原文地址：http://blog.sina.com.cn/s/blog_8d8425f301014dls.html

1.設定Pojo為實體

[email protected] //標識這個pojo是一個jpa實體    

02.public class Users implements Serializable {    

03.} 

2.設定表名

[email protected]    

[email protected](name = "users") //指定表名為users    

03.public class Users implements Serializable {    

04.} 

3.設定主鍵

01.public class Users implements Serializable {    

[email protected]    

03.private String userCode; 

4. 設定欄位型別

通過@Column註解設定，包含的設定如下

.name：欄位名

.unique：是否唯一

.nullable：是否可以為空

.inserttable：是否可以插入

.updateable：是否可以更新

.columnDefinition: 定義建表時建立此列的DDL

.secondaryTable: 從表名。如果此列不建在主表上（預設建在主表），該屬性定義該列所在從表的名字。

[email protected](name = "user_code", nullable = false, length=32)//設定屬性userCode對應的欄位為user_code，長度為32，非空   

02.private String userCode;    

[email protected](name = "user_wages", nullable = true, precision=12, scale=2)//設定屬性wages對應的欄位為user_wages，12位數字可保留兩位小數，可以為空   

04.private double wages;    

[email protected]

06.private Date joinDate; 

5.欄位排序

在載入資料的時候可以為其指定順序，使用@OrderBy註解實現

[email protected](name = "USERS")    

02.public class User {    

[email protected](name = "group_name ASC, name DESC")    

04.private List books = new ArrayList();    

05.} 

6.主鍵生成策略

01.public class Users implements Serializable {    

[email protected]    

[email protected](strategy=GenerationType.IDENTITY)//主鍵自增，注意，這種方式依賴於具體的資料庫，如果資料庫不支援自增主鍵，那麼這個型別是沒法用的   

[email protected](name = "user_id", nullable = false)    

05.private int userId;    

06. 

07. 

08.public class Users implements Serializable {    

[email protected]    

[email protected](strategy=GenerationType.TABLE)//通過一個表來實現主鍵id的自增，這種方式不依賴於具體的資料庫，可以解決資料遷移的問題   

[email protected](name = "user_code", nullable = false)    

12.private String userCode;    

13. 

14. 

15.public class Users implements Serializable {    

[email protected]    

[email protected](strategy=GenerationType.SEQUENCE)//通過Sequence來實現表主鍵自增，這種方式依賴於資料庫是否有SEQUENCE，如果沒有就不能用   

[email protected](name="seq_user")    

[email protected](name = "user_id", nullable = false)    

20.private int userId; 

7.一對多對映關係

有T_One和T_Many兩個表，他們是一對多的關係，註解範例如下

主Pojo

[email protected]    

[email protected](name = "T_ONE")    

03.public class One implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "ONE_ID", nullable = false)    

07.private String oneId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

[email protected](cascade = CascadeType.ALL, mappedBy = "oneId")//指向多的那方的pojo的關聯外來鍵欄位   

11.private Collection manyCollection;  

子Pojo

[email protected]    

[email protected](name = "T_MANY")    

03.public class Many implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "MANY_ID", nullable = false)    

07.private String manyId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

10. 

[email protected](name = "ONE_ID", referencedColumnName = "ONE_ID")//設定對應資料表的列名和引用的資料表的列名   

[email protected]//設定在“一方”pojo的外來鍵欄位上    

13.private One oneId;  

8.多對多對映關係

貌似多對多關係不需要設定級聯，以前用hibernate的時候著實為多對多的級聯頭疼了一陣子，JPA的多對多還需要實際的嘗試一下才能有所體會。

估計JPA的多對多也是可以轉換成兩個一對多的。

第一個Pojo

[email protected]    

[email protected](name = "T_MANYA")    

03.public class ManyA implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "MANYA_ID", nullable = false)    

07.private String manyaId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

[email protected]    

[email protected](name = "TMANY1_TMANY2", joinColumns = {@JoinColumn(name = "MANYA_ID", referencedColumnName = "MANYA_ID")}, inverseJoinColumns = {@JoinColumn(name = "MANYB_ID", referencedColumnName = "MANYB_ID")})   

12.private Collection manybIdCollection;  

第二個Pojo

[email protected]    

[email protected](name = "T_MANYB")    

03.public class ManyB implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "MANYB_ID", nullable = false)    

07.private String manybId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

[email protected](mappedBy = "manybIdCollection")    

11.private Collection manyaIdCollection;  

9.一對一對映關係

主Pojo

[email protected]    

[email protected](name = "T_ONEA")    

03.public class OneA implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "ONEA_ID", nullable = false)    

07.private String oneaId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

[email protected](cascade = CascadeType.ALL, mappedBy = "oneA")//主Pojo這方的設定比較簡單，只要設定好級聯和對映到從Pojo的外來鍵就可以了。   

11.private OneB oneB;   

從Pojo

[email protected]    

[email protected](name = "T_ONEB")    

03.public class OneB implements Serializable {    

04.private static final long serialVersionUID = 1L;    

[email protected]    

[email protected](name = "ONEA_ID", nullable = false)    

07.private String oneaId;    

[email protected](name = "DESCRIPTION")    

09.private String description;    

[email protected](name = "ONEA_ID", referencedColumnName = "ONEA_ID", insertable = false, updatable = false)//設定從方指向主方的關聯外來鍵，這個ONEA_ID其實是表T_ONEA的主鍵   

[email protected]    

12.private OneA oneA;  

10 大欄位

[email protected] //對應Blob欄位型別    

[email protected](name = "PHOTO")    

03.private Serializable photo;    

[email protected] //對應Clob欄位型別    

[email protected](name = "DESCRIPTION")    

06.private String description; 

11.瞬時欄位

不需要與資料庫對映的欄位，在儲存的時候不需要儲存倒資料庫

[email protected]    

02.private int tempValue;    

03. 

04.public int getTempValue(){    

05.get tempValue;    

06.}    

07. 

08.public void setTempValue(int value){    

09.this.tempValue = value;    

10.}