During schema generation, ignore duplicate column references.

[vnayar]

When creating @manytoone relationships, the column is typically located Currently, during schema generation, a check is put in place to avoid attempting to create duplicate columns. For example:

@Entity class Pet {
  @Id string name;
  @ManyToOne @JoinColumn("person_fk") Person owner; // OK
  // @ManyToOne @JoinColumn("person_fk") Person owner2; // ERROR - Duplicate column
}

@Entity class Person {
  @Id string name;
}

The code above represents the tables, which allow many pets to reference the same person.

CREATE TABLE pet (name varchar, person_fk varchar);
CREATE TABLE person (name varchar);

However, there are new ways to have a @manytomany relationship when using an @EmbeddedId. Consider RoboPets, which have a name, and also a manufacturer. Many RoboPets can share the same manufacturer, there's a many-to-one relationship.

@Embeddable class PetPk {
  string name;
  string manufacturerId;
}

@Entity class RoboPet {
  @EmbeddedId PetPk pk;
  // This works on existing schemas, but fails while generating schemas with the error:
  // "duplicate column name robot_pet.manufacturer_id in schema"
  @ManyToOne @JoinColumn("manufacturer_id") Manufacturer manufacturerId;
}

@Entity class Manufacturer {
  @Id string manufacturerId;
}

The code above represents the tables:

CREATE TABLE robo_pet(name varchar, manufacturer_id varchar);
CREATE TABLE manufacturer(manufacturer_id);

The many-to-one relationship is established because many RoboPets share the same "manufacturer_id" column. This relationship is currently flagged as an error, because the column mentioned in @joincolumn on RoboPet.manufacturerId is already defined, however, it is a valid relationship and functions on schemas that have already been created.