Develop Reference

Mapping JPA native query to DTO returns column name instead of column value [duplicate]

Consider the following method on a Spring Data JPA interface:
#Query("select distinct :columnName from Item i")
List<Item> findByName(#Param("columnName") String columnName);
I would like to use such a method for performing queries dynamically using different column names on the same entity. How can this be done?

You can't. You'll have to implement such a method by yourself. And you won't be able to use parameters: you'll have to use String concatenation or the criteria API. What you'll pass won't be a column name but a field/property name. And it won't return a List<Item>, since you only select one field.

You can use QueryDSL support built into Spring Data. See this tutorial to get started.

First of all you must implement custom Spring Data repository by adding interface:
public interface ItemCustomRepository {
List<Item> findBy(String columnName, String columnValue);
}
then you must extend your current Spring Data repository interface with newly created i.e.:
public interface ItemRepository extends JpaRepository<Item, Long>, ItemCustomRepository, QueryDslPredicateExecutor {
}
and then you must implement your interface using Query DSL dynamic expression feature (the name ItemRepositoryImpl is crucial - it will let you use original Spring Data repository implementation):
public class ItemRepositoryImpl implements ItemCustomRepository {
#Autowired
private ItemRepository itemRepository;
public List<Item> findBy(final String columnName, final String columnValue) {
Path<Item> item = Expressions.path(Item.class, "item");
Path<String> itemColumnName = Expressions.path(String.class, item, columnName);
Expression<String> itemValueExpression = Expressions.constant(columnValue);
BooleanExpression fieldEqualsExpression = Expressions.predicate(Ops.EQ, itemColumnName, itemValueExpression);
return itemRepository.findAll(fieldEqualsExpression);
}
}

Perform a query in Spring Data JPA using an Example to get only the First Projection

In Spring Data, I have a Projection to get only Id of my Entity.
public interface IdOnly<T> {
T getId();
}
I need to filter my query by many fields, many of them can be null and the null values should not be used in the query. So, I intended to use Query by Example to do the job. Besides that, I need to get just the first one Id, limiting the query on database. But unfortunately, it seems like there is no way to do that.
// It works
IdOnly<Long> findFirstBy();
// It works too (although it does not limit the query in database)
Entity findOne(Example<Entity> example);
// It doesn't work. Throws an ClassCastException when returns a non-null entity
IdOnly<Long> findOne(Example<Entity> example);
// That's the one I really need
// The server doesn't start (IndexOutOfBoundsException)
IdOnly<Long> findFirstBy(Example<Entity> example);
So, is there any way to do a query like that, using an Example, to get only the First projection?
Projections
Query by Example
Limiting query results

I really couldn't do it using all three Spring's facilities. So, I preferred to use Example because of the extensive code to test all the fields and get the predicates. This way it's easier to make different queries using the ease of Query By Example.
import static org.springframework.data.jpa.convert.QueryByExamplePredicateBuilder.getPredicate;
#Service
public class EntityService {
#PersistenceContext
private EntityManager em;
public Long findFirstById(Entity entity) {
return getId(em, Entity_.id, Example.of(entity), Entity.class, Long.class);
}
private static <E, N> N getId(EntityManager em, SingularAttribute<E, N> id, Example<E> example, Class<E> entityClass, Class<N> idClass) {
final CriteriaBuilder cb = em.getCriteriaBuilder();
final CriteriaQuery<N> cq = cb.createQuery(idClass);
final Root<E> from = cq.from(entityClass);
cq.select(from.get(id)).where(getPredicate(from, cb, example));
return em.createQuery(cq).setMaxResults(1).getResultList().stream().findFirst().orElse(null);
}
}

How to use createQuery with spring boot

Normally I use annotiations:#Query("SELECT c FROM Country c") with JpaRepositoryor predefined methods like findAll
but in my case I want to generate dynamic query.
String baseQuery =SELECT c FROM Country c`
if(age!=null)
baseQuery+="WHERE c.age=20"
I need to perform same query from code level like this:
Query q1 = em.createQuery("SELECT c FROM Country c");
but I dont use EntityManager in spring boot
How can I generate query from code level?

If you would like to create dynamic queries from code you can take advantage of Spring's JdbcTemplate. Using spring boot it is as simple as injecting JdbcOperations bean to your repository class (assuming you have provided spring-boot-starter-jdbc module to your project).
But remember! This solution uses SQL, not JPQL. That's why you have to use proper tables and columns names in queries and properly map result to objects (i.e. using RowMapper)
This simple example worked fine for me (with different entity, but in same manner - I've adapted it to your example):
#Repository
public class CountryRepository {
#Autowired
private JdbcOperations jdbcOperations;
private static String BASIC_QUERY = "SELECT * FROM COUNTRY";
public List<Country> selectCoutry(Long age){
String query = BASIC_QUERY;
if (age != null){
query += " WHERE AGE = ";
query += age.toString();
}
//let's pretend that Country has constructor Conutry(String name, int age)
return jdbcOperations.query(query, (rs, rowNum) ->
{ return new Country(rs.getString("NAME"), rs.getInt("AGE");}
);
};
}
Then in service or whatever you inject CountryRepository and call method.

Since you're using Spring Boot, you can use Spring Data to create queries in your repository:
#Repository
public interface CountryRepository extends JpaRepository<Country, Long> {
}
Not a 100% on syntax, but should be something similar.
Now you can autowire this class:
#Autowired
public CountryRepository countryRepo;
And all basic methods are already available to you like:
countryRepo.findOne(id);
countryRepo.find();
If you want to make more advanced queries, you can use Spring Data e.g.:
#Repository
public interface CountryRepository extends JpaRepository<Country, Long> {
public Country findByNameAndContinent(String name, String continent);
}
This is just an example (a stupid one) of course and assumes your Country class has the field names 'name' and 'continent' and both are strings. More is available here:
http://docs.spring.io/spring-data/jpa/docs/current/reference/html/
Section 5.3 more specifically.
PS: Make sure your Country class has the #Entity annotation

Spring Data - Spring Repository: selecting distinct values for a class attribute

I have the following
#Entity
public class Restaurant{
#ManyToOne
private City c;
// more
}
#Entity
public class City{
private String name;
// more
}
I also have a repository
public interface RestaurantRepository extends JPARepository<Restaurant, Long> {
// something to put here
}
===EDIT====
I have the list of ALL cities, but only some of them are associated to restaurants.
I need to write a method in such respository to extract all cities that are referred by a Restaurant.
In SQL I would just do the following:
SELECT CITY.id, CITY.name FROM CITY WHERE ID NOT IN (SELECT DISTINCT(city_id) FROM RESTAURANT)
Is there a way to obtain the same result using the name conventions as of http://docs.spring.io/spring-data/jpa/docs/1.4.3.RELEASE/reference/html/jpa.repositories.html ?
Thanks.

As I do not see any support for the IS [NOT] EMPTY operator listed in Spring Data's documentation around query creation strategy, you can try manually defining a JPQL query:
public interface CityRepository extends JPARepository<City, Long> {
#Query(value="SELECT c FROM CITY c WHERE c.restaurants IS EMPTY")
List<City> findCitiesWithNoRestaurants();
}
which I'm guessing will translate to the native SQL you gave as an example.
Otherwise looks like you would need to find a list of cities with Restraunts and then pass that to a method which used the NotIn pattern. This will obviously be less performant than the above.
http://docs.spring.io/spring-data/jpa/docs/current/reference/html/#jpa.query-methods.query-creation

Filter child object in Spring Data Query

I have a following domain model:
Playlist -> List<PlaylistItem> -> Video
#Entity
class Playlist{
// id, name, etc
List<PlaylistItem> playlistItems;
// getters and setters
}
#Entity
class PlaylistItem{
// id, name, etc.
Video video;
// getters and setters
}
#Entity
class Video{
// id, name, etc.
boolean isDeleted;
// getters and setters
}
And my repository:
public interface PlaylistRepository extends JpaRepository<Playlist, Long> {
List<Playlist> findAll();
}
Now, how do I return a playlist with only existing videos, ie, if there are three videos in the database assigned to that playlist item and one of those videos has isDeleted set to true, then I need to get only two items instead.

All you have to do is declare this method on your PlaylistRepository interface:
List<Playlist> findByPlaylistItemsVideoIsDeleted(boolean isDeleted);
And call it like this:
playListRepository.findByPlaylistItemsVideoIsDeleted(false);
That will return all playlist with videos that are not removed.

You may have already resolved this issue, but I thought I would contribute this in hopes it might help you, or anyone else visiting this page.
Using Spring JPA Specifications, you would:
Enable your PlaylistRepository to use JPA Specifications
Write the Specification as a reusable method
Make use of the Specification as the query
Here are the details.
1. Implement JpaSpecificationExecutor
Update PlaylistRepository to implement JpaSpecificationExecutor. This adds find* methods that accept Specification<T> parameters to your PlaylistRepository.
public interface PlaylistRepository extends JpaRepository<Playlist, Long>,
JpaSpecificationExecutor<Playlist> {
}
2. Create the Specification
Create a class with a static method for use in creating a reusable Specification.
public final class PlaylistSpecifications {
private PlaylistSpecifications() {}
public static Specification<Playlist> hasExistingVideos() {
return (root, query, cb) -> {
return cb.equal(root.join("playlistItems").join("video")
.get("isDeleted"), false);
};
}
}
Using root.join (and subsequent joins) is similar to using JOIN in SQL. Here, we are joining on the fields of classes, instead of on columns of tables.
3. Issue the Query
I don't know how you plan to issue your query, but below is an example of how it could be done in a "service" class:
#Service
public class PlaylistService {
#Autowired
private PlaylistRepository playlistRepository;
public List<Playlist> findPlaylistsWithExistingVideos() {
Specification<Playlist> spec = PlaylistSpecifications.hasExistingVideos();
return playlistRepository.findAll(spec);
}
}
Hope this helps!

Maksim, you could use the #query annotation like this :
public interface PlaylistRepository extends JpaRepository<Playlist, Long> {
#Query("select playlist from Playlist playlist
fetch join playlist.playlistItems itens
fetch join itens.video as video
where video.isDeleted = false")
List<Playlist> findAll();
}
Or even better way :
public interface PlaylistRepository extends JpaRepository<Playlist, Long> {
#Query("select playlist from Playlist playlist
fetch join playlist.playlistItems itens
fetch join itens.video as video
where video.isDeleted = :hasVideo ")
List<Playlist> findPlayList(#Param("hasVideo") boolean hasVideo);
}

You can look into Spring Data Specifications. You use them by calling repository.findAll(s);
Specifications allow you add on arbitrary conditions to your query, including the filter you want to add. Another nice thing about Specifications is that they can be type-safe. See here:
http://docs.spring.io/spring-data/jpa/docs/current/reference/html/#specifications