Mapping a Single Entity to Multiple Tables in JPA

1. Introduction

JPA makes dealing with relational database models from our Java applications less painful. Things are simple when we map every table to a single entity class. But, sometimes we have reasons to model our entities and tables differently:

In this short tutorial, we’ll see how to tackle this last scenario.

2. Data Model

Let’s say we run a restaurant, and we want to store data about every meal we serve:

  • name

  • description

  • price

  • what kind of allergens it contains

Since there are many possible allergens, we’re going to group this dataset together. Furthermore, we’ll also model this using the following table definitions:

image

Now let’s see how can we map these tables to entities using standard JPA annotations.

3. Creating Multiple Entities

The most obvious solution is to create an entity for both classes.

Let’s start by defining the Meal entity:

@Entity
@Table(name = "meal")
class Meal {

    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "id")
    Long id;

    @Column(name = "name")
    String name;

    @Column(name = "description")
    String description;

    @Column(name = "price")
    BigDecimal price;

    @OneToOne(mappedBy = "meal")
    Allergens allergens;

    // standard getters and setters
}

Next, we’ll add the Allergens entity:

@Entity
@Table(name = "allergens")
class Allergens {

    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "meal_id")
    Long mealId;

    @OneToOne
    @PrimaryKeyJoinColumn(name = "meal_id")
    Meal meal;

    @Column(name = "peanuts")
    boolean peanuts;

    @Column(name = "celery")
    boolean celery;

    @Column(name = "sesame_seeds")
    boolean sesameSeeds;

    // standard getters and setters
}

In the above example, we can see that meal_id is both the primary key and also the foreign key. That means we need to define the one-to-one relationship column using @PrimaryKeyJoinColumn.

However, this solution has two problems:

  • We always want to store allergens for a meal, and this solution doesn’t enforce this rule

  • The meal and allergen data belong together logically – therefore we might want to store this information in the same Java class even though we created multiple tables for them

One possible resolution to the first problem is to add the @NotNull annotation to the allergens field on our Meal entity. JPA won’t let us persist the Meal if we have a null Allergens.

However, this is not an ideal solution; we want a more restrictive one, where we don’t even have the opportunity to try to persist a Meal without Allergens.

4. Creating a Single Entity with @SecondaryTable

We can create a single entity specifying that we have columns in different tables using the @SecondaryTable annotation:

@Entity
@Table(name = "meal")
@SecondaryTable(name = "allergens", pkJoinColumns = @PrimaryKeyJoinColumn(name = "meal_id"))
class Meal {

    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "id")
    Long id;

    @Column(name = "name")
    String name;

    @Column(name = "description")
    String description;

    @Column(name = "price")
    BigDecimal price;

    @Column(name = "peanuts", table = "allergens")
    boolean peanuts;

    @Column(name = "celery", table = "allergens")
    boolean celery;

    @Column(name = "sesame_seeds", table = "allergens")
    boolean sesameSeeds;

    // standard getters and setters

}

Behind the scenes, JPA joins the primary table with the secondary table and populates the fields. This solution is similar to the @OneToOne relationship, but this way, we can have all of the properties in the same class.

It’s important to note that if we have a column that is in a secondary table, we have to specify it with the table argument of the @Column annotation. If a column is in the primary table, we can omit the table argument as JPA looks for columns in the primary table by default.

Also, note that we can have multiple secondary tables if we embed them in @SecondaryTables. Alternatively, from Java 8, we can mark the entity with multiple @SecondaryTable annotations since it’s a repeatable annotation.

5. Combining @SecondaryTable With @Embedded

As we’ve seen, @SecondaryTable maps multiple tables to the same entity. We also know that @Embedded and @Embeddable to do the opposite and map a single table to multiple classes.

Let’s see what we get when we combine @SecondaryTable with @Embedded and @Embeddable:

@Entity
@Table(name = "meal")
@SecondaryTable(name = "allergens", pkJoinColumns = @PrimaryKeyJoinColumn(name = "meal_id"))
class Meal {

    @Id
    @GeneratedValue(strategy = GenerationType.IDENTITY)
    @Column(name = "id")
    Long id;

    @Column(name = "name")
    String name;

    @Column(name = "description")
    String description;

    @Column(name = "price")
    BigDecimal price;

    @Embedded
    Allergens allergens;

    // standard getters and setters

}

@Embeddable
class Allergens {

    @Column(name = "peanuts", table = "allergens")
    boolean peanuts;

    @Column(name = "celery", table = "allergens")
    boolean celery;

    @Column(name = "sesame_seeds", table = "allergens")
    boolean sesameSeeds;

    // standard getters and setters

}

It’s a similar approach to what we saw using @OneToOne. However, it has a couple of advantages:

  • JPA manages the two tables together for us, so we can be sure that there will be a row for each meal in both tables

  • Also, the code is a bit simpler, since we need less configuration

Nevertheless, this one-to-one like solution works only when the two tables have matching ids.

It’s worth mentioning that if we want to reuse the Allergens class, it would be better if we defined the columns of the secondary table in the Meal class with @AttributeOverride.

6. Conclusion

In this short tutorial, we’ve seen how we can map multiple tables to the same entity using the @SecondaryTable JPA annotation.

We also saw the advantages of combining @SecondaryTable with @Embedded and @Embeddable to get a relationship similar to one-to-one.

As usual, the examples are available over on GitHub.