Introduction

In this article, I’m going to show you the best way to write a Spring Data Exists Query that’s efficient from an SQL perspective.

While doing consulting, I’ve encountered several options that are commonly used without developers knowing there are actually better alternatives.

Domain Model

Let’s assume we have the following Post entity:

The slug property is a business key, meaning it has a unique constraint, and, for this reason, we can annotate it with the @NaturalId Hibernate annotation:

@Entity
@Entity
@Table(
    name = "post",
    uniqueConstraints = @UniqueConstraint(
        name = "UK_POST_SLUG",
        columnNames = "slug"
    )
)
public class Post {

    @Id
    private Long id;

    private String title;

    @NaturalId
    private String slug;

    public Long getId() {
        return id;
    }

    public Post setId(Long id) {
        this.id = id;
        return this;
    }

    public String getTitle() {
        return title;
    }

    public Post setTitle(String title) {
        this.title = title;
        return this;
    }

    public Post setSlug(String slug) {
        this.slug = slug;
        return this;
    }
}

How NOT to write an Exists query with Spring Data

First, let’s start with various methods which, while popular, you are better off avoiding.

Emulating existence with a findBy query

Spring Data provides a way to derive queries from method names, so you can write a findBy query to emulate existence, like this:

@Repository
public interface PostRepository 
        extends JpaRepository<Post, Long> {

    Optional<Post> findBySlug(String slug);   
}

Since the findBySlug method is meant for fetching a Post entity, I’ve seen cases when this method was used for equality checks, like in the following example:

assertTrue(
    postRepository.findBySlug(slug).isPresent()
);

The problem with this approach is that an entity is actually fetched just for the sake of checking whether there is an associated record for the provided filtering criteria:

SELECT 
    p.id AS id1_0_,
    p.slug AS slug2_0_,
    p.title AS title3_0_
FROM 
    post p
WHERE 
    p.slug = 'high-performance-java-persistence'

Using a fidnBy query to fetch an entity to check for its existence is a waste of resources since not only you cannot use a covering query in case you have an index on the slug property, but you have to send the entity result set over the network to the JDBC Driver, only to silently discard it.

Checking existence using Query By Example

Another very popular, but inefficient, way to check existence is using the Query By Example feature:

assertTrue(
    postRepository.exists(
        Example.of(
            new Post().setSlug(slug),
            ExampleMatcher.matching()
                .withIgnorePaths(Post_.ID)
                .withMatcher(Post_.SLUG, exact())
        )
    )
);

The Query By Example feature builds a Post entity that is going to be used as a reference when matching the properties given by the provided ExampleMatcher specification.

When executing the above Query By Example method, Spring Data generates the same SQL query that was generated by the previous findBy method:

SELECT 
    p.id AS id1_0_,
    p.slug AS slug2_0_,
    p.title AS title3_0_
FROM 
    post p
WHERE 
    p.slug = 'high-performance-java-persistence'

While the Query By Example feature might be useful for fetching entities, it’s not very efficient to use it with the exists generic method of the Spring Data JPA Repository.

How to write an Exists query with Spring Data

There are better ways to write a Spring Data Exists Query.

Checking existence with an existsBy query method

Spring Data offers an existsBy query method, which we can define in the PostRepository, as follows:

@Repository
public interface PostRepository 
        extends JpaRepository<Post, Long> {

    boolean existsBySlug(String slug);
}

When calling the existsBySlug method on PostgreSQL or MySQL:

assertTrue(
    postRepository.existsBySlug(slug)
);

Spring Data generates the following SQL query:

SELECT 
    p.id AS col_0_0_
FROM 
    post p
WHERE 
    p.slug = 'high-performance-java-persistence'
LIMIT 1

The PostgreSQL Execution Plan for this query looks as follows:

Limit  
    (cost=0.28..8.29 rows=1 width=8) 
    (actual time=0.021..0.021 rows=1 loops=1)
  ->  Index Scan using uk_post_slug on post p  
      (cost=0.28..8.29 rows=1 width=8) 
      (actual time=0.020..0.020 rows=1 loops=1)
        Index Cond: ((slug)::text = 'high-performance-java-persistence'::text)
        
Planning Time: 0.088 ms
Execution Time: 0.033 ms

And, the MySQL one, like this:

-> Limit: 1 row(s)  
   (cost=0.00 rows=1) 
   (actual time=0.001..0.001 rows=1 loops=1)
    -> Rows fetched before execution  
       (cost=0.00 rows=1) 
       (actual time=0.000..0.000 rows=1 loops=1)

So, the query is very fast, and the extra LIMIT operation doesn’t really affect performance since it’s done on a one-record result set anyway.

Checking existence with a COUNT SQL query

Another option to emulate existence is using a COUNT query:

@Repository
public interface PostRepository 
        extends JpaRepository<Post, Long> {

    @Query(value = """
        select count(p.id) = 1 
        from Post p
        where p.slug = :slug
        """
    )
    boolean existsBySlugWithCount(@Param("slug") String slug);
}

The COUNT query works fine in this particular case since we are matching a UNIQUE column value.

However, generally, for queries that return result sets having more than one record, you should prefer using EXISTS instead of COUNT, as explained by Lukas Eder in this article.

When calling the existsBySlugWithCount method on PostgreSQL and MySQL:

assertTrue(
    postRepository.existsBySlugWithCount(slug)
);

Spring Data executes the following SQL query:

SELECT 
    count(p.id) > 0 AS col_0_0_
FROM 
    post p
WHERE 
    p.slug = 'high-performance-java-persistence'

And, the PostgreSQL Execution Plan for this query looks as follows:

Aggregate  
  (cost=8.29..8.31 rows=1 width=1) 
  (actual time=0.023..0.024 rows=1 loops=1)
  ->  Index Scan using uk_post_slug on post p  
      (cost=0.28..8.29 rows=1 width=8) 
      (actual time=0.019..0.020 rows=1 loops=1)
        Index Cond: ((slug)::text = 'high-performance-java-persistence'::text)
        
Planning Time: 0.091 ms
Execution Time: 0.044 ms

And on MySQL:

-> Aggregate: count('1')  
   (actual time=0.002..0.002 rows=1 loops=1)
    -> Rows fetched before execution  
       (cost=0.00 rows=1) 
       (actual time=0.000..0.000 rows=1 loops=1)

Although there’s an extra Aggregate step for the COUNT operator, that step is very fast since there’s a single record to count.

Checking existence with a CASE WHEN EXISTS SQL query

And the last option to emulate existence is using a CASE WHEN EXISTS native SQL query:

@Repository
public interface PostRepository 
        extends JpaRepository<Post, Long> {

    @Query(value = """
        SELECT 
            CASE WHEN EXISTS (
                SELECT 1 
                FROM post 
                WHERE slug = :slug
            ) 
            THEN 'true' 
            ELSE 'false'
            END
        """,
        nativeQuery = true
    )
    boolean existsBySlugWithCase(@Param("slug") String slug);
}

And, we can call the existsBySlugWithCase method like this:

assertTrue(
    postRepository.existsBySlugWithCase(slug)
);

The PostgreSQL Execution Plan for this query looks as follows:

Result  
  (cost=8.29..8.29 rows=1 width=1) 
  (actual time=0.021..0.022 rows=1 loops=1)
  InitPlan 1 (returns $0)
    ->  Index Only Scan using uk_post_slug on post  
          (cost=0.27..8.29 rows=1 width=0) 
          (actual time=0.020..0.020 rows=1 loops=1)
          Index Cond: (slug = 'high-performance-java-persistence'::text)
          Heap Fetches: 1
          
Planning Time: 0.097 ms
Execution Time: 0.037 ms

And on MySQL:

-> Rows fetched before execution  
   (cost=0.00 rows=1) 
   (actual time=0.000..0.000 rows=1 loops=1)
-> Select #2 (subquery in projection; run only once)
    -> Limit: 1 row(s)  
        (cost=0.00 rows=1) 
        (actual time=0.000..0.001 rows=1 loops=1)
        -> Rows fetched before execution  
           (cost=0.00 rows=1) 
           (actual time=0.000..0.000 rows=1 loops=1)

So, this is just as fast as the previous LIMIT and COUNT queries. On other databases, you might want to check out to see if there’s any difference.

I'm running an online workshop on the 11th of October about High-Performance SQL.

If you enjoyed this article, I bet you are going to love my Book and Video Courses as well.

Conclusion

Therefore, if you want to check a record’s existence with Spring Data, the easiest way to do so is using the existsBy query method.

And, if the query is more complex and you cannot express it with the Spring Data query methods, you can use either a COUNT or a CASE WHEN EXISTS query since they are just as fast.

Insert details about how the information is going to be processed

DOWNLOAD NOW