Reverse Engineering

The menu on the top of the window allows you to configure:

• DB connection

• Source root and package where the generated entities will be saved

• Whether indexes and constraints need to be migrated

• Whether schema name should be specified in the @Table annotation

Also, from the Other settings drop-down list, you can move to the entity declaration and reverse engineering settings.

On the left side of the window, you can see:

• Mapped Relations: tables and views mapped to JPA entities

• Tables: tables that exist in the DB but are not mapped to entities

• Views: views that exist in the DB but are not mapped to entities

After selecting any element from the tree, a panel for migrating attributes from columns will appear. Also, you will be able to define a class name in the corresponding field.

The main part of the window allows you to configure everything related to attributes. You can choose which attributes you want to add and change all their params, except Column Name. Mapping type and attribute/converter/hibernate types are represented as drop-down lists.

All attributes are divided into 3 categories:

• Migrated Columns - columns that are already present in the entity (available only for mapped relations)

• Columns - new columns that are not yet mapped in the entity or parent @MappedSuperclass

• References - optional associations that are not represented as a column in the observed table

Parent Entities

IntelliJ IDEA offers the ability to define a parent entity by selecting a class annotated with @MappedSuperclass from the Parent drop-down box. This allows the generated entities to extend from the parent class and automatically inherit all attributes that have the same name and type.

In cases where the column name in the @MappedSuperclass doesn't match the child entity's table, we can still inherit the attribute using the @AttributeOverride annotation. By simply selecting the attribute name and choosing the one to override, IntelliJ IDEA assists in managing the inheritance.

During entity generation, IntelliJ IDEA alerts us if any inherited attributes from the @MappedSuperclass are missing in the database. To align the model with the database, access the Generate DDL by Entities action in the JPA Structure menu and select the Existing DB update option.

// TODO Comments

If you want to postpone an attribute creation for specific columns, you can choose //todo comment as the mapping type. IntelliJ IDEA will generate the //todo comment with the corresponding quick-fix actions depending on the column type. You can call these actions by pressing :

• For known basic and association types you can:

  • Uncomment as is

  • Remove column mapping

• For unknown column type you can:

  • Define target Java type

  • Uncomment as is

  • Remove column mapping

Here is an example of generated //todo comment for the attribute with an unknown column type:

/*
  TODO [Reverse Engineering] create field to map the 'description' column
   Available actions: Define target Java type | Uncomment as is | Remove column mapping
  @Column(name = "description", columnDefinition = "jsonb")
  private java.lang.Object description;
*/

After calling the Define target Java type action, the following window will appear:

IntelliJ IDEA will remember data mappings for the subsequent reverse engineering actions. You can always change them in the settings.

IntelliJ IDEA follows all best practices providing the most efficient mapping for DB views while reverse engineering:

1. As DB views do not have a primary key, IntelliJ IDEA allows you to select a field or a set of fields to use as the identifier for the target entity.

2. Most DB views are immutable. So, IntelliJ IDEA adds @Immutable annotation to the entity and generates getters only. This helps to improve application performance.

3. IntelliJ IDEA generates only a no-arg protected constructor for entities that are mapped to a DB view, as per JPA specifications, which prevents developers from creating a new instance of such entities in the business logic code

There are two situations where we can confidently assume the cardinality of the relation as @OneToOne:

1. Table has a column with the unique constraint that refers to the primary key of another table

2. Primary key of the table is a foreign key

Case 1:

CREATE TABLE profiles
  (
      id        BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
      join_date date,
      user_id BIGINT,
      status    VARCHAR(255),
      bio       VARCHAR(255),
      CONSTRAINT pk_profiles PRIMARY KEY (id)
  );

  CREATE TABLE users
  (
      id         BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
      last_name  VARCHAR(255),
      first_name VARCHAR(255),
      CONSTRAINT pk_users PRIMARY KEY (id)
  );

  ALTER TABLE profiles
      ADD CONSTRAINT uc_profiles_user UNIQUE (user_id);

  ALTER TABLE profiles
      ADD CONSTRAINT FK_PROFILES_ON_USER FOREIGN KEY (user_id) REFERENCES users (id);

IntelliJ IDEA will generate a @OneToOne association with a @JoinColumn annotation in the User entity, and a @OneToOne association with a mappedBy parameter in the Profile entity:

@Entity
@Table(name = "users")
public class User {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  @Column(name = "id", nullable = false)
  private Long id;

  @OneToOne(fetch = FetchType.LAZY)
  @JoinColumn(name = "profile_id")
  private Profile profile;
}

@Entity
@Table(name = "profiles")
public class Profile {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  @Column(name = "id", nullable = false)
  private Long id;

  @OneToOne(fetch = FetchType.LAZY, mappedBy = "profile")
  private User users;
}

Case №2:

CREATE TABLE users
(
    id BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
    last_name VARCHAR(255),
    first_name VARCHAR(255),
    CONSTRAINT pk_users PRIMARY KEY (id)
);

CREATE TABLE profiles
(
    user_id BIGINT NOT NULL,
    status VARCHAR(255),
    bio VARCHAR(255),
    join_date date,
    CONSTRAINT pk_profiles PRIMARY KEY (user_id)
);

ALTER TABLE profiles
    ADD CONSTRAINT FK_PROFILES_ON_USER FOREIGN KEY (user_id) REFERENCES users (id);

Since @Id should not be a persistence entity, IntelliJ IDEA will generate:

• id attribute of a basic type and mark it with @Id annotation

• users @OneToOne association and mark it with @MapsId annotation

@Entity
@Table(name = "profiles")
public class Profile {
  @Id
  @Column(name = "user_id", nullable = false)
  private Long id;

  @MapsId
  @OneToOne(fetch = FetchType.LAZY, optional = false)
  @JoinColumn(name = "user_id", nullable = false)
  private User users;

  //...
}

@Entity
@Table(name = "users")
public class User {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  @Column(name = "id", nullable = false)
  private Long id;

  @OneToOne(fetch = FetchType.LAZY, mappedBy = "user")
  private Profile profiles;

  //...
}

If a table has the column that refers to the primary key of another table, it is most likely a @ManyToOne association. But you are also able to change cardinality to @OneToOne if required. So, depending on which table you call the reverse engineering action, IntelliJ IDEA will detect mapping type as @OneToMany or @ManyToOne:

CREATE TABLE users
(
    id BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
    last_name VARCHAR(255),
    first_name VARCHAR(255),
    CONSTRAINT pk_users PRIMARY KEY (id)
);

CREATE TABLE profiles
(
    id BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
    join_date date,
    status VARCHAR(255),
    bio VARCHAR(255),
    user_id BIGINT,
    CONSTRAINT pk_profiles PRIMARY KEY (id)
);

ALTER TABLE profiles
    ADD CONSTRAINT FK_PROFILES_ON_USER FOREIGN KEY (user_id) REFERENCES users (id);

IntelliJ IDEA will generate the following code:

@Entity
@Table(name = "users")
public class User {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  @Column(name = "id", nullable = false)
  private Long id;

  @OneToMany(mappedBy = "user")
  private Set<Profile> profiles = new LinkedHashSet<>();

  //...
}

@Entity
@Table(name = "profiles")
public class Profile {
  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  @Column(name = "id", nullable = false)
  private Long id;

  @ManyToOne(fetch = FetchType.LAZY)
  @JoinColumn(name = "user_id")
  private User user;

  //...
}

To establish a many-to-many relationship between two tables, you need to use a junction table. The junction table, in this case, contains only two columns - foreign keys. IntelliJ IDEA can automatically detect such a table and identify the relation cardinality between the two tables whose ids are represented as foreign keys in the junction table as @ManyToMany.

CREATE TABLE users
(
    id BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
    last_name VARCHAR(255),
    first_name VARCHAR(255),
    CONSTRAINT pk_users PRIMARY KEY (id)
);

CREATE TABLE profiles
(
    id BIGINT GENERATED BY DEFAULT AS IDENTITY NOT NULL,
    join_date date,
    status VARCHAR(255),
    bio VARCHAR(255),
    CONSTRAINT pk_profiles PRIMARY KEY (id)
);

CREATE TABLE profiles_users
(
    profile_id BIGINT NOT NULL,
    users_id BIGINT NOT NULL,
    CONSTRAINT pk_profiles_users PRIMARY KEY (profile_id, users_id)
);

ALTER TABLE profiles_users
    ADD CONSTRAINT fk_prouse_on_profile FOREIGN KEY (profile_id) REFERENCES profiles (id);

ALTER TABLE profiles_users
    ADD CONSTRAINT fk_prouse_on_user FOREIGN KEY (users_id) REFERENCES users (id);

If this association does not exist in any of the entities, IntelliJ IDEA will generate it in the entity for which the reverse engineering action was called.

@Entity
@Table(name = "users")
public class User {
   @Id
   @GeneratedValue(strategy = GenerationType.IDENTITY)
   @Column(name = "id", nullable = false)
   private Long id;

   @ManyToMany
   @JoinTable(name = "profiles_users",
      joinColumns = @JoinColumn(name = "users_id"),
      inverseJoinColumns = @JoinColumn(name = "profile_id"))
   private Set<Profile> profiles = new LinkedHashSet<>();

   //...
}

If this association already exists in one of the entities, then IntelliJ IDEA will generate the @ManyToMany attribute with the mappedBy parameter.

@Entity
@Table(name = "profiles")
public class Profile {
   @Id
   @GeneratedValue(strategy = GenerationType.IDENTITY)
   @Column(name = "id", nullable = false)
   private Long id;

   @ManyToMany(mappedBy = "profiles")
   private Set<User> users = new LinkedHashSet<>();

   //...
}

1. Fetch Type. To follow best practices and avoid potential performance issues, IntelliJ IDEA sets FetchType.LAZY for @OneToOne and @ManyToOne associations by default.

2. Validation Annotations. Validation annotations give you another layer of protection in addition to the DB constraints. By default, IntelliJ IDEA will apply such annotations over entity attributes while reverse engineering.

3. Pluralization. By default, IntelliJ IDEA uses the singular form for entity names. For example, if you have a table called users, IntelliJ IDEA will generate a User entity. If you disable this option, IntelliJ IDEA will keep the original name of the table and only capitalize the first letter – Users.

4. Basic type attribute. When this option is enabled, IntelliJ IDEA will analyze the ORM references in the database schema and generate basic type attributes instead of creating associations or relationships between entities. This can be useful in certain scenarios where you prefer to have simple attribute types instead of complex associations.

To preserve comments added to the database objects, IntelliJ IDEA transfers them to the corresponding entity using the Hibernate @Comment annotation or JavaDocs, depending on your settings.

tip

Please note that only @Comment annotations on entities can be included in the generated DDL scripts, depending on the settings, JavaDocs will be ignored anyway.

When the application works with several DBMSs, your schema might have slightly different data types for each of them.

Let's say the application needs to support both PostgreSQL and MS SQL and you need to store Unicode characters in string data. PostgreSQL supports Unicode chars in VARCHAR, but MS SQL has a separate NVARCHAR data type for it.

IntelliJ IDEA lets you specify type mappings for each DBMS. It is also possible to set mappings for JPA Converters and Hibernate Types:

See how you can configure type mappings for reverse engineering in IntelliJ IDEA to make use of the @JavaType annotation from Hibernate 6: