Develop Reference

Different bidirectional #ManyToOne parent for 2 Entities extending some base Entity, but keeping base entity data in one table

This title is closest I could get to what I am trying to do. :)
Let's start with how it should look in database (tables with columns):
Email
- id : PK
- email : String
CompanyEmail
- email_id : FK
- company_id : FK
PersonEmail
- email_id : FK
- person_id : FK
Company
- id : PK
Person
- id : PK
Now let's look at model:
#Entity
public class Company
{
#Id
#GeneratedValue
private Long id;
#OneToMany
private List<CompanyEmail> emails = new ArrayList<>();
}
#Entity
public class Person
{
#Id
#GeneratedValue
private Long id;
#OneToMany
private List<PersonEmail> emails = new ArrayList<>();
}
// Should this be #Entity? Maybe #MappedSuperclass? What strategy to use to be able to extend it?
public class Email // This is base email class/entity
{
#Id
#GeneratedValue
private Long id;
#javax.validation.constraints.Email
private String email;
}
#Entity
public class PersonEmail // This needs to somehow extend Email, so all Email data is also accessible here.
{
#ManyToOne
private Person person;
}
#Entity
public class ComanyEmail // This needs to somehow extend Email, so all Email data is also accessible here.
{
#ManyToOne
private Company company;
}
Now my question is - is it possible in Hibernate (latest) to achieve such structure?
Key points when designing above (what drove me):
Keep ALL emails in one table, for sanity checks (uniqueness).
Have smallest database footprint - above gives just one table for Email and then 2 join tables with 2 FK.
Still be able to keep the model design Hibernate-friendly (basically don't use any special queries, just JPA annotations). This means that both Company and Person can easily LAZY load their specific emails (subclasses), and also those subclassed emails can correspond to them (PersonEmail to Person, and CompanyEmail to Company) - making model bidirectional.
Note: I've also considered creating something like Contactable base class for Comapny and Person which would have Email list, but that doesn't suit my needs (neither class per table, nor same table designs).
My question is - is it possible? Even if I don't get answer with example based on classes I gave, just the fact will give me hope and I will find it.
Edit 1
Should I maybe use discriminator in Email table and then keep FK there?
Email
- id : PK
- email : String
- companyOrPerson : FK
- discriminator : Decides what FK points at
Company
- id : PK
Person
- id : PK
Here I am grabbing straws - I have no idea if such thing is possible, can discriminator decide about parent table (companyOrPerson)? Does bidirectional work here (like I mentioned, maybe I should make some base class Contactable)?
I am open to suggestions on how to do it well.

Favour composition over inheritence. It tends to simplify things.
If you had main entities of
Person
Company
Email
Then a couple of composites
PersonEmail (contains a Person and an Email)
CompanyEmail (contains a Company and an Email)
Would that not fit what you need?
e.g.
#Entity
public class Person {
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private long id;
#OneToMany
#JoinColumn(name="person_id")
private List<PersonEmail> emails = new ArrayList<>();
}
#Entity
public class Company {
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private long id;
#OneToMany()
#JoinColumn(name="company_id")
private List<CompanyEmail> emails = new ArrayList<>();
}
#Entity
public class Email {
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private long id;
#javax.validation.constraints.Email
private String email;
}
#Entity
public class PersonEmail {
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private long id;
#OneToOne(cascade=CascadeType.ALL)
#JoinColumn(name="person_id", referencedColumnName="id")
private Person person;
#OneToOne(cascade=CascadeType.ALL)
#JoinColumn(name="email_id", referencedColumnName="id")
private Email email;
}
#Entity
public class CompanyEmail
{
#Id
#GeneratedValue(strategy=GenerationType.IDENTITY)
private long id;
#OneToOne(cascade=CascadeType.ALL)
#JoinColumn(name="company_id", referencedColumnName="id")
private Company company;
#OneToOne(cascade=CascadeType.ALL)
#JoinColumn(name="email_id", referencedColumnName="id")
private Email email;
}

Hibernate Mapping Only one class

Hibernate Mapping
How to implement such a code?
Each company has two properties, they are company name and estimated annual earnings.
There are two types of companies: 1- Main company, 2 - Subsidiary company.
The company can belong only to one company but can have a few child companies.
public class Company {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private Integer id;
private String companyName;
private double estimatedAnnualEarnings;
private Company company; // here need to do a #OneToOne
private List<Company> subsidiaryCompany; // here need to do a #OneToMany
}

In your Implementation you should use :
The #Entity annotation in your class level, so the entity can be persisted to database.
The #Column annotation with the companyName and estimatedAnnualEarnings properties, so they can be persisted as columns in the database.
#ManyToOne annotation with the company field, so it can be mapped with a self-reference relationship.
The same goes with the subsidiaryCompany List which needs to be mapped with #OneToMany annotation to have a relationship too.
This is how should be your code:
#Entity
public class Company {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private Integer id;
#Column
private String companyName;
#Column
private double estimatedAnnualEarnings;
#ManyToOne(cascade={CascadeType.ALL})
#JoinColumn(name="mainCompanyId")
private Company mainCompany;
#OneToMany(mappedBy="mainCompany")
private List<Company> subsidiaryCompanies;
//getters and setters goes here
}
Note
I changed the name of company field to mainCompany and
subsidiaryCompaniy to subsidiaryCompanies for better readability
and to make it fit the logic better.
If you want to give your entity a different name in the database you
should use #Table(name="differentName") in the class level with
#Entity annotation, the smae thing with the columns you can add
name property to the #Column annotation i.e
#Column(name="company_name") if you want different names.

JPA/Hibernate: Multiple unnecessary entires in a table

I'm new here and I'm also new in JPA! I developed a little JPA Application with the following entities:
A Customer(Name, Prename, ID) has got an Address(ZIP_Code, city, ID). The relation between these entities is ManyToOne (Customer's perpective) and OneToMany (Address' perspective).
The code (parts) are the following: (without getters/setters)
#Entity
public class AddressEntity implements Serializable{
private static final long serialVersionUID = 1L;
#Id
#GeneratedValue(strategy=GenerationType.AUTO)
private int id;
private String city;
private int zipCode;
#Column
#ElementCollection(targetClass=CustomerEntity.class)
private List<CustomerEntity> customers;
public AddressEntity() {}
#OneToMany(cascade=CascadeType.ALL,
fetch=FetchType.EAGER,
mappedBy="addressentity")
public List<CustomerEntity> getCustomers() {
return customers;
}
#Entity
public class CustomerEntity implements Serializable {
private static final long serialVersionUID = 1L;
#Id
#GeneratedValue(strategy=GenerationType.AUTO)
private int primaryKey;
private String preName, surName;
#ManyToOne(targetEntity=AddressEntity.class, cascade=CascadeType.ALL)
#JoinColumn(name="Address_ID")
private AddressEntity address;
public CustomerEntity() {}
}
Now i want to add two persons:
1) Bart Simpson 1234 Springfield
2) Homer Simpson 1234 Springfield
The problem is that in the address table Springfield appears twice. But that's not the sense of normalized database! How can I realize that an AddressEntity is only added when it does not exists before!
Thanks for answering,
eniac

Persisting new addresses
It is possible to have many AddressEntities with the same City and Zipcode. If you require that a given CustomerEntity is related to the same address, then you must use that specific entity. So use the same AddressEntity that you used for Bart for Homer also
CustomerEntity bart = new CustomerEntity();
//Set Barts’s fields.
CustomerEntity homer = new CustomerEntity();
//Set Homers’s fields.
AddressEntity simpsonsPlace = new AddressEntity();
//set 1234 and Springfield.
bart.setAdress(simpsonsPlace);
homer.setAdress(simpsonsPlace);
List<CustomerEntity> simpsons = new ArrayList<CustomerEntity>();
simpsons.add(bart);
simpsons.add(homer);
simpsonsPlace.setCustomers(simpsons);
yourEntityManager.persist(bart);
yourEntityManager.persist(homer);
Associating with existing addresses.
If you are creating homer sometime after bart then you will want to associate bart to an existing address. It maybe that you will pass in the address from the client. This will have the id field populated and will be a detached entity. You can use this entity with the em.merge() operation;
CustomerEntity homer = new CustomerEntity();
homer.setAdress(detachedAddress);
yourEntityManager.merge(homer);
A better way might be to form a compound PK of Zip/HouseNumber for example. This ensures that any AddressEntity with Zip/HouseNumber combination already in the DB will be treated as a detached object, and will of course only appear once in the DB.
Embeddables
You have both #ElementCollection and #OneToMany targeted to your CustomerEntity Entity.
As per JPA Spec 11.1.14;
The ElementCollection annotation defines a collection of instances of
a basic type or embeddable class.
It is therefore incorrect to target an entity with #ElementCollection.
You require a relation between your two entities, therefore, in you your AddressEntity you just need to use the #OneToMany and drop the #ElementCollection. You also need to map (with mappedBy) to the field of the target class and not the target class itself;
#Entity
public class AddressEntity implements Serializable{
private static final long serialVersionUID = 1L;
#Id
#GeneratedValue(strategy=GenerationType.AUTO)
private int id;
private String city;
private int zipCode;
#OneToMany(cascade=CascadeType.ALL,
fetch=FetchType.EAGER,
mappedBy="address") //Not addressentity
private List<CustomerEntity> customers;
public AddressEntity() {}
public List<CustomerEntity> getCustomers() {
return customers;
}
Alternatively you can define your CustomerEntity as an #Embeddable and not an #entity. In this way you can use #ElementCollection to target CustomerEntity from AddressEntity, but CustomerEntity would not be an entity in its own right and cannot be processed independently as it is dependent on AddressEntity

Why in this Hibernate mapping it is used #ManyToOne instead #OneToOne?

I am absolutly new in Hibernate development and I have the following problem.
I have 2 entity classes that maps 2 DB tables:
1) The first entity class (the main one) is named KM_ProjectInfo and map a DB table named KM_PROJECT.
2) The second entity class is named KM_ProjectInfoStatus and map a DB table named KM_PROJECT_INFO_STATUS.
So the second one represent a specific field of the first one (a status of the row representd by an instance of the KM_ProjectInfo class). Infact I have something like this:
1) KM_ProjectInfo class:
#Entity
#Table(name = "KM_PROJECT")
public class KM_ProjectInfo implements Serializable {
#Id
#GeneratedValue
private Long idProjectInfo;
#Column(name = "name")
private String name;
#Column(name = "technology")
private String technology;
#ManyToOne
#JoinColumn(name = "idCountry")
private KMCountry country;
#Column(name = "power")
private long power;
#Column(name = "cod")
private String cod;
#ManyToOne
#JoinColumn(name = "idProjectInfoStatus")
private KM_ProjectInfoStatus status;
// GETTERS & SETTERS
}
2) KM_ProjectInfoStatus:
#Entity
#Table(name = "KM_PROJECT_INFO_STATUS")
public class KM_ProjectInfoStatus implements Serializable {
#Id
#GeneratedValue
private Long idProjectInfoStatus;
#Column(name = "foldertech")
private Long foldertech;
#Column(name = "folderproject")
private Long folderproject;
// GETTERS & SETTERS
}
So, as you can see in the previous snippet, the KM_ProjectInfoStatuss is a field of the KM_ProjectInfo because I want that it contains the primary key of this table as foreign key.
In the logic of my application I want that at one row of the KM_PROJECT table (so at one instance of the KM_ProjectInfo entity class) is associated a single row of the KM_PROJECT_INFO_STATUS (one instance of the KM_ProjectInfoStatus entity class) because it represent a specific status for the KM_PROJECT row.
In my code I have:
#ManyToOne
#JoinColumn(name = "idProjectInfoStatus")
private KM_ProjectInfoStatus status;
but I think that is wrong because at one row of my first table it is associated a specific single row of the second table. But maybe I am missing something about how Hibernate work.
Can you help me to understand what I am missing? What it work? Why I have #ManyToOne instead #OneToOne?
Tnx

It all depends on how you want to model things. In terms of Database structure, OneToOne and ManyToOne are implemented in the same way:
One or more JoinColumns which makes a foreign key pointing to the primary key of the other table.
So both solutions correctly map to your database, but it depends if you want to allow several KM_ProjectInfo to point to the same KM_ProjectInfoStatus, or only allow a single one.
Note that, even though you would declare a OneToOne, you could still end up with multiple KM_ProjectInfo pointing to the same KM_ProjectInfoStatus if you don't manipulate Hibernate properly.
Here you did not declare the reverse relationship, but if you did, the declaration would have to be different:
In case of a OneToOne, you would have a KM_ProjectInfo member
In case of a OneToMany (reverse of ManyToOne), you would have a Collection<KM_ProjectInfo> member

From the description it seems you want to have one-to-one relationship. That is the project entity should have its very own status not shared by any other project. You could achieve this by using #OneToOne as below.
#Entity
#Table(name = "KM_PROJECT")
public class KM_ProjectInfo implements Serializable {
#Id
#GeneratedValue
private Long idProjectInfo;
#OneToOne
#JoinColumn(name = "idProjectInfoStatus")
private KM_ProjectInfoStatus status;
}
#Entity
#Table(name = "KM_PROJECT_INFO_STATUS")
public class KM_ProjectInfoStatus implements Serializable {
#Id
#GeneratedValue
private Long idProjectInfoStatus;
#OneToOne(mappedBy="idProjectInfoStatus")
private KM_ProjectInfo project;
}
This way you can have specific status for the KM_PROJECT.
Coming back to #ManyToOne, you will want to have this if you want to share the same status with multiple projects, but that's not what you want in your case. I have tried to explain mappings in simple way here One-to-One mapping.

Persist an entity that have associated a list of entities that use #idclass

I have an Evaluation entity that has an associated list of EvaluationEvaluator. I need to explicitly create that entity because it required an extra column "STATUS". Before I continue evaluation. I do: evaluation.setEvaluationEvaluator(listEvaluator) where listEvaluator is a list of EvaluationEvaluator type. Then persist(evaluation). When I run this, it does not throw any kind of exception. But in the database, it inserts in the Evaluation table, and not inserted into the EvaluationEvaluator table.
Below my Evaluation entity.
#Entity
public class Evaluation implements Serializable{
#Id
#GeneratedValue
private Long id;
//MORE FIELDS
#OneToMany(mappedBy="evaluation")
private List<EvaluationEvaluator> evaluators;
//CONSTRUCTORS
//GETTER AND SETTERS
}
This is my EvalutionEvaluator Entity:
#Entity
#Table(name= "EVALUATION_EVALUATOR")
#IdClass(EvaluationEvaluatorId.class)
public class EvaluationEvaluator implements Serializable{
#Id
#Column(name="EMPLOYEE_ID", insertable=false , updatable=false)
private Long EmployeeID;
#Id
#Column(name="EVALUATION_ID", insertable=false, updatable=false)
private Long EvaluationID;
#ManyToOne
#JoinColumn(name"EMPLOYEE_ID")
private Employee employee;
#ManyToOne
#JoinColumn(name"EVALUATION_ID")
private Evaluation evaluation;
#NotNull
private String status;
//CONSTRUCTORS
//GETTER AND SETTERS
}
This is my EvaluationEvaluatorId class
public class EvaluationEvaluatorId implements Serializable{
private Long employeeID;
private Long evaluationID;
//CONSTRUCTOR
//GETTER AND SETTERS
}
And finally, this is my EvaluationBean class
#Stateful
#Named
#LocalBean
#ConversationScoped
public class EvaluationBean {
#PersistentContext(type= PersistenceContextType.EXTENDED)
private EntityManager em;
#Inject
Conversation conversation;
private Evaluation evaluation;
//IN MY WEBPAGE I IMPLEMENT PRIMEFACES PICKLIST AND IT REQUIRE DUALIST TO HANDLE
private DualListModel<Employe> evaluators;
private EvaluationEvaluator evaluationEvaluator;
private List<EvaluationEvaluator> listEvaluators;
#Inject
private EmployeeList employeeList;
//GETTER AND SETTERS
public String begin(){
if (conversation.isTransient()){
converstaion.begin();
}
evaluationEvaluator = new EvaluationEvaluator();
listEvaluators = new ArrayList<EvaluationEvaluator>();
evaluation = new Evaluation();
List<Employee> source = employeeList.findAll();
target = new ArrayList<Employee>();
evaluators = new DualListModel<Employee>(source, target);
return "/evalution/evaluationAsig.xhtml"
}
public String save(){
Iterator<Employee> iterator = evaluators.getTarget().iterator();
while (iterator.hasNext()){
EvaluationEvaluator ev = new EvaluationEvaluator();
ev.setEmployee(iterator.next());
listEvaluators.add(ev);
}
evalution.setEvaluationEvaluators(listEvaluators);
if(evaluation.getId()==null){
em.persist(evalution);
} else{
em.merge(evalution);
}
if(!conversation.isTransient()){
convesation.end();
}
return "/evalution/evaluationsAsig.xhtml"
}
}
When I debug my application,apparently everything is correct, but I mentioned above, doesn't persist in EvaluationEvaluator table.

Your #OneToMany association is missing cascading configuration.
Add cascade = CascadeType.ALL or cascade = {CascadeType.PERSIST, CascadeType.MERGE} to the #OneToMany annotation. JPA assumes no cascading by default so you would need to persist each EvaluationEvaluator by yourself explicitely otherwise.
UPDATE
There is another thing wrong with the code - the Ids of EvaluationEvaluators are never assigned. You have a complex key made of two Long columns. Both are marked not insertable nor updatable which tells to JPA that the id is going to be somehow generated on database level and it should not care about it. There is however no sequence configured explicitely in your entity (although it is not necessarily required) and also from your comment:
I did what you recommended but it throws the following exception. "A different object with same identifier was already associated with the session"
I assume that this is not the case and both id column values default to null or zero and are same for all EvaluationEvaluators you are trying to persist. If you'd like the database to generate the id for you automatically use #GeneratedValue - Configure JPA to let PostgreSQL generate the primary key value - here you can find explanation how to do this (the database part is database dependent, this is for PostgreSQL). The most common use case however, is to configure the sequence but let hibernate pick the next value, instructions here - Hibernate sequence on oracle, #GeneratedValue(strategy = GenerationType.AUTO)