在房间中使用嵌套数据库最正确的方法是什么？

Question

我正在考虑创建一个包含一对多和多对多元素的模式。由于我在这方面有点新手，我搜索了一些资源，发现了两种类型的示例；1.将下层对象的内容放入上层，然后进行处理。第二种类型是创建外键，您认为哪一种更正确？

1.

@Entity
public class User {
   @PrimaryKey
   public final int id;
   public final String login;
   public final String avatarUrl;

}

@Entity
public class Tweet {
   @PrimaryKey
   public long id;
   public String body;
   public String createdAt;
   public long userId;

   // this field will be ignored by Room, but still can be used in other places in the Twitter app
   @Ignore 
   public User user;
}

@Entity
public class User {
    @PrimaryKey
    public final int id;
    public final String login;
    public final String avatarUrl;
}


@Entity(foreignKeys = @ForeignKey(entity = User.class,
                                  parentColumns = "id",
                                  childColumns = "userId",
                                  onDelete = CASCADE))
public class Tweet {
    @PrimaryKey
    public long id;
    public String body;
    public String createdAt;
    public long userId;
}

这两个选项中，哪一个是正确且有效的？

另外，这是我想要创建的数据库模式。。你觉得我该怎么办。。请用java展示我不认识kotlin的例子

{
 "UserDatabase":
 [
   {
     "sql_id": 0,
     "name": "Test Name",
     "username": "null",
     "gender": "male",
     "profile_photo_path": "path",
     "pro_version": true,
     "date_of_birth": "01/01/0000",
     "weight": 100,
     "height": 186,
     "waist_size": 0,
     "neck_size": 0,
     "hip_size": 0,
     "bmr_formula": "Benedict",
     "macro_formula": "Fitness",
     "starting_date": "18/02/2022",
     "saved_items":
     [
       {
         "sql_id": 0,
         "parent_id": 0,
         "item_id": 0,
         "item_name": "name",
         "item_image": "item image",
         "item_url": "url",
         "item_editor": "editor"
       }
     ],
     "days":
     [
       {
         "sql_id": 0,
         "parent_id": 0,
         "id": 0,
         "date": "16/02/2022",
         "weight": 96,
         "steps": 25252525,
         "taken_kcal": 257.56,
         "burned_kcal": 27.86,
         "carb": 25.52,
         "protein": 5.2,
         "fat": 8.4,
         "water_ml": "3250",
         "water_percent": 27.68,
         "daily_water_list":
         [
           {
             "sql_id": 0,
             "parent_id": 0,
             "id": 0,
             "time": 1644999314504,
             "water_ml": 300
           }
         ],
         "daily_meals":
         [
           {
             "sql_id": 0,
             "parent_id": 0,
             "id": 0,
             "food_meal_time": "Meal Time",
             "food_name": "Name",
             "food_image": "image url",
             "food_kcal": 300.56,
             "food_portion": "Portion"
           }
         ]
       }
     ]
   }
 ]
}

Answer 1

更好的方法是使用外键，因为这样可以加强引用完整性。

除此之外，除了在子（ren）（Tweet）中包含父（用户）之外，没有什么区别。这是不需要的（因为你引用了父），它与正常化相矛盾（即它复制数据）。

更具体地说

添加外键引入了约束（规则），这些约束（规则）表示引用父表（用户）的子表（鸣叫）中的值必须是父表中存在的值，否则会发生冲突。

CASCADE的onDelete操作将自动删除父级，从而帮助/简化维护引用完整性。