[
https://issues.apache.org/jira/browse/OPENJPA-2913?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dario Napolitano updated OPENJPA-2913:
--------------------------------------
Summary: TableJDBCSeq can generate duplicate IDs if allocate() is called
(was: TableJDBCSeq can genarate duplicate IDs if allocate() is called)
> TableJDBCSeq can generate duplicate IDs if allocate() is called
> ---------------------------------------------------------------
>
> Key: OPENJPA-2913
> URL: https://issues.apache.org/jira/browse/OPENJPA-2913
> Project: OpenJPA
> Issue Type: Bug
> Components: kernel
> Affects Versions: 3.1.2
> Reporter: Dario Napolitano
> Priority: Major
>
> When using the allocate() method of the ID generator of an entity using table
> based generation (TableJDBCSeq), the generator will start from 1 no matter
> what is the current value of the sequence, if it's the first allocation. If a
> new allocation takes place, the generator is likely to use values already
> allocated to other entities.
> Example:
> OpenJPAEntityManager oem = (OpenJPAEntityManager) em;
> Generator idGenerator = oem.getIdGenerator(User.class);
> idGenerator.allocate(1000);
>
> This is because the TableJDBCSeq does not update the current sequence value
> during allocate() calls, but only during next() calls. However, the
> allocate() method is exposed via the Generator interface to request
> preallocation for efficiency.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
