Related
Is it possible to use a DB sequence for some column that is not the identifier/is not part of a composite identifier?
I'm using hibernate as jpa provider, and I have a table that has some columns that are generated values (using a sequence), although they are not part of the identifier.
What I want is to use a sequence to create a new value for an entity, where the column for the sequence is NOT (part of) the primary key:
#Entity
#Table(name = "MyTable")
public class MyEntity {
//...
#Id //... etc
public Long getId() {
return id;
}
//note NO #Id here! but this doesn't work...
#GeneratedValue(strategy = GenerationType.AUTO, generator = "myGen")
#SequenceGenerator(name = "myGen", sequenceName = "MY_SEQUENCE")
#Column(name = "SEQ_VAL", unique = false, nullable = false, insertable = true, updatable = true)
public Long getMySequencedValue(){
return myVal;
}
}
Then when I do this:
em.persist(new MyEntity());
the id will be generated, but the mySequenceVal property will be also generated by my JPA provider.
Just to make things clear: I want Hibernate to generate the value for the mySequencedValue property. I know Hibernate can handle database-generated values, but I don't want to use a trigger or any other thing other than Hibernate itself to generate the value for my property. If Hibernate can generate values for primary keys, why can't it generate for a simple property?
Looking for answers to this problem, I stumbled upon this link
It seems that Hibernate/JPA isn't able to automatically create a value for your non-id-properties. The #GeneratedValue annotation is only used in conjunction with #Id to create auto-numbers.
The #GeneratedValue annotation just tells Hibernate that the database is generating this value itself.
The solution (or work-around) suggested in that forum is to create a separate entity with a generated Id, something like this:
#Entity
public class GeneralSequenceNumber {
#Id
#GeneratedValue(...)
private Long number;
}
#Entity
public class MyEntity {
#Id ..
private Long id;
#OneToOne(...)
private GeneralSequnceNumber myVal;
}
I found that #Column(columnDefinition="serial") works perfect but only for PostgreSQL. For me this was perfect solution, because second entity is "ugly" option.
A call to saveAndFlush on the entity is also necessary, and save won't be enough to populate the value from the DB.
I know this is a very old question, but it's showed firstly upon the results and jpa has changed a lot since the question.
The right way to do it now is with the #Generated annotation. You can define the sequence, set the default in the column to that sequence and then map the column as:
#Generated(GenerationTime.INSERT)
#Column(name = "column_name", insertable = false)
Hibernate definitely supports this. From the docs:
"Generated properties are properties which have their values generated by the database. Typically, Hibernate applications needed to refresh objects which contain any properties for which the database was generating values. Marking properties as generated, however, lets the application delegate this responsibility to Hibernate. Essentially, whenever Hibernate issues an SQL INSERT or UPDATE for an entity which has defined generated properties, it immediately issues a select afterwards to retrieve the generated values."
For properties generated on insert only, your property mapping (.hbm.xml) would look like:
<property name="foo" generated="insert"/>
For properties generated on insert and update your property mapping (.hbm.xml) would look like:
<property name="foo" generated="always"/>
Unfortunately, I don't know JPA, so I don't know if this feature is exposed via JPA (I suspect possibly not)
Alternatively, you should be able to exclude the property from inserts and updates, and then "manually" call session.refresh( obj ); after you have inserted/updated it to load the generated value from the database.
This is how you would exclude the property from being used in insert and update statements:
<property name="foo" update="false" insert="false"/>
Again, I don't know if JPA exposes these Hibernate features, but Hibernate does support them.
I fixed the generation of UUID (or sequences) with Hibernate using #PrePersist annotation:
#PrePersist
public void initializeUUID() {
if (uuid == null) {
uuid = UUID.randomUUID().toString();
}
}
Looks like thread is old, I just wanted to add my solution here(Using AspectJ - AOP in spring).
Solution is to create a custom annotation #InjectSequenceValue as follows.
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface InjectSequenceValue {
String sequencename();
}
Now you can annotate any field in entity, so that the underlying field (Long/Integer) value will be injected at runtime using the nextvalue of the sequence.
Annotate like this.
//serialNumber will be injected dynamically, with the next value of the serialnum_sequence.
#InjectSequenceValue(sequencename = "serialnum_sequence")
Long serialNumber;
So far we have marked the field we need to inject the sequence value.So we will look how to inject the sequence value to the marked fields, this is done by creating the point cut in AspectJ.
We will trigger the injection just before the save/persist method is being executed.This is done in the below class.
#Aspect
#Configuration
public class AspectDefinition {
#Autowired
JdbcTemplate jdbcTemplate;
//#Before("execution(* org.hibernate.session.save(..))") Use this for Hibernate.(also include session.save())
#Before("execution(* org.springframework.data.repository.CrudRepository.save(..))") //This is for JPA.
public void generateSequence(JoinPoint joinPoint){
Object [] aragumentList=joinPoint.getArgs(); //Getting all arguments of the save
for (Object arg :aragumentList ) {
if (arg.getClass().isAnnotationPresent(Entity.class)){ // getting the Entity class
Field[] fields = arg.getClass().getDeclaredFields();
for (Field field : fields) {
if (field.isAnnotationPresent(InjectSequenceValue.class)) { //getting annotated fields
field.setAccessible(true);
try {
if (field.get(arg) == null){ // Setting the next value
String sequenceName=field.getAnnotation(InjectSequenceValue.class).sequencename();
long nextval=getNextValue(sequenceName);
System.out.println("Next value :"+nextval); //TODO remove sout.
field.set(arg, nextval);
}
} catch (Exception e) {
e.printStackTrace();
}
}
}
}
}
}
/**
* This method fetches the next value from sequence
* #param sequence
* #return
*/
public long getNextValue(String sequence){
long sequenceNextVal=0L;
SqlRowSet sqlRowSet= jdbcTemplate.queryForRowSet("SELECT "+sequence+".NEXTVAL as value FROM DUAL");
while (sqlRowSet.next()){
sequenceNextVal=sqlRowSet.getLong("value");
}
return sequenceNextVal;
}
}
Now you can annotate any Entity as below.
#Entity
#Table(name = "T_USER")
public class UserEntity {
#Id
#SequenceGenerator(sequenceName = "userid_sequence",name = "this_seq")
#GeneratedValue(strategy = GenerationType.SEQUENCE,generator = "this_seq")
Long id;
String userName;
String password;
#InjectSequenceValue(sequencename = "serialnum_sequence") // this will be injected at the time of saving.
Long serialNumber;
String name;
}
As a followup here's how I got it to work:
#Override public Long getNextExternalId() {
BigDecimal seq =
(BigDecimal)((List)em.createNativeQuery("select col_msd_external_id_seq.nextval from dual").getResultList()).get(0);
return seq.longValue();
}
If you are using postgresql
And i'm using in spring boot 1.5.6
#Column(columnDefinition = "serial")
#Generated(GenerationTime.INSERT)
private Integer orderID;
Although this is an old thread I want to share my solution and hopefully get some feedback on this. Be warned that I only tested this solution with my local database in some JUnit testcase. So this is not a productive feature so far.
I solved that issue for my by introducing a custom annotation called Sequence with no property. It's just a marker for fields that should be assigned a value from an incremented sequence.
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface Sequence
{
}
Using this annotation i marked my entities.
public class Area extends BaseEntity implements ClientAware, IssuerAware
{
#Column(name = "areaNumber", updatable = false)
#Sequence
private Integer areaNumber;
....
}
To keep things database independent I introduced an entity called SequenceNumber which holds the sequence current value and the increment size. I chose the className as unique key so each entity class wil get its own sequence.
#Entity
#Table(name = "SequenceNumber", uniqueConstraints = { #UniqueConstraint(columnNames = { "className" }) })
public class SequenceNumber
{
#Id
#Column(name = "className", updatable = false)
private String className;
#Column(name = "nextValue")
private Integer nextValue = 1;
#Column(name = "incrementValue")
private Integer incrementValue = 10;
... some getters and setters ....
}
The last step and the most difficult is a PreInsertListener that handles the sequence number assignment. Note that I used spring as bean container.
#Component
public class SequenceListener implements PreInsertEventListener
{
private static final long serialVersionUID = 7946581162328559098L;
private final static Logger log = Logger.getLogger(SequenceListener.class);
#Autowired
private SessionFactoryImplementor sessionFactoryImpl;
private final Map<String, CacheEntry> cache = new HashMap<>();
#PostConstruct
public void selfRegister()
{
// As you might expect, an EventListenerRegistry is the place with which event listeners are registered
// It is a service so we look it up using the service registry
final EventListenerRegistry eventListenerRegistry = sessionFactoryImpl.getServiceRegistry().getService(EventListenerRegistry.class);
// add the listener to the end of the listener chain
eventListenerRegistry.appendListeners(EventType.PRE_INSERT, this);
}
#Override
public boolean onPreInsert(PreInsertEvent p_event)
{
updateSequenceValue(p_event.getEntity(), p_event.getState(), p_event.getPersister().getPropertyNames());
return false;
}
private void updateSequenceValue(Object p_entity, Object[] p_state, String[] p_propertyNames)
{
try
{
List<Field> fields = ReflectUtil.getFields(p_entity.getClass(), null, Sequence.class);
if (!fields.isEmpty())
{
if (log.isDebugEnabled())
{
log.debug("Intercepted custom sequence entity.");
}
for (Field field : fields)
{
Integer value = getSequenceNumber(p_entity.getClass().getName());
field.setAccessible(true);
field.set(p_entity, value);
setPropertyState(p_state, p_propertyNames, field.getName(), value);
if (log.isDebugEnabled())
{
LogMF.debug(log, "Set {0} property to {1}.", new Object[] { field, value });
}
}
}
}
catch (Exception e)
{
log.error("Failed to set sequence property.", e);
}
}
private Integer getSequenceNumber(String p_className)
{
synchronized (cache)
{
CacheEntry current = cache.get(p_className);
// not in cache yet => load from database
if ((current == null) || current.isEmpty())
{
boolean insert = false;
StatelessSession session = sessionFactoryImpl.openStatelessSession();
session.beginTransaction();
SequenceNumber sequenceNumber = (SequenceNumber) session.get(SequenceNumber.class, p_className);
// not in database yet => create new sequence
if (sequenceNumber == null)
{
sequenceNumber = new SequenceNumber();
sequenceNumber.setClassName(p_className);
insert = true;
}
current = new CacheEntry(sequenceNumber.getNextValue() + sequenceNumber.getIncrementValue(), sequenceNumber.getNextValue());
cache.put(p_className, current);
sequenceNumber.setNextValue(sequenceNumber.getNextValue() + sequenceNumber.getIncrementValue());
if (insert)
{
session.insert(sequenceNumber);
}
else
{
session.update(sequenceNumber);
}
session.getTransaction().commit();
session.close();
}
return current.next();
}
}
private void setPropertyState(Object[] propertyStates, String[] propertyNames, String propertyName, Object propertyState)
{
for (int i = 0; i < propertyNames.length; i++)
{
if (propertyName.equals(propertyNames[i]))
{
propertyStates[i] = propertyState;
return;
}
}
}
private static class CacheEntry
{
private int current;
private final int limit;
public CacheEntry(final int p_limit, final int p_current)
{
current = p_current;
limit = p_limit;
}
public Integer next()
{
return current++;
}
public boolean isEmpty()
{
return current >= limit;
}
}
}
As you can see from the above code the listener used one SequenceNumber instance per entity class and reserves a couple of sequence numbers defined by the incrementValue of the SequenceNumber entity. If it runs out of sequence numbers it loads the SequenceNumber entity for the target class and reserves incrementValue values for the next calls. This way I do not need to query the database each time a sequence value is needed.
Note the StatelessSession that is being opened for reserving the next set of sequence numbers. You cannot use the same session the target entity is currently persisted since this would lead to a ConcurrentModificationException in the EntityPersister.
Hope this helps someone.
I run in the same situation like you and I also didn't find any serious answers if it is basically possible to generate non-id propertys with JPA or not.
My solution is to call the sequence with a native JPA query to set the property by hand before persisiting it.
This is not satisfying but it works as a workaround for the moment.
Mario
I've found this specific note in session 9.1.9 GeneratedValue Annotation from JPA specification:
"[43] Portable applications should not use the GeneratedValue annotation on other persistent fields or properties."
So, I presume that it is not possible to auto generate value for non primary key values at least using simply JPA.
You can do exactly what you are asking.
I've found it is possible to adapt Hibernate's IdentifierGenerator implementations by registering them with an Integrator. With this you should be able to use any id sequence generator provided by Hibernate to generate sequences for non-id fields (presumably the non-sequential id generators would work as well).
There are quite a few options for generating ids this way. Check out some of the implementations of IdentifierGenerator, specifically SequenceStyleGenerator and TableGenerator. If you have configured generators using the #GenericGenerator annotation, then the parameters for these classes may be familiar to you. This would also have the advantage of using Hibernate to generate the SQL.
Here is how I got it working:
import org.hibernate.Session;
import org.hibernate.boot.Metadata;
import org.hibernate.engine.spi.SessionFactoryImplementor;
import org.hibernate.id.IdentifierGenerator;
import org.hibernate.id.enhanced.TableGenerator;
import org.hibernate.integrator.spi.Integrator;
import org.hibernate.internal.SessionImpl;
import org.hibernate.service.spi.SessionFactoryServiceRegistry;
import org.hibernate.tuple.ValueGenerator;
import org.hibernate.type.LongType;
import java.util.Properties;
public class SequenceIntegrator implements Integrator, ValueGenerator<Long> {
public static final String TABLE_NAME = "SEQUENCE_TABLE";
public static final String VALUE_COLUMN_NAME = "NEXT_VAL";
public static final String SEGMENT_COLUMN_NAME = "SEQUENCE_NAME";
private static SessionFactoryServiceRegistry serviceRegistry;
private static Metadata metadata;
private static IdentifierGenerator defaultGenerator;
#Override
public void integrate(Metadata metadata, SessionFactoryImplementor sessionFactoryImplementor, SessionFactoryServiceRegistry sessionFactoryServiceRegistry) {
//assigning metadata and registry to fields for use in a below example
SequenceIntegrator.metadata = metadata;
SequenceIntegrator.serviceRegistry = sessionFactoryServiceRegistry;
SequenceIntegrator.defaultGenerator = getTableGenerator(metadata, sessionFactoryServiceRegistry, "DEFAULT");
}
private TableGenerator getTableGenerator(Metadata metadata, SessionFactoryServiceRegistry sessionFactoryServiceRegistry, String segmentValue) {
TableGenerator generator = new TableGenerator();
Properties properties = new Properties();
properties.setProperty("table_name", TABLE_NAME);
properties.setProperty("value_column_name", VALUE_COLUMN_NAME);
properties.setProperty("segment_column_name", SEGMENT_COLUMN_NAME);
properties.setProperty("segment_value", segmentValue);
//any type should work if the generator supports it
generator.configure(LongType.INSTANCE, properties, sessionFactoryServiceRegistry);
//this should create the table if ddl auto update is enabled and if this function is called inside of the integrate method
generator.registerExportables(metadata.getDatabase());
return generator;
}
#Override
public Long generateValue(Session session, Object o) {
// registering additional generators with getTableGenerator will work here. inserting new sequences can be done dynamically
// example:
// TableGenerator classSpecificGenerator = getTableGenerator(metadata, serviceRegistry, o.getClass().getName());
// return (Long) classSpecificGenerator.generate((SessionImpl)session, o);
return (Long) defaultGenerator.generate((SessionImpl)session, o);
}
#Override
public void disintegrate(SessionFactoryImplementor sessionFactoryImplementor, SessionFactoryServiceRegistry sessionFactoryServiceRegistry) {
}
}
You would need to register this class in the META-INF/services directory. Here is what the Hibernate documentation has to say about registering an Integrator:
For the integrator to be automatically used when Hibernate starts up, you will need to add a META-INF/services/org.hibernate.integrator.spi.Integrator file to your jar. The file should contain the fully qualified name of the class implementing the interface.
Because this class implements the ValueGenerator class, it can be used with the #GeneratorType annotation to automatically generate the sequential values. Here is how your class might be configured:
#Entity
#Table(name = "MyTable")
public class MyEntity {
//...
#Id //... etc
public Long getId() {
return id;
}
#GeneratorType(type = SequenceIntegrator.class, when = GenerationTime.INSERT)
#Column(name = "SEQ_VAL", unique = false, nullable = false, insertable = true, updatable = true)
public Long getMySequencedValue(){
return myVal;
}
}
I want to provide an alternative next to #Morten Berg's accepted solution, which worked better for me.
This approach allows to define the field with the actually desired Number type - Long in my use case - instead of GeneralSequenceNumber. This can be useful, e.g. for JSON (de-)serialization.
The downside is that it requires a little more database overhead.
First, we need an ActualEntity in which we want to auto-increment generated of type Long:
// ...
#Entity
public class ActualEntity {
#Id
// ...
Long id;
#Column(unique = true, updatable = false, nullable = false)
Long generated;
// ...
}
Next, we need a helper entity Generated. I placed it package-private next to ActualEntity, to keep it an implementation detail of the package:
#Entity
class Generated {
#Id
#GeneratedValue(strategy = SEQUENCE, generator = "seq")
#SequenceGenerator(name = "seq", initialValue = 1, allocationSize = 1)
Long id;
}
Finally, we need a place to hook in right before we save the ActualEntity. There, we create and persist aGenerated instance. This then provides a database-sequence generated id of type Long. We make use of this value by writing it to ActualEntity.generated.
In my use case, I implemented this using a Spring Data REST #RepositoryEventHandler, which get's called right before the ActualEntity get's persisted. It should demonstrate the principle:
#Component
#RepositoryEventHandler
public class ActualEntityHandler {
#Autowired
EntityManager entityManager;
#Transactional
#HandleBeforeCreate
public void generate(ActualEntity entity) {
Generated generated = new Generated();
entityManager.persist(generated);
entity.setGlobalId(generated.getId());
entityManager.remove(generated);
}
}
I didn't test it in a real-life application, so please enjoy with care.
"I don't want to use a trigger or any other thing other than Hibernate itself to generate the value for my property"
In that case, how about creating an implementation of UserType which generates the required value, and configuring the metadata to use that UserType for persistence of the mySequenceVal property?
This is not the same as using a sequence. When using a sequence, you are not inserting or updating anything. You are simply retrieving the next sequence value. It looks like hibernate does not support it.
If you have a column with UNIQUEIDENTIFIER type and default generation needed on insert but column is not PK
#Generated(GenerationTime.INSERT)
#Column(nullable = false , columnDefinition="UNIQUEIDENTIFIER")
private String uuidValue;
In db you will have
CREATE TABLE operation.Table1
(
Id INT IDENTITY (1,1) NOT NULL,
UuidValue UNIQUEIDENTIFIER DEFAULT NEWID() NOT NULL)
In this case you will not define generator for a value which you need (It will be automatically thanks to columnDefinition="UNIQUEIDENTIFIER"). The same you can try for other column types
I have found a workaround for this on MySql databases using #PostConstruct and JdbcTemplate in a Spring application. It may be doable with other databases but the use case that I will present is based on my experience with MySql, as it uses auto_increment.
First, I had tried defining a column as auto_increment using the ColumnDefinition property of the #Column annotation, but it was not working as the column needed to be an key in order to be auto incremental, but apparently the column wouldn't be defined as an index until after it was defined, causing a deadlock.
Here is where I came with the idea of creating the column without the auto_increment definition, and adding it after the database was created. This is possible using the #PostConstruct annotation, which causes a method to be invoked right after the application has initialized the beans, coupled with JdbcTemplate's update method.
The code is as follows:
In My Entity:
#Entity
#Table(name = "MyTable", indexes = { #Index(name = "my_index", columnList = "mySequencedValue") })
public class MyEntity {
//...
#Column(columnDefinition = "integer unsigned", nullable = false, updatable = false, insertable = false)
private Long mySequencedValue;
//...
}
In a PostConstructComponent class:
#Component
public class PostConstructComponent {
#Autowired
private JdbcTemplate jdbcTemplate;
#PostConstruct
public void makeMyEntityMySequencedValueAutoIncremental() {
jdbcTemplate.update("alter table MyTable modify mySequencedValue int unsigned auto_increment");
}
}
I was struggling with this today, was able to solve using this
#Generated(GenerationTime.INSERT)
#Column(name = "internal_id", columnDefinition = "serial", updatable = false)
private int internalId;
#Column(name = "<column name>", columnDefinition = "serial")
Works for mySQL
I've made a separate entity table for generating id and used it in to set this non-primay key id in the service that holds that id.
Entity:
import lombok.Data;
#Entity
#Data
public class GeneralSeqGenerator {
#Id
#GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "my_gen")
#SequenceGenerator(name = "my_gen", sequenceName= "my_seq", allocationSize = 1, initialValue = 100000)
private long seqNumber;
}
Repository:
public interface GeneralSeqGeneratorRepository extends JpaRepository<GeneralSeqGenerator, Long>{
}
Implementation of the service that holds non-primary id:
...
public void saveNewEntity(...) {
...
newEntity.setNonPrimaryId(generalSeqGeneratorRepository.save(new GeneralSeqGenerator()).getSeqNumber());
...
}
...
I've been in a situation like you (JPA/Hibernate sequence for non #Id field) and I ended up creating a trigger in my db schema that add a unique sequence number on insert. I just never got it to work with JPA/Hibernate
After spending hours, this neatly helped me to solve my problem:
For Oracle 12c:
ID NUMBER GENERATED as IDENTITY
For H2:
ID BIGINT GENERATED as auto_increment
Also make:
#Column(insertable = false)
I am creating a spring boot application and use it's build in JpaRepository interface to store my entities. I have the following two entities (removed getters and setters for readability):
Profile entity
#Entity
public class Profile {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private long id;
#OneToMany(mappedBy = "profileOne", orphanRemoval = true)
private List<Match> matchOnes;
#OneToMany(mappedBy = "profileTwo", orphanRemoval = true)
private List<Match> matchTwos;
}
Match entity
#Entity
#Table(uniqueConstraints={
#UniqueConstraint(columnNames = { "profileOne_id", "profileTwo_id" })
})
public class Match {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private long id;
#ManyToOne
#JoinColumn(name = "profileOne_id")
private Profile profileOne;
#ManyToOne
#JoinColumn(name = "profileTwo_id")
private Profile profileTwo;
}
To understand the JpaRepository behavior I wrote the following unit test.
#RunWith(SpringRunner.class)
#DataJpaTest
public class IProfileDaoTest {
#Autowired
private IProfileDao profileDao; //This class extends JpaRepository<Profile, long>
#Autowired
private IMatchDao matchDao; //This class extends JpaRepository<Match, long>
#Test
public void saveProfileTest() throws Exception {
#Test
public void profileMatchRelationTest() throws Exception {
//Test if matches stored by the IMatchDao are retrievable from the IProfileDao
Profile profileOne = new Profile("Bob"),
profileTwo = new Profile("Alex");
profileDao.saveAndFlush(profileOne);
profileDao.saveAndFlush(profileTwo);
matchDao.saveAndFlush(new Match(profileOne, profileTwo));
profileOne = profileDao.getOne(profileOne.getId());
Assert.assertEquals("Match not retrievable by profile.", 1, profileOne.getMatchOnes().size());
}
}
Now I expected the matches to have appeared in the profile entity but they do not. I also tried adding CascadeType.ALL to the #ManyToOne annotation in the match entity and adding FetchType.EAGER to the #OneToMany annotation in the profile entity.
Is it possible to get the matches saved with the matchDao by requesting a profile in the profileDao? Or should I find the matches with a profile with a separate function?
Spring Data repositories don't write to the database immediately for performance (and probably other) reasons. In tests, if you need to test query methods you need to use the TestEntityManager provided by #DataJpaTest (it's just the entity manager that the repositories use anyway in the background, but with a few convenience methods for testing).
Update 1:
The matches aren't added to the profile. To make sure the relationship is bidirectional the matches should have the profiles but the profiles should also have the matches.
Try this:
#RunWith(SpringRunner.class)
#DataJpaTest
public class IProfileDaoTest {
#Autowired
private IProfileDao profileDao; //This class extends JpaRepository<Profile, long>
#Autowired
private IMatchDao matchDao; //This class extends JpaRepository<Match, long>
#Autowired
private TestEntityManager testEntityManager;
#Test
public void saveProfileTest() throws Exception {
#Test
public void profileMatchRelationTest() throws Exception {
//Test if matches stored by the IMatchDao are retrievable from the IProfileDao
Profile profileOne = new Profile("Bob"),
profileTwo = new Profile("Alex");
//Persist the profiles so they exist when they are added to the match
entityManager.persistAndFlush(profileOne);
entityManager.persistAndFlush(profileTwo);
//Create and persist the match with two profiles
Match yourMatch = entityManager.persistFlushFind(new Match(profileOne, profileTwo));
//Add the match to both profiles and persist them again.
profileOne.matchOnes.add(yourMatch);
entityManager.persistAndFlush(profileOne);
profileTwo.matchTwos.add(yourMatch);
entityManager.persistAndFlush(profileTwo);
profileOne = profileDao.getOne(profileOne.getId());
Assert.assertEquals("Match not retrievable by profile.", 1, profileOne.getMatchOnes().size());
}
}
Everything in your test happens in the same JPA session. Such a session guarantees that every entity is included only once. So when you execute
profileOne = profileDao.getOne(profileOne.getId());
you are getting the exact instance back you created 5 lines above.
Hibernate nor any other JPA implementation will change anything in the entity for loading.
If you want to actually reload an entity you'll have to either evict it first from the entity manager or use a fresh Session/EntityManager.
For more details see chapter 3 of the JPA specification.
I'm seeing some videos about API Rest with Spring Boot and so far I've done some basics and when I tried to increase the complexity I'm getting caught.
My idea is in the Post / class, create a new class with students getting the following json:
{
"nome": "Primeira Serie - A".
"alunos": [
"João",
"José",
"Maria"
]
}
And return:
{
"id_classe": 101
}
It happens that it saves the class, but it does not save the students and I have no idea how to show only the id of the class.
I have created the following classes in Java:
Model
Classe.java
package com.example.classe.model;
//Import's suppressed
#Entity
#Table(name = "classe")
public class Classe {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private Integer id;
private String nome;
#OneToMany(mappedBy = "classe")
private Set<Aluno> alunos = new HashSet<Aluno>();
//Get's e Set's suppressed
}
Aluno.java
package com.example.classe.model;
//Import's suppressed
#Entity
#Table(name = "aluno")
public class Aluno {
private static int tempID = 0;
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private Integer id;
private String nome;
#ManyToOne
#JoinColumn(name = "id_classe")
#JsonBackReference
private Classe classe;
public Aluno(String nome) {
tempID++;
this.id = tempID;
this.nome = nome;
}
public Aluno() {
}
//Get's e Set's suppressed
}
Repository
ClasseRepository.java
package com.example.classe.repository;
//Import's suppressed
#Repository
public interface ClasseRepository extends JpaRepository<Classe, Integer> {
public List<Classe> findAll();
}
Controller
ClasseController.java
package com.example.classe.controller;
//Import's suppressed
#RestController
#RequestMapping("/classe")
public class ClasseController {
#Autowired
private ClasseRepository classeRepo;
#RequestMapping(method = RequestMethod.GET, produces = MediaType.APPLICATION_JSON_UTF8_VALUE)
public ResponseEntity<Collection<Classe>> getClasse() {
return new ResponseEntity<>(classeRepo.findAll(), HttpStatus.OK);
}
#RequestMapping(method = RequestMethod.POST)
public ResponseEntity<?> salvarClasse(#RequestBody Classe classe) {
return new ResponseEntity<>(classeRepo.saveAndFlush(classe), HttpStatus.CREATED);
}
}
Am I doing everything wrong or did I not understand the concept? But I wanted to understand how to do it that way.
Thanks in advance.
Cesar Sturion
What you want to achieve is totally doable, but requires several changes.
I split my answer into 2 parts:
Save the students
There are several problems with saving:
On POST your incoming json deserialized into objects in which Classe has a reference to Anuli, but Anuli doesn't have a reference toClasse. To check it you can add a break point at the line: return new ResponseEntity<>(... , run in debug mode and check fields of Anuli in Classe. To fix it you can add #JsonManagedReference on aluni field in Classe. Related question
Hibernate can't save referenced objects by default. You have to save them one by one after saving your Classe object or just turn on Cascade persisting. Related question
So, to fix 1 and 2 Classe should have:
#OneToMany(mappedBy = "classe", cascade = CascadeType.PERSIST)
#JsonManagedReference
private Set<Aluno> alunos = new HashSet<Aluno>();
You have to remove custom id generation in Alumi (I am talking about static int tempID). Annotation #GeneratedValue will perfectly generate id for you as soon as you persist an object. This custom generation breaks Hibernate support. I even not talking about that it also breaks the app after restart, not threadsafe etc.
Return id only
On POST returned json represent what was returned in classeRepo.saveAndFlush(classe) so it's an object of Classe.
If you want to return exactly this:
{
"id_classe": 101
}
Then create new class like this:
public class ClasseIdVO {
#JsonProperty("id_casse")
private Integer id;
// Constructors, getter, setter
VO - means View Object, so this object only for representation, not for persisting, etc.
You can use field name id_casse, but it's against Java code convention, so better add #JsonProperty.
Also change your saving code to new ClasseIdVO(classeRepo.saveAndFlush(classe).getId())
Or you can just return id as a number: classeRepo.saveAndFlush(classe).getId()
In JPA (Hibernate), when we automatically generate the ID field, it is assumed that the user has no knowledge about this key. So, when obtaining the entity, user would query based on some field other than ID. How do we obtain the entity in that case (since em.find() cannot be used).
I understand we can use a query and filter the results later. But, is there a more direct way (because this is a very common problem as I understand).
It is not a "problem" as you stated it.
Hibernate has the built-in find(), but you have to build your own query in order to get a particular object. I recommend using Hibernate's Criteria :
Criteria criteria = session.createCriteria(YourClass.class);
YourObject yourObject = criteria.add(Restrictions.eq("yourField", yourFieldValue))
.uniqueResult();
This will create a criteria on your current class, adding the restriction that the column "yourField" is equal to the value yourFieldValue. uniqueResult() tells it to bring a unique result. If more objects match, you should retrive a list.
List<YourObject> list = criteria.add(Restrictions.eq("yourField", yourFieldValue)).list();
If you have any further questions, please feel free to ask. Hope this helps.
if you have repository for entity Foo and need to select all entries with exact string value boo (also works for other primitive types or entity types). Put this into your repository interface:
List<Foo> findByBoo(String boo);
if you need to order results:
List<Foo> findByBooOrderById(String boo);
See more at reference.
Basically, you should add a specific unique field. I usually use xxxUri fields.
class User {
#Id
// automatically generated
private Long id;
// globally unique id
#Column(name = "SCN", nullable = false, unique = true)
private String scn;
}
And you business method will do like this.
public User findUserByScn(#NotNull final String scn) {
CriteriaBuilder builder = manager.getCriteriaBuilder();
CriteriaQuery<User> criteria = builder.createQuery(User.class);
Root<User> from = criteria.from(User.class);
criteria.select(from);
criteria.where(builder.equal(from.get(User_.scn), scn));
TypedQuery<User> typed = manager.createQuery(criteria);
try {
return typed.getSingleResult();
} catch (final NoResultException nre) {
return null;
}
}
Best practice is using #NaturalId annotation. It can be used as the business key for some cases it is too complicated, so some fields are using as the identifier in the real world.
For example, I have user class with user id as primary key, and email is also unique field. So we can use email as our natural id
#Entity
#Table(name="user")
public class User {
#Id
#Column(name="id")
private int id;
#NaturalId
#Column(name="email")
private String email;
#Column(name="name")
private String name;
}
To get our record, just simply use 'session.byNaturalId()'
Session session = sessionFactory.getCurrentSession();
User user = session.byNaturalId(User.class)
.using("email","huchenhai#qq.com")
.load()
This solution is from Beginning Hibernate book:
Query<User> query = session.createQuery("from User u where u.scn=:scn", User.class);
query.setParameter("scn", scn);
User user = query.uniqueResult();
I solved a similar problem, where I wanted to find a book by its isbnCode not by your id(primary key).
#Entity
public class Book implements Serializable {
private static final long serialVersionUID = 1L;
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private Integer id;
private String isbnCode;
...
In the repository the method was created like #kamalveer singh mentioned. Note that the method name is findBy+fieldName (in my case: findByisbnCode):
#Repository
public interface BookRepository extends JpaRepository<Book, Integer> {
Book findByisbnCode(String isbnCode);
}
Then, implemented the method in the service:
#Service
public class BookService {
#Autowired
private BookRepository repo;
public Book findByIsbnCode(String isbnCode) {
Book obj = repo.findByisbnCode(isbnCode);
return obj;
}
}
Write a custom method like this:
public Object findByYourField(Class entityClass, String yourFieldValue)
{
CriteriaBuilder criteriaBuilder = entityManager.getCriteriaBuilder();
CriteriaQuery<Object> criteriaQuery = criteriaBuilder.createQuery(entityClass);
Root<Object> root = criteriaQuery.from(entityClass);
criteriaQuery.select(root);
ParameterExpression<String> params = criteriaBuilder.parameter(String.class);
criteriaQuery.where(criteriaBuilder.equal(root.get("yourField"), params));
TypedQuery<Object> query = entityManager.createQuery(criteriaQuery);
query.setParameter(params, yourFieldValue);
List<Object> queryResult = query.getResultList();
Object returnObject = null;
if (CollectionUtils.isNotEmpty(queryResult)) {
returnObject = queryResult.get(0);
}
return returnObject;
}
Edit: Just realized that #Chinmoy was getting at basically the same thing, but I think I may have done a better job ELI5 :)
If you're using a flavor of Spring Data to help persist / fetch things from whatever kind of Repository you've defined, you can probably have your JPA provider do this for you via some clever tricks with method names in your Repository interface class. Allow me to explain.
(As a disclaimer, I just a few moments ago did/still am figuring this out for myself.)
For example, if I am storing Tokens in my database, I might have an entity class that looks like this:
#Data // << Project Lombok convenience annotation
#Entity
public class Token {
#Id
#Column(name = "TOKEN_ID")
private String tokenId;
#Column(name = "TOKEN")
private String token;
#Column(name = "EXPIRATION")
private String expiration;
#Column(name = "SCOPE")
private String scope;
}
And I probably have a CrudRepository<K,V> interface defined like this, to give me simple CRUD operations on that Repository for free.
#Repository
// CrudRepository<{Entity Type}, {Entity Primary Key Type}>
public interface TokenRepository extends CrudRepository<Token, String> { }
And when I'm looking up one of these tokens, my purpose might be checking the expiration or scope, for example. In either of those cases, I probably don't have the tokenId handy, but rather just the value of a token field itself that I want to look up.
To do that, you can add an additional method to your TokenRepository interface in a clever way to tell your JPA provider that the value you're passing in to the method is not the tokenId, but the value of another field within the Entity class, and it should take that into account when it is generating the actual SQL that it will run against your database.
#Repository
// CrudRepository<{Entity Type}, {Entity Primary Key Type}>
public interface TokenRepository extends CrudRepository<Token, String> {
List<Token> findByToken(String token);
}
I read about this on the Spring Data R2DBC docs page, and it seems to be working so far within a SpringBoot 2.x app storing in an embedded H2 database.
No, you don't need to make criteria query it would be boilerplate code you just do simple thing if you working in Spring-boot:
in your repo declare a method name with findBy[exact field name].
Example-
if your model or document consist a string field myField and you want to find by it then your method name will be:
findBymyField(String myField);
All the answers require you to write some sort of SQL/HQL/whatever. Why? You don't have to - just use CriteriaBuilder:
Person.java:
#Entity
class Person {
#Id #GeneratedValue
private int id;
#Column(name = "name")
private String name;
#Column(name = "age")
private int age;
...
}
Dao.java:
public class Dao {
public static Person getPersonByName(String name) {
SessionFactory sessionFactory = new Configuration().configure().buildSessionFactory();
Session session = sessionFactory.openSession();
session.beginTransaction();
CriteriaBuilder cb = session.getCriteriaBuilder();
CriteriaQuery<Person> cr = cb.createQuery(Person.class);
Root<Person> root = cr.from(Person.class);
cr.select(root).where(cb.equal(root.get("name"), name)); //here you pass a class field, not a table column (in this example they are called the same)
Query query = session.createQuery(cr);
query.setMaxResults(1);
List<Person> resultList = query.getResultList();
Person result = resultList.get(0);
return result;
}
}
example of use:
public static void main(String[] args) {
Person person = Dao.getPersonByName("John");
System.out.println(person.getAge()); //John's age
}
Have a look at:
JPA query language: The Java Persistence Query Language
JPA Criteria API: Using the Criteria API to Create Queries
I've written a library that helps do precisely this. It allows search by object simply by initializing only the fields you want to filter by: https://github.com/kg6zvp/GenericEntityEJB
Refer - Spring docs for query methods
We can add methods in Spring Jpa by passing diff params in methods like:
List<Person> findByEmailAddressAndLastname(EmailAddress emailAddress, String lastname);
// Enabling static ORDER BY for a query
List<Person> findByLastnameOrderByFirstnameAsc(String lastname);
In my Spring Boot app I resolved a similar type of issue like this:
#Autowired
private EntityManager entityManager;
public User findByEmail(String email) {
User user = null;
Query query = entityManager.createQuery("SELECT u FROM User u WHERE u.email=:email");
query.setParameter("email", email);
try {
user = (User) query.getSingleResult();
} catch (Exception e) {
// Handle exception
}
return user;
}
This is very basic query :
Entity : Student
#Entity
#Data
#NoArgsConstructor
public class Student{
#Id
#GeneratedValue(generator = "uuid2", strategy = GenerationType.IDENTITY)
#GenericGenerator(name = "uuid2", strategy = "uuid2")
private String id;
#Column(nullable = false)
#Version
#JsonIgnore
private Integer version;
private String studentId;
private String studentName;
private OffsetDateTime enrollDate;
}
Repository Interface : StudentRepository
#Repository
public interface StudentRepository extends JpaRepository<Student, String> {
List<Student> findByStudentName(String studentName);
List<Student> findByStudentNameOrderByEnrollDateDesc(String studentName);
#Transactional
#Modifying
void deleteByStudentName(String studentName);
}
Note:
findByColumnName : give results by criteria
List findByStudentName(String studentName)
Internally convert into query : select * from Student where name='studentName'
#Transactional
#Modifying
Is useful when you want to remove persisted data from database.
Using CrudRepository and JPA query works for me:
import org.springframework.data.jpa.repository.Query;
import org.springframework.data.repository.CrudRepository;
import org.springframework.data.repository.query.Param;
public interface TokenCrudRepository extends CrudRepository<Token, Integer> {
/**
* Finds a token by using the user as a search criteria.
* #param user
* #return A token element matching with the given user.
*/
#Query("SELECT t FROM Token t WHERE LOWER(t.user) = LOWER(:user)")
public Token find(#Param("user") String user);
}
and you invoke the find custom method like this:
public void destroyCurrentToken(String user){
AbstractApplicationContext context = getContext();
repository = context.getBean(TokenCrudRepository.class);
Token token = ((TokenCrudRepository) repository).find(user);
int idToken = token.getId();
repository.delete(idToken);
context.close();
}
Is it possible to use a DB sequence for some column that is not the identifier/is not part of a composite identifier?
I'm using hibernate as jpa provider, and I have a table that has some columns that are generated values (using a sequence), although they are not part of the identifier.
What I want is to use a sequence to create a new value for an entity, where the column for the sequence is NOT (part of) the primary key:
#Entity
#Table(name = "MyTable")
public class MyEntity {
//...
#Id //... etc
public Long getId() {
return id;
}
//note NO #Id here! but this doesn't work...
#GeneratedValue(strategy = GenerationType.AUTO, generator = "myGen")
#SequenceGenerator(name = "myGen", sequenceName = "MY_SEQUENCE")
#Column(name = "SEQ_VAL", unique = false, nullable = false, insertable = true, updatable = true)
public Long getMySequencedValue(){
return myVal;
}
}
Then when I do this:
em.persist(new MyEntity());
the id will be generated, but the mySequenceVal property will be also generated by my JPA provider.
Just to make things clear: I want Hibernate to generate the value for the mySequencedValue property. I know Hibernate can handle database-generated values, but I don't want to use a trigger or any other thing other than Hibernate itself to generate the value for my property. If Hibernate can generate values for primary keys, why can't it generate for a simple property?
Looking for answers to this problem, I stumbled upon this link
It seems that Hibernate/JPA isn't able to automatically create a value for your non-id-properties. The #GeneratedValue annotation is only used in conjunction with #Id to create auto-numbers.
The #GeneratedValue annotation just tells Hibernate that the database is generating this value itself.
The solution (or work-around) suggested in that forum is to create a separate entity with a generated Id, something like this:
#Entity
public class GeneralSequenceNumber {
#Id
#GeneratedValue(...)
private Long number;
}
#Entity
public class MyEntity {
#Id ..
private Long id;
#OneToOne(...)
private GeneralSequnceNumber myVal;
}
I found that #Column(columnDefinition="serial") works perfect but only for PostgreSQL. For me this was perfect solution, because second entity is "ugly" option.
A call to saveAndFlush on the entity is also necessary, and save won't be enough to populate the value from the DB.
I know this is a very old question, but it's showed firstly upon the results and jpa has changed a lot since the question.
The right way to do it now is with the #Generated annotation. You can define the sequence, set the default in the column to that sequence and then map the column as:
#Generated(GenerationTime.INSERT)
#Column(name = "column_name", insertable = false)
Hibernate definitely supports this. From the docs:
"Generated properties are properties which have their values generated by the database. Typically, Hibernate applications needed to refresh objects which contain any properties for which the database was generating values. Marking properties as generated, however, lets the application delegate this responsibility to Hibernate. Essentially, whenever Hibernate issues an SQL INSERT or UPDATE for an entity which has defined generated properties, it immediately issues a select afterwards to retrieve the generated values."
For properties generated on insert only, your property mapping (.hbm.xml) would look like:
<property name="foo" generated="insert"/>
For properties generated on insert and update your property mapping (.hbm.xml) would look like:
<property name="foo" generated="always"/>
Unfortunately, I don't know JPA, so I don't know if this feature is exposed via JPA (I suspect possibly not)
Alternatively, you should be able to exclude the property from inserts and updates, and then "manually" call session.refresh( obj ); after you have inserted/updated it to load the generated value from the database.
This is how you would exclude the property from being used in insert and update statements:
<property name="foo" update="false" insert="false"/>
Again, I don't know if JPA exposes these Hibernate features, but Hibernate does support them.
I fixed the generation of UUID (or sequences) with Hibernate using #PrePersist annotation:
#PrePersist
public void initializeUUID() {
if (uuid == null) {
uuid = UUID.randomUUID().toString();
}
}
Looks like thread is old, I just wanted to add my solution here(Using AspectJ - AOP in spring).
Solution is to create a custom annotation #InjectSequenceValue as follows.
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface InjectSequenceValue {
String sequencename();
}
Now you can annotate any field in entity, so that the underlying field (Long/Integer) value will be injected at runtime using the nextvalue of the sequence.
Annotate like this.
//serialNumber will be injected dynamically, with the next value of the serialnum_sequence.
#InjectSequenceValue(sequencename = "serialnum_sequence")
Long serialNumber;
So far we have marked the field we need to inject the sequence value.So we will look how to inject the sequence value to the marked fields, this is done by creating the point cut in AspectJ.
We will trigger the injection just before the save/persist method is being executed.This is done in the below class.
#Aspect
#Configuration
public class AspectDefinition {
#Autowired
JdbcTemplate jdbcTemplate;
//#Before("execution(* org.hibernate.session.save(..))") Use this for Hibernate.(also include session.save())
#Before("execution(* org.springframework.data.repository.CrudRepository.save(..))") //This is for JPA.
public void generateSequence(JoinPoint joinPoint){
Object [] aragumentList=joinPoint.getArgs(); //Getting all arguments of the save
for (Object arg :aragumentList ) {
if (arg.getClass().isAnnotationPresent(Entity.class)){ // getting the Entity class
Field[] fields = arg.getClass().getDeclaredFields();
for (Field field : fields) {
if (field.isAnnotationPresent(InjectSequenceValue.class)) { //getting annotated fields
field.setAccessible(true);
try {
if (field.get(arg) == null){ // Setting the next value
String sequenceName=field.getAnnotation(InjectSequenceValue.class).sequencename();
long nextval=getNextValue(sequenceName);
System.out.println("Next value :"+nextval); //TODO remove sout.
field.set(arg, nextval);
}
} catch (Exception e) {
e.printStackTrace();
}
}
}
}
}
}
/**
* This method fetches the next value from sequence
* #param sequence
* #return
*/
public long getNextValue(String sequence){
long sequenceNextVal=0L;
SqlRowSet sqlRowSet= jdbcTemplate.queryForRowSet("SELECT "+sequence+".NEXTVAL as value FROM DUAL");
while (sqlRowSet.next()){
sequenceNextVal=sqlRowSet.getLong("value");
}
return sequenceNextVal;
}
}
Now you can annotate any Entity as below.
#Entity
#Table(name = "T_USER")
public class UserEntity {
#Id
#SequenceGenerator(sequenceName = "userid_sequence",name = "this_seq")
#GeneratedValue(strategy = GenerationType.SEQUENCE,generator = "this_seq")
Long id;
String userName;
String password;
#InjectSequenceValue(sequencename = "serialnum_sequence") // this will be injected at the time of saving.
Long serialNumber;
String name;
}
As a followup here's how I got it to work:
#Override public Long getNextExternalId() {
BigDecimal seq =
(BigDecimal)((List)em.createNativeQuery("select col_msd_external_id_seq.nextval from dual").getResultList()).get(0);
return seq.longValue();
}
Although this is an old thread I want to share my solution and hopefully get some feedback on this. Be warned that I only tested this solution with my local database in some JUnit testcase. So this is not a productive feature so far.
I solved that issue for my by introducing a custom annotation called Sequence with no property. It's just a marker for fields that should be assigned a value from an incremented sequence.
#Retention(RetentionPolicy.RUNTIME)
#Target(ElementType.FIELD)
public #interface Sequence
{
}
Using this annotation i marked my entities.
public class Area extends BaseEntity implements ClientAware, IssuerAware
{
#Column(name = "areaNumber", updatable = false)
#Sequence
private Integer areaNumber;
....
}
To keep things database independent I introduced an entity called SequenceNumber which holds the sequence current value and the increment size. I chose the className as unique key so each entity class wil get its own sequence.
#Entity
#Table(name = "SequenceNumber", uniqueConstraints = { #UniqueConstraint(columnNames = { "className" }) })
public class SequenceNumber
{
#Id
#Column(name = "className", updatable = false)
private String className;
#Column(name = "nextValue")
private Integer nextValue = 1;
#Column(name = "incrementValue")
private Integer incrementValue = 10;
... some getters and setters ....
}
The last step and the most difficult is a PreInsertListener that handles the sequence number assignment. Note that I used spring as bean container.
#Component
public class SequenceListener implements PreInsertEventListener
{
private static final long serialVersionUID = 7946581162328559098L;
private final static Logger log = Logger.getLogger(SequenceListener.class);
#Autowired
private SessionFactoryImplementor sessionFactoryImpl;
private final Map<String, CacheEntry> cache = new HashMap<>();
#PostConstruct
public void selfRegister()
{
// As you might expect, an EventListenerRegistry is the place with which event listeners are registered
// It is a service so we look it up using the service registry
final EventListenerRegistry eventListenerRegistry = sessionFactoryImpl.getServiceRegistry().getService(EventListenerRegistry.class);
// add the listener to the end of the listener chain
eventListenerRegistry.appendListeners(EventType.PRE_INSERT, this);
}
#Override
public boolean onPreInsert(PreInsertEvent p_event)
{
updateSequenceValue(p_event.getEntity(), p_event.getState(), p_event.getPersister().getPropertyNames());
return false;
}
private void updateSequenceValue(Object p_entity, Object[] p_state, String[] p_propertyNames)
{
try
{
List<Field> fields = ReflectUtil.getFields(p_entity.getClass(), null, Sequence.class);
if (!fields.isEmpty())
{
if (log.isDebugEnabled())
{
log.debug("Intercepted custom sequence entity.");
}
for (Field field : fields)
{
Integer value = getSequenceNumber(p_entity.getClass().getName());
field.setAccessible(true);
field.set(p_entity, value);
setPropertyState(p_state, p_propertyNames, field.getName(), value);
if (log.isDebugEnabled())
{
LogMF.debug(log, "Set {0} property to {1}.", new Object[] { field, value });
}
}
}
}
catch (Exception e)
{
log.error("Failed to set sequence property.", e);
}
}
private Integer getSequenceNumber(String p_className)
{
synchronized (cache)
{
CacheEntry current = cache.get(p_className);
// not in cache yet => load from database
if ((current == null) || current.isEmpty())
{
boolean insert = false;
StatelessSession session = sessionFactoryImpl.openStatelessSession();
session.beginTransaction();
SequenceNumber sequenceNumber = (SequenceNumber) session.get(SequenceNumber.class, p_className);
// not in database yet => create new sequence
if (sequenceNumber == null)
{
sequenceNumber = new SequenceNumber();
sequenceNumber.setClassName(p_className);
insert = true;
}
current = new CacheEntry(sequenceNumber.getNextValue() + sequenceNumber.getIncrementValue(), sequenceNumber.getNextValue());
cache.put(p_className, current);
sequenceNumber.setNextValue(sequenceNumber.getNextValue() + sequenceNumber.getIncrementValue());
if (insert)
{
session.insert(sequenceNumber);
}
else
{
session.update(sequenceNumber);
}
session.getTransaction().commit();
session.close();
}
return current.next();
}
}
private void setPropertyState(Object[] propertyStates, String[] propertyNames, String propertyName, Object propertyState)
{
for (int i = 0; i < propertyNames.length; i++)
{
if (propertyName.equals(propertyNames[i]))
{
propertyStates[i] = propertyState;
return;
}
}
}
private static class CacheEntry
{
private int current;
private final int limit;
public CacheEntry(final int p_limit, final int p_current)
{
current = p_current;
limit = p_limit;
}
public Integer next()
{
return current++;
}
public boolean isEmpty()
{
return current >= limit;
}
}
}
As you can see from the above code the listener used one SequenceNumber instance per entity class and reserves a couple of sequence numbers defined by the incrementValue of the SequenceNumber entity. If it runs out of sequence numbers it loads the SequenceNumber entity for the target class and reserves incrementValue values for the next calls. This way I do not need to query the database each time a sequence value is needed.
Note the StatelessSession that is being opened for reserving the next set of sequence numbers. You cannot use the same session the target entity is currently persisted since this would lead to a ConcurrentModificationException in the EntityPersister.
Hope this helps someone.
If you are using postgresql
And i'm using in spring boot 1.5.6
#Column(columnDefinition = "serial")
#Generated(GenerationTime.INSERT)
private Integer orderID;
I run in the same situation like you and I also didn't find any serious answers if it is basically possible to generate non-id propertys with JPA or not.
My solution is to call the sequence with a native JPA query to set the property by hand before persisiting it.
This is not satisfying but it works as a workaround for the moment.
Mario
I've found this specific note in session 9.1.9 GeneratedValue Annotation from JPA specification:
"[43] Portable applications should not use the GeneratedValue annotation on other persistent fields or properties."
So, I presume that it is not possible to auto generate value for non primary key values at least using simply JPA.
I want to provide an alternative next to #Morten Berg's accepted solution, which worked better for me.
This approach allows to define the field with the actually desired Number type - Long in my use case - instead of GeneralSequenceNumber. This can be useful, e.g. for JSON (de-)serialization.
The downside is that it requires a little more database overhead.
First, we need an ActualEntity in which we want to auto-increment generated of type Long:
// ...
#Entity
public class ActualEntity {
#Id
// ...
Long id;
#Column(unique = true, updatable = false, nullable = false)
Long generated;
// ...
}
Next, we need a helper entity Generated. I placed it package-private next to ActualEntity, to keep it an implementation detail of the package:
#Entity
class Generated {
#Id
#GeneratedValue(strategy = SEQUENCE, generator = "seq")
#SequenceGenerator(name = "seq", initialValue = 1, allocationSize = 1)
Long id;
}
Finally, we need a place to hook in right before we save the ActualEntity. There, we create and persist aGenerated instance. This then provides a database-sequence generated id of type Long. We make use of this value by writing it to ActualEntity.generated.
In my use case, I implemented this using a Spring Data REST #RepositoryEventHandler, which get's called right before the ActualEntity get's persisted. It should demonstrate the principle:
#Component
#RepositoryEventHandler
public class ActualEntityHandler {
#Autowired
EntityManager entityManager;
#Transactional
#HandleBeforeCreate
public void generate(ActualEntity entity) {
Generated generated = new Generated();
entityManager.persist(generated);
entity.setGlobalId(generated.getId());
entityManager.remove(generated);
}
}
I didn't test it in a real-life application, so please enjoy with care.
You can do exactly what you are asking.
I've found it is possible to adapt Hibernate's IdentifierGenerator implementations by registering them with an Integrator. With this you should be able to use any id sequence generator provided by Hibernate to generate sequences for non-id fields (presumably the non-sequential id generators would work as well).
There are quite a few options for generating ids this way. Check out some of the implementations of IdentifierGenerator, specifically SequenceStyleGenerator and TableGenerator. If you have configured generators using the #GenericGenerator annotation, then the parameters for these classes may be familiar to you. This would also have the advantage of using Hibernate to generate the SQL.
Here is how I got it working:
import org.hibernate.Session;
import org.hibernate.boot.Metadata;
import org.hibernate.engine.spi.SessionFactoryImplementor;
import org.hibernate.id.IdentifierGenerator;
import org.hibernate.id.enhanced.TableGenerator;
import org.hibernate.integrator.spi.Integrator;
import org.hibernate.internal.SessionImpl;
import org.hibernate.service.spi.SessionFactoryServiceRegistry;
import org.hibernate.tuple.ValueGenerator;
import org.hibernate.type.LongType;
import java.util.Properties;
public class SequenceIntegrator implements Integrator, ValueGenerator<Long> {
public static final String TABLE_NAME = "SEQUENCE_TABLE";
public static final String VALUE_COLUMN_NAME = "NEXT_VAL";
public static final String SEGMENT_COLUMN_NAME = "SEQUENCE_NAME";
private static SessionFactoryServiceRegistry serviceRegistry;
private static Metadata metadata;
private static IdentifierGenerator defaultGenerator;
#Override
public void integrate(Metadata metadata, SessionFactoryImplementor sessionFactoryImplementor, SessionFactoryServiceRegistry sessionFactoryServiceRegistry) {
//assigning metadata and registry to fields for use in a below example
SequenceIntegrator.metadata = metadata;
SequenceIntegrator.serviceRegistry = sessionFactoryServiceRegistry;
SequenceIntegrator.defaultGenerator = getTableGenerator(metadata, sessionFactoryServiceRegistry, "DEFAULT");
}
private TableGenerator getTableGenerator(Metadata metadata, SessionFactoryServiceRegistry sessionFactoryServiceRegistry, String segmentValue) {
TableGenerator generator = new TableGenerator();
Properties properties = new Properties();
properties.setProperty("table_name", TABLE_NAME);
properties.setProperty("value_column_name", VALUE_COLUMN_NAME);
properties.setProperty("segment_column_name", SEGMENT_COLUMN_NAME);
properties.setProperty("segment_value", segmentValue);
//any type should work if the generator supports it
generator.configure(LongType.INSTANCE, properties, sessionFactoryServiceRegistry);
//this should create the table if ddl auto update is enabled and if this function is called inside of the integrate method
generator.registerExportables(metadata.getDatabase());
return generator;
}
#Override
public Long generateValue(Session session, Object o) {
// registering additional generators with getTableGenerator will work here. inserting new sequences can be done dynamically
// example:
// TableGenerator classSpecificGenerator = getTableGenerator(metadata, serviceRegistry, o.getClass().getName());
// return (Long) classSpecificGenerator.generate((SessionImpl)session, o);
return (Long) defaultGenerator.generate((SessionImpl)session, o);
}
#Override
public void disintegrate(SessionFactoryImplementor sessionFactoryImplementor, SessionFactoryServiceRegistry sessionFactoryServiceRegistry) {
}
}
You would need to register this class in the META-INF/services directory. Here is what the Hibernate documentation has to say about registering an Integrator:
For the integrator to be automatically used when Hibernate starts up, you will need to add a META-INF/services/org.hibernate.integrator.spi.Integrator file to your jar. The file should contain the fully qualified name of the class implementing the interface.
Because this class implements the ValueGenerator class, it can be used with the #GeneratorType annotation to automatically generate the sequential values. Here is how your class might be configured:
#Entity
#Table(name = "MyTable")
public class MyEntity {
//...
#Id //... etc
public Long getId() {
return id;
}
#GeneratorType(type = SequenceIntegrator.class, when = GenerationTime.INSERT)
#Column(name = "SEQ_VAL", unique = false, nullable = false, insertable = true, updatable = true)
public Long getMySequencedValue(){
return myVal;
}
}
"I don't want to use a trigger or any other thing other than Hibernate itself to generate the value for my property"
In that case, how about creating an implementation of UserType which generates the required value, and configuring the metadata to use that UserType for persistence of the mySequenceVal property?
This is not the same as using a sequence. When using a sequence, you are not inserting or updating anything. You are simply retrieving the next sequence value. It looks like hibernate does not support it.
If you have a column with UNIQUEIDENTIFIER type and default generation needed on insert but column is not PK
#Generated(GenerationTime.INSERT)
#Column(nullable = false , columnDefinition="UNIQUEIDENTIFIER")
private String uuidValue;
In db you will have
CREATE TABLE operation.Table1
(
Id INT IDENTITY (1,1) NOT NULL,
UuidValue UNIQUEIDENTIFIER DEFAULT NEWID() NOT NULL)
In this case you will not define generator for a value which you need (It will be automatically thanks to columnDefinition="UNIQUEIDENTIFIER"). The same you can try for other column types
I have found a workaround for this on MySql databases using #PostConstruct and JdbcTemplate in a Spring application. It may be doable with other databases but the use case that I will present is based on my experience with MySql, as it uses auto_increment.
First, I had tried defining a column as auto_increment using the ColumnDefinition property of the #Column annotation, but it was not working as the column needed to be an key in order to be auto incremental, but apparently the column wouldn't be defined as an index until after it was defined, causing a deadlock.
Here is where I came with the idea of creating the column without the auto_increment definition, and adding it after the database was created. This is possible using the #PostConstruct annotation, which causes a method to be invoked right after the application has initialized the beans, coupled with JdbcTemplate's update method.
The code is as follows:
In My Entity:
#Entity
#Table(name = "MyTable", indexes = { #Index(name = "my_index", columnList = "mySequencedValue") })
public class MyEntity {
//...
#Column(columnDefinition = "integer unsigned", nullable = false, updatable = false, insertable = false)
private Long mySequencedValue;
//...
}
In a PostConstructComponent class:
#Component
public class PostConstructComponent {
#Autowired
private JdbcTemplate jdbcTemplate;
#PostConstruct
public void makeMyEntityMySequencedValueAutoIncremental() {
jdbcTemplate.update("alter table MyTable modify mySequencedValue int unsigned auto_increment");
}
}
I was struggling with this today, was able to solve using this
#Generated(GenerationTime.INSERT)
#Column(name = "internal_id", columnDefinition = "serial", updatable = false)
private int internalId;
#Column(name = "<column name>", columnDefinition = "serial")
Works for mySQL
I've made a separate entity table for generating id and used it in to set this non-primay key id in the service that holds that id.
Entity:
import lombok.Data;
#Entity
#Data
public class GeneralSeqGenerator {
#Id
#GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "my_gen")
#SequenceGenerator(name = "my_gen", sequenceName= "my_seq", allocationSize = 1, initialValue = 100000)
private long seqNumber;
}
Repository:
public interface GeneralSeqGeneratorRepository extends JpaRepository<GeneralSeqGenerator, Long>{
}
Implementation of the service that holds non-primary id:
...
public void saveNewEntity(...) {
...
newEntity.setNonPrimaryId(generalSeqGeneratorRepository.save(new GeneralSeqGenerator()).getSeqNumber());
...
}
...
I've been in a situation like you (JPA/Hibernate sequence for non #Id field) and I ended up creating a trigger in my db schema that add a unique sequence number on insert. I just never got it to work with JPA/Hibernate
After spending hours, this neatly helped me to solve my problem:
For Oracle 12c:
ID NUMBER GENERATED as IDENTITY
For H2:
ID BIGINT GENERATED as auto_increment
Also make:
#Column(insertable = false)