On 1/22/2013 7:00 AM, michael keith wrote:
> Ideally we would have some kind of annotation that just pointed to the other attribute to signify that the column was
> defined by it and that only one of the mappings should write to the column or contribute to schema gen (a bit like the
> @MapsId situation, but in the opposite direction). Since in practice something already works, though, and there's not
> really any time left to add new features to the spec in jsr 338 it won't likely happen in this round. Maybe in the
> future if enough people ask for it?
>
Right.
BTW, IIRC, in past we ran up against the issue that vendors differed in their use of this technique for
overlapping PK/FK mappings -- which was a problem that we solved with the introduction of @MapsId.
> On 22/01/2013 8:50 AM, Nicolas Seyvet wrote:
>> I mean to say that it was not the intent of the spec to detail this. Let me try to ask this question differently:
>> Would it be good to standardize how the vendors support this as part of jsr-338 scope?
>>
>> /Nicolas
>>
>> -----Original Message-----
>> From: michael keith [mailto:michael.keith_at_oracle.com]
>> Sent: Tuesday, January 22, 2013 2:24 PM
>> To: jsr338-experts_at_jpa-spec.java.net
>> Cc: Nicolas Seyvet
>> Subject: Re: [jsr338-experts] Re: insertable = false, updatable = false& field re-use
>>
>> No, I never said the spec "will not clarify the re-use of a table column". I just filled in some of the history and
>> how we got to the current state. There is nothing in the spec that disallows reusing a column in this way, and this is
>> how the vendors support it, so you shouldn't encounter any problems in your application.
>>
>> -Mike
>>
>> On 22/01/2013 7:06 AM, Nicolas Seyvet wrote:
>>> " and is what most vendors support since it at least uses standard options instead of a vendor-specific read-only
>>> annotation."
>>> Ok, but this is still not a specification.
>>>
>>> I guess the point is that this specification will not clarify the re-use of a table column within the same entity. Is
>>> that correct?
>>>
>>> /Nicolas
>>>
>>> -----Original Message-----
>>> From: michael keith [mailto:michael.keith_at_oracle.com]
>>> Sent: Monday, January 21, 2013 2:29 PM
>>> To: jsr338-experts_at_jpa-spec.java.net
>>> Cc: Linda DeMichiel
>>> Subject: [jsr338-experts] Re: insertable = false, updatable = false&
>>> field re-use
>>>
>>> In JPA 1.0 we did not get around to specifying a read-only mapping (in fact we still haven't).
>>> Because of that, using the insertable=false, updatable=false combination in the @Column mapping became the de facto
>>> way of setting a mapping to be read-only. It is not ideal since it requires setting multiple physical column/join
>>> column attributes when a logical read-only option would be the more appropriate option, but for the most part it has
>>> stuck and is what most vendors support since it at least uses standard options instead of a vendor-specific read-only
>>> annotation.
>>>
>>> -Mike
>>>
>>> On 17/01/2013 3:21 PM, Linda DeMichiel wrote:
>>>> Hi Nicolas,
>>>>
>>>> On 1/14/2013 12:01 AM, Nicolas Seyvet wrote:
>>>>> Hi,
>>>>>
>>>>> After using Hibernate for a while, I had assumed that insertable =
>>>>> false and updatable = false where useful for supporting reuse of
>>>>> columns across fields. But I got into an argument this week end
>>>>> with someone saying that the JPA spec does not define how/if this is
>>>>> mandated.
>>>>>
>>>>> Example:
>>>>> @Entity
>>>>> public class Company {
>>>>>
>>>>> @Id
>>>>> @GeneratedValue(strategy = GenerationType.IDENTITY)
>>>>> private Long id = null;
>>>>> private String name = null;
>>>>>
>>>>> @OneToMany(cascade = CascadeType.ALL, mappedBy = "company")
>>>>> private Set<Employee> employees = new HashSet<Employee>();
>>>>>
>>>>> [..]
>>>>> }
>>>>>
>>>>> @Entity
>>>>> public class Employee {
>>>>>
>>>>> @Id
>>>>> @GeneratedValue(strategy = GenerationType.IDENTITY)
>>>>> private Long id = null;
>>>>> private String name = null;
>>>>>
>>>>> @ManyToOne()
>>>>> @JoinColumn(name = "COMPANY_ID")
>>>>> private Company company = null;
>>>>>
>>>>> // Reuse of the field for the FK id
>>>>> @Column(name = "COMPANY_ID", insertable = false, updatable = false)
>>>>> private Long companyId;
>>>>>
>>>>> }
>>>>>
>>>>> I went through the JPA 2.1 draft, can could not find anything
>>>>> specific to this topic.
>>>>>
>>>>> Should this be clarified? Or was it discussed before I joined?
>>>>>
>>>> This was discussed back in the JPA 1.0 days. The semantics of
>>>> insertable / updatable are defined in the Column annotation and are
>>>> consistent with your use of them in the example above.
>>>>
>>>> -Linda
>>>>
>>>>
>>>>> Best regards,
>>>>> Nicolas Seyvet