I’m writing an import and need to save two types of data (parent & child) to their respective tables and a link table (parentChild) to join them up. I know the error is pretty self explanatory and there have been lots of other people asking similar questions – but this is different. Between here, chatGpt, searching the project for examples and my own stuff I haven’t been able to come up with something that works. I think it’s a problem with the FK constraint but it’s the same pattern I’ve used elsewhere. It clearly *thinks *I’m sending null but somehow it’s not pulling the ids from the two objects being linked.

I can’t post the actual code but here’s a noddy version that recreates the situation.

DDL (run in mssql):

CREATE SEQUENCE Project.TableParentIdSeq START WITH 1 INCREMENT BY 1;
CREATE SEQUENCE Project.TableChildIdSeq START WITH 1 INCREMENT BY 1;
CREATE SEQUENCE Project.TableParentChildIdSeq START WITH 1 INCREMENT BY 1;

CREATE TABLE Project.TableParent
(
    ParentId BIGINT NOT NULL CONSTRAINT DF_Project_TableParent_TableParentIdSeq DEFAULT (NEXT VALUE FOR Project.TableParentIdSeq),
    ParentName VARCHAR(10)  NOT NULL
);

CREATE TABLE Project.TableChild
(
    ChildId BIGINT NOT NULL CONSTRAINT DF_Project_TableChild_TableChildIdSeq DEFAULT (NEXT VALUE FOR Project.TableChildIdSeq),
    ChildName VARCHAR(10)  NOT NULL
);

CREATE TABLE Project.TableParentChild
(
    ParentChildId BIGINT NOT NULL CONSTRAINT DF_Project_TableParentChild_TableParentChildIdSeq DEFAULT (NEXT VALUE FOR Project.TableParentChildIdSeq),
    ParentId BIGINT NOT NULL,
    ChildId BIGINT NOT NULL
);

ALTER TABLE Project.TableParent
    ADD
        CONSTRAINT PK_TableParent PRIMARY KEY (ParentId);

ALTER TABLE Project.TableChild
    ADD
        CONSTRAINT PK_TableChild PRIMARY KEY (ChildId);

ALTER TABLE Project.TableParentChild
    ADD
        CONSTRAINT PK_TableParentChild PRIMARY KEY (ParentChildId),
        CONSTRAINT FK_TableParentChild_ParentId FOREIGN KEY (ParentId) REFERENCES Project.TableParent (ParentId),
        CONSTRAINT FK_TableParentChild_ChildId FOREIGN KEY (ChildId) REFERENCES Project.TableChild (ChildId);

Entities (each with @entity/getter/setter/all args/no args/table):

public class ParentEntity  implements Serializable {
    @Id
    @Column(name = "ParentId")
    @GeneratedValue(generator = "TableParentIdSeq")
    @SequenceGenerator(name = "TableParentIdSeq", sequenceName = "Project.TableParentIdSeq", allocationSize = 1)
    private Long id;

    private String parentName;

    @OneToMany(cascade = {CascadeType.ALL}, mappedBy = "parent", orphanRemoval = true)
    private List<ParentChildEntity> children = new ArrayList<>();

    public ParentEntity(String parentName) {
        this.parentName = parentName;
    }
}

public class ChildEntity implements Serializable {
    @Id
    @Column(name = "ChildId")
    @GeneratedValue(generator = "TableChildIdSeq")
    @SequenceGenerator(name = "TableChildIdSeq", sequenceName = "Project.TableChildIdSeq", allocationSize = 1)
    private Long id;

    private String childName;

    public ChildEntity(String childName) {
        this.childName = childName;
    }
}

public class ParentChildEntity  implements Serializable {
    @Id
    @Column(name = "ParentChildId")
    @GeneratedValue(generator = "TableParentChildIdSeq")
    @SequenceGenerator(name = "TableParentChildIdSeq", sequenceName = "Project.TableParentChildIdSeq", allocationSize = 1)
    private Long id;

    @ManyToOne //(cascade = {CascadeType.ALL})
    @JoinColumn(name = "ParentId", referencedColumnName = "ParentId", insertable = false, updatable = false)
    private ParentEntity parent;

    @ManyToOne
    @JoinColumn(name = "ChildId", referencedColumnName = "ChildId", insertable = false, updatable = false)
    private ChildEntity child;

    public ParentChildEntity (ParentEntity parent, ChildEntity child) {
        this.parent =  parent;
        this.child = child;
    }
}

Each table has its own repo extending JpaRepository and using only the save().

Service layer code:

    @Transactional
    public void populateTables() {
        ParentEntity parentSaved = parentRepository.saveAndFlush(new ParentEntity("Parent 1"));
        ChildEntity childSaved = childRepository.saveAndFlush(new ChildEntity("Child A"));
        parentChildRepository.save(new ParentChildEntity(parentSaved, childSaved));
    }

If I don’t do that final save to the link table the parent and child are saved successfully.
In debug I can see that both parent and child have ids and the parentChild save() is being called with everything in place.
I had thought the child class should have a OneToMany list as well like the parent, as it has a similar relationship to the link table as the parent but chatGpt said otherwise.
Making the columns nullable (not what I want to do) just means they get set to null when saved with data.
The id generation works on all three tables.

Running that gives this error in the console:
Cannot insert the value NULL into column ‘ParentId’, table ‘Project.TableParentChild’; column does not allow nulls. INSERT fails.

Evaluating that final save() in debug as a saveAndFlush() gives:
could not execute statement; SQL [n/a]; constraint [null]

Any ideas? Hopefully nothing too stupid on my part.