hibernate 下的oracle Id Generator sequence

hiaf055u6m

　　@SequenceGenerator(name="COMMONSEQ",sequenceName="COMMON_SEQ")
 @GeneratedValue(strategy = GenerationType.SEQUENCE,generator="COMMONSEQ")
　　
　　
　　基本配置
    @Id
    @GeneratedValue(strategy=GenerationType.SEQUENCE) 
    public Integer getId() {
        return this.id;
    }
 
 这样hibernate自动生成一个名叫hibernate_sequence的sequence
　　SQL> select * from USER_SEQUENCES;
　　
　　
　　hibernate使用sequence的后台动作:
下面是一个save()动作,hibernate 先发出一个 select hibernate_sequence.nextval from dual;
Hibernate:
    select
        hibernate_sequence.nextval
    from
        dual
Hibernate:
    insert
    into
        DvdType
        (name, id)
    values
        (?, ?)
1:typeoracle5
 
　　
用自动创建的hibernate_sequence的缺陷: 会使得多个表共用一个sequnce,造成同一个表序号不连续
TYPE表:
1:TypeExample21
2:TypeExample21
3:TypeExample21
4:TypeExample22
5:TypeExample23
6:TypeExample24
7:TypeExample25
8:TypeExample26
9:TypeExample27
10:TypeExample28
11:TypeExample29
12:TypeExample210
 QUALITY表
13:Qualityoracle5
14:QualityExample21
15:QualityExample22
16:QualityExample23
17:QualityExample24
18:QualityExample25
19:QualityExample26
20:QualityExample27
21:QualityExample28
22:QualityExample29
23:QualityExample210
 
 
所以建议,还是为每个表定义具体的sequence name
　　
自定义sequence
@Entity
@SequenceGenerator(name="TypeSEQ",sequenceName="TypeSEQ_DB")
public class DvdType implements java.io.Serializable {
　　@Id
    //@GeneratedValue            
    @GeneratedValue(strategy=GenerationType.SEQUENCE, generator="TypeSEQ")
 @Entity
@SequenceGenerator(name="QualitySEQ",sequenceName="QualitySEQ_DB")
public class QualityType implements java.io.Serializable {
　　@Id
    //@GeneratedValue
    @GeneratedValue(strategy=GenerationType.SEQUENCE, generator="QualitySEQ")
 
00:07:46,437 INFO  [STDOUT] Hibernate:
    select
        TypeSEQ_DB.nextval
    from
        dual
 00:08:52,437 INFO  [STDOUT] Hibernate:
    select
        QualitySEQ_DB.nextval
    from
        dual
 
　　
　　hibernate的sequence ,生成的id,事务间缺省是50 increment
现象:
hibernate自动生成的sequence,在oracle下看,increment_by还是1
　　SQL> select * from USER_SEQUENCES;
　　SEQUENCE_NAME                   MIN_VALUE  MAX_VALUE INCREMENT_BY C O CACHE_SIZE LAST_NUMBER
------------------------------ ---------- ---------- ------------ - - ---------- -----------
SEQUENCE_DVD_TABLE_1                    1 1.0000E+27            1 N N          0        1511
USERSEQ_DB                              1 1.0000E+27            1 N N         20           1
GROUPSEQ_DB                             1 1.0000E+27            1 N N         20          21
 
但通过hibernate/JPA进行save/persist()操作,ID却是increment +50
　　SQL> insert into USERGROUP(ID,NAME) values(GROUPSEQ_DB.nextval,'group6');
　　1 row created.
　　SQL> select * from usergroup;
　　ID NAME
---------- --------------------------------------------------
       200 group1
       250 group1
       300 group3
       350 group4
       400 group5
         9 group6               //这是手工insert的,上面5条记录是hibernate persist的
　　6 rows selected.
　　SQL> select GROUPSEQ_DB.nextval from dual;
　　NEXTVAL
----------
        10
 
原因:
1.hibernate 对sequence参数以自己为准,不读oracle的参数
2.hibernate increment缺省就是50,大概跟 hibernate集群应用有关
 
解决:hibernate提供了allocationSize参数来设      
@Entity
@SequenceGenerator(name="GroupSEQ",sequenceName="GroupSEQ_DB",allocationSize=1)
public class UserGroup {
　　
SQL> select * from usergroup;
　　ID NAME
---------- --------------------------------------------------
       200 group1
       250 group1
       300 group3
       350 group4
       400 group5
        11 group8               这两条是用hibernate新存的
        12 group9
         9 group6