
    .createNamedQuery("PostWithCommentByRank")
    .setParameter(
        "titlePattern",
        "High-Performance Java Persistence %"
    )
    .setParameter(
        "rank",
        2
    )
    .unwrap(NativeQuery.class)
    .setResultTransformer(
        new DistinctPostResultTransformer(entityManager)
    )
    .getResultList();
});

assertEquals(2, posts.size());

Post post1 = posts.get(0);

post1.addComment(
    new PostComment()
    .setId((post1.getId() - 1) * COMMENT_COUNT)
    .setReview("Awesome!")
    .setCreatedOn(
        Timestamp.valueOf(LocalDateTime.now())
    )
);

Post post2 = posts.get(1);
post2.removeComment(post2.getComments().get(0));

doInJPA(entityManager -> {
    entityManager.merge(post1);
    entityManager.merge(post2);
});

And Hibernate will properly propagate the changes to the database:

INSERT INTO post_comment (
    created_on, 
    post_id, 
    review, 
    id
) 
VALUES (
    '2019-01-09 10:47:32.134', 
    1, 
    'Awesome!', 
    0
)

DELETE FROM post_comment 
WHERE id = 6

Awesome, right?

Online Workshops

If you enjoyed this article, I bet you are going to love my upcoming Online Workshops!

• Caching Best Practices with JPA and Hibernate (2.5 hours) on the 30th of September
• High-Performance SQL (4 hours) on the 6th of October in collaboration with Voxxed Days Ticino
• High-Performance SQL (12 hours) starting on the 28th of October in collaboration with Bouvet

Conclusion

So, to fix the HHH000104 issue, you have two options. Either you execute two queries and fetch the entities in either read-write or read-only mode, or you use a single query with Window Functions to fetch the entities in read-only mode.

Insert details about how the information is going to be processed

DOWNLOAD NOW