Derived Query Methods in Spring Data JPA Repositories

1. Introduction

For simple queries, it’s easy to derive what the query should be just by looking at the corresponding method name in our code.

In this tutorial, we’ll explore how Spring Data JPA leverages this idea in the form of a method naming convention.

2. Structure of Derived Query Methods in Spring

Derived method names have two main parts separated by the first By keyword:

List<User> findByName(String name)

The first part – like find – is the introducer and the rest – like ByName – is the criteria.

Spring Data JPA supports find, read, query, count and get. So, for example, we could have done queryByName and Spring Data would behave the same.

We can also use Distinct, First, or Top to remove duplicates or limit our result set:

List<User> findTop3ByAge()

The criteria part contains the entity-specific condition expressions of the query. We can use the condition keywords along with the entity’s property names. We can also concatenate the expressions with And and Or, as well see in just a moment.

3. Sample Application

First, we’ll, of course, need an application using Spring Data JPA.

In that application, let’s define an entity class:

@Table(name = "users")
@Entity
class User {
    @Id
    @GeneratedValue
    private Integer id;
    
    private String name;
    private Integer age;
    private ZonedDateTime birthDate;
    private Boolean active;

    // standard getters and setters
}

And, let’s also define a repository. It’ll extend JpaRepository, one of the Spring Data Repository types:

interface UserRepository extends JpaRepository<User, Integer> {}

This is where we’ll place all our derived query methods.

4. Equality Condition Keywords

Exact equality is one of the most-used conditions in queries. We have several options to express = or IS operators in the query.

We can just append the property name without any keyword for an exact match condition:

List<User> findByName(String name);

And we can add Is or Equals for readability:

List<User> findByNameIs(String name);
List<User> findByNameEquals(String name);

This extra readability comes in handy when we need to express inequality instead:

List<User> findByNameIsNot(String name);

This is quite a bit more readable than findByNameNot(String)!

As null equality is a special case, we shouldn’t use the = operator. Spring Data JPA handles null parameters by default. So, when we pass a null value for an equality condition, Spring interprets the query as IS NULL in the generated SQL.

We can also use the IsNull keyword to add IS NULL criteria to the query:

List<User> findByNameIsNull();
List<User> findByNameIsNotNull();

Note that, neither IsNull nor IsNotNull requires a method argument.

There are also two more keywords that don’t require any arguments. We can use True and False keywords to add equality conditions for boolean types:

List<User> findByActiveTrue();
List<User> findByActiveFalse();

Of course, sometimes we want something more lenient than exact equality, let’s see what else we can do.

5. Similarity Condition Keywords

When we need to query the results with a pattern of a property, we have a few options.

We can find names that start with a value using StartingWith:

List<User> findByNameStartingWith(String prefix);

Roughly, this translates to “WHERE name LIKE ‘value%’“.

If we want names that end with a value, then EndingWith is what we want:

List<User> findByNameEndingWith(String suffix);

Or, we can find which names contain a value with Containing:

List<User> findByNameContaining(String infix);

Note that all conditions above are called predefined pattern expressions. So, we don’t need to add % operator inside the argument when these methods are called.

But, let’s suppose we are doing something more complex. Say we need to fetch the users whose names start with an a, contain b, and end with c.

For that, we can add our own LIKE with the Like keyword:

List<User> findByNameLike(String likePattern);

And we can then hand in our LIKE pattern when we call the method:

String likePattern = "a%b%c";
userRepository.findByNameLike(likePattern);

That’s enough about names for now. Let’s try some other values in User.

6. Comparison Condition Keywords

Furthermore, we can use LessThan and LessThanEqual keywords to compare the records with the given value using the < and <= operators:

List<User> findByAgeLessThan(Integer age);
List<User> findByAgeLessThanEqual(Integer age);

On the other hand, in the opposite situation, we can use GreaterThan and GreaterThanEqual keywords:

List<User> findByAgeGreaterThan(Integer age);
List<User> findByAgeGreaterThanEqual(Integer age);

Or, we can find users who are between two ages with Between:

List<User> findByAgeBetween(Integer startAge, Integer endAge);

We can also supply a collection of ages to match against using In:

List<User> findByAgeIn(Collection<Integer> ages);

Since we know the users’ birthdates, we might want to query for users who were born before or after a given date. We’d use Before and After for that:

List<User> findByBirthDateAfter(ZonedDateTime birthDate);
List<User> findByBirthDateBefore(ZonedDateTime birthDate);

7. Multiple Condition Expressions

We can combine as many expressions as we need by using And and Or keywords:

List<User> findByNameOrBirthDate(String name, ZonedDateTime birthDate);
List<User> findByNameOrBirthDateAndActive(String name, ZonedDateTime birthDate, Boolean active);

The precedence order is And then Or, just like Java.

While Spring Data JPA imposes no limit to how many expressions we can add, we shouldn’t go crazy here. Long names are unreadable and hard to maintain. For complex queries, take a look at the @Query annotation instead.

8. Sorting the Results

Next up is sorting. We could ask that the users be sorted alphabetically by their name using OrderBy:

List<User> findByNameOrderByName(String name);
List<User> findByNameOrderByNameAsc(String name);

Ascending order is the default sorting option, but we can use Desc instead to sort them in reverse:

List<User> findByNameOrderByNameDesc(String name);

9. findOne vs findById in a CrudRepository

The Spring team made some major changes in CrudRepository with Spring Boot 2.x. One of them is renaming findOne to findById.

Previously with Spring Boot 1.x, we’d call findOne when we wanted to retrieve an entity by its primary key:

User user = userRepository.findOne(1);

Since Spring Boot 2.x we can do the same with findById:

User user = userRepository.findById(1);

Note that the findById() method is already defined in CrudRepository for us. So we don’t have to define it explicitly in custom repositories that extend CrudRepository.

10. Conclusion

In this article, we explained the query derivation mechanism in Spring Data JPA. We used the property condition keywords to write derived query methods in Spring Data JPA repositories.

The source code of this tutorial is available on the Github project.