Hi, I am trying to migrate from 4.3.6.Final to 5.1.0 Final to get java 8 date time api mapping and unrelated join entities features.
But unfortunately I got an issue related with sequence generation.
Here is my test:
{code}
for (int i = 0; i < 50; i++) {
Client client = new Client();
client.setName("Test#" + i);
clientRepository.save(client);
System.out.println(client.getId());
}
{code}
{code}
@Entity
@Table(name = "client")
@Getter
@Setter
class Client {
@Id
@GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "c_seq")
@SequenceGenerator(name = "c_seq", sequenceName = "c_seq", allocationSize = 20)
@Column(name = "ID_CLIENT")
public Integer id;
....
}
{code}
CREATE SEQUENCE c_seq start 1 increment 1;
Console output:
Hibernate:
select
nextval ('c_seq')
Hibernate:
select
nextval ('c_seq')
1
2
Hibernate:
select
nextval ('c_seq')
-16
-15
-14
-13
-12
-11
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
org.springframework.dao.DataIntegrityViolationException: A different object with the same identifier value was already associated with the session : [igor.domain.Client#1]
I've investigated what's going on. Actually you have changed sequence strategy generation
old: SequenceHiLoGenerator
new: SequenceStyleGenerator
And SequenceStyleGenerator uses PooledOptimizer.
I've spent a lot of time debugging PooledOptimizer and got that it happens due to this piece of code
{code}
else if ( generationState.value.gt( generationState.hiValue ) ) {
generationState.hiValue = callback.getNextValue();
generationState.value = generationState.hiValue.copy().subtract( incrementSize - 1 );
}
{code}
But it works fine with Hibernate 4.3.6.Final
Console output:
Hibernate:
select
nextval ('c_seq')
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Hibernate:
select
nextval ('c_seq')
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
Hibernate:
select
nextval ('c_seq')
60
61
62
63
64
65
66
67
68
69
Hibernate:
/* insert igor.domain.Client
*/ insert
into
client
(age, name, id_client)
values
(?, ?, ?) (x50 times)
As we can see everything is fine, allocationSize works and there is no exception with 4 . 3.6.
I am really wondering about negative ids and this issue |
|