[hibernate-issues] [Hibernate-JIRA] Updated: (HHH-3930) one-to-one causes redundant select query
Strong Liu (JIRA)
noreply at atlassian.com
Wed Oct 19 13:52:19 EDT 2011
[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-3930?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Strong Liu updated HHH-3930:
----------------------------
Description:
{code}
@Entity
public class Address {
@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private Long id;
@OneToOne(mappedBy = "address")
private Customer customer;
}
@Entity
public class Customer {
@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private Long id;
@OneToOne(fetch = FetchType.EAGER, cascade = CascadeType.ALL)
private Address address = new Address();
public Long getId() {
return id;
}
public Address getAddress() {
return address;
}
}
{code}
This mapping causes 2 instead of the expected 1 query to retrieve a Customer and its Address from the db:
select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.id=?
select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.address_id=?
Changing the mapping to a LAZY fetch type:
@OneToOne(fetch = FetchType.LAZY, cascade = CascadeType.ALL)
private Address address = new Address();
Causes 3 select queries instead of the expected 2 queries to retrieve a Customer (and its Address) from the db:
select * from Customer customer0_ where customer0_.id=?
select * from Address address0_ left outer join Customer customer1_ on address0_.id=customer1_.address_id where address0_.id=?
select * from Customer customer0_ where customer0_.address_id=?
The third select is superfluous because the relationship is already completely known: you already have the customer, so why not just set it on the address entity?
Making the address field in Customer a @ManyToOne doesn't make a difference.
Making the customer field in Address a @OneToMany does remove the extra select, but forces our model to change the relationship from Customer to List<Customer> where we *know* there'll be only 1 element.
Apparently Hibernate can figure out the reverse relationship with a @ManyToOne - @OneToMany without the need for additional queries, can't this be extended to @OneToOne bidirectional relationships as well?
was:
@Entity
public class Address {
@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private Long id;
@OneToOne(mappedBy = "address")
private Customer customer;
}
@Entity
public class Customer {
@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private Long id;
@OneToOne(fetch = FetchType.EAGER, cascade = CascadeType.ALL)
private Address address = new Address();
public Long getId() {
return id;
}
public Address getAddress() {
return address;
}
}
This mapping causes 2 instead of the expected 1 query to retrieve a Customer and its Address from the db:
select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.id=?
select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.address_id=?
Changing the mapping to a LAZY fetch type:
@OneToOne(fetch = FetchType.LAZY, cascade = CascadeType.ALL)
private Address address = new Address();
Causes 3 select queries instead of the expected 2 queries to retrieve a Customer (and its Address) from the db:
select * from Customer customer0_ where customer0_.id=?
select * from Address address0_ left outer join Customer customer1_ on address0_.id=customer1_.address_id where address0_.id=?
select * from Customer customer0_ where customer0_.address_id=?
The third select is superfluous because the relationship is already completely known: you already have the customer, so why not just set it on the address entity?
Making the address field in Customer a @ManyToOne doesn't make a difference.
Making the customer field in Address a @OneToMany does remove the extra select, but forces our model to change the relationship from Customer to List<Customer> where we *know* there'll be only 1 element.
Apparently Hibernate can figure out the reverse relationship with a @ManyToOne - @OneToMany without the need for additional queries, can't this be extended to @OneToOne bidirectional relationships as well?
> one-to-one causes redundant select query
> ----------------------------------------
>
> Key: HHH-3930
> URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-3930
> Project: Hibernate Core
> Issue Type: Bug
> Components: core
> Affects Versions: 3.3.1
> Reporter: Martijn Dashorst
>
> {code}
> @Entity
> public class Address {
> @Id
> @GeneratedValue(strategy = GenerationType.AUTO)
> private Long id;
> @OneToOne(mappedBy = "address")
> private Customer customer;
> }
> @Entity
> public class Customer {
> @Id
> @GeneratedValue(strategy = GenerationType.AUTO)
> private Long id;
> @OneToOne(fetch = FetchType.EAGER, cascade = CascadeType.ALL)
> private Address address = new Address();
> public Long getId() {
> return id;
> }
> public Address getAddress() {
> return address;
> }
> }
> {code}
> This mapping causes 2 instead of the expected 1 query to retrieve a Customer and its Address from the db:
> select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.id=?
> select * from Customer customer0_ left outer join Address address1_ on customer0_.address_id=address1_.id where customer0_.address_id=?
> Changing the mapping to a LAZY fetch type:
> @OneToOne(fetch = FetchType.LAZY, cascade = CascadeType.ALL)
> private Address address = new Address();
> Causes 3 select queries instead of the expected 2 queries to retrieve a Customer (and its Address) from the db:
> select * from Customer customer0_ where customer0_.id=?
> select * from Address address0_ left outer join Customer customer1_ on address0_.id=customer1_.address_id where address0_.id=?
> select * from Customer customer0_ where customer0_.address_id=?
> The third select is superfluous because the relationship is already completely known: you already have the customer, so why not just set it on the address entity?
> Making the address field in Customer a @ManyToOne doesn't make a difference.
> Making the customer field in Address a @OneToMany does remove the extra select, but forces our model to change the relationship from Customer to List<Customer> where we *know* there'll be only 1 element.
> Apparently Hibernate can figure out the reverse relationship with a @ManyToOne - @OneToMany without the need for additional queries, can't this be extended to @OneToOne bidirectional relationships as well?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the hibernate-issues
mailing list