Updating aggregate with no "column" members (only "Id" and relations) does not work

[gushakov]

Hello Spring Data JDBC team,

I might have stumbled on a little peculiarity (maybe a bug). If we have an aggregate which does not contain any fields annotated with @Column but does contain an ID field and possibly some one-to-one relations to some other entity, like this:

@Table
public class Foobar {

    @Id
    private String id;

    private Wambaz wambaz;
    
    // getters, setters, all-argument constructor omitted
}

@Table
public class Wambaz {

    @Column
    private String name;

    // getters, setters, all-argument constructor omitted
}

Then we can insert this aggregate OK, but we cannot update it (say by specifying a new value for a relation). For example:

String foobarId = UUID.randomUUID().toString();
// this works
jdbcAggregateTemplate.insert(new Foobar(foobarId, new Wambaz("wam")));

// this will give an error
jdbcAggregateTemplate.update(new Foobar(foobarId, new Wambaz("baz")));

Here is the typical error:

Caused by: org.springframework.jdbc.BadSqlGrammarException: PreparedStatementCallback; bad SQL grammar [UPDATE "foobar" WHERE "foobar"."id" = ?]; nested exception is org.postgresql.util.PSQLException: ERROR: syntax error at or near "WHERE"

I've checked it with Spring Data JDBC v. 2.4.2 (from spring-boot-starter-data-jdbc v. 2.7.2) and Postgres database.

Thank you for your attention and you work in this excellent framework.

George