[ 
https://issues.apache.org/jira/browse/OPENJPA-2390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Di Wu Lau updated OPENJPA-2390:
-------------------------------

    Attachment: OPENJPA-2390.patch

Problem is fixed by adding LIMIT 1 to the nextSequenceQuery String in 
HSQLDictionary. This force the sequence to hand back only 1 value at a time. 

Any version that is lower than 2.1.0 does not need to add LIMIT 1. The older 
versions do not have the skipping problem.

To make the code more compatible, I added some conditions in HSQLDictionary to 
check the HSQL versions so that only version 2.1.0 and beyond will add LIMIT 1 
to the nextSequenceQuery.

Uploadded is my patch file. In this file, I added two test cases that use 
allocationSize=1 and allocationSize=2 to demonstrate the problem.

Thanks.

                
> HSQLDB SELECT NEXT VALUE FOR Sequence will skip 1 sequence value if 
> allocation size is 1
> ----------------------------------------------------------------------------------------
>
>                 Key: OPENJPA-2390
>                 URL: https://issues.apache.org/jira/browse/OPENJPA-2390
>             Project: OpenJPA
>          Issue Type: Bug
>          Components: jdbc
>    Affects Versions: 2.2.1
>            Reporter: René Link
>         Attachments: HSQLSequenceTest.java, OPENJPA-2390.patch
>
>
> The HSQLDictionary sets the DBDictionary.nextSequenceQuery to "SELECT NEXT 
> VALUE FOR {0} FROM INFORMATION_SCHEMA.SYSTEM_SEQUENCES" when the 
> HSQLDictionary is initialized.
> The NativeJDBCSeq uses
>  _select = MessageFormat.format(format, new Object[]{name, 
> String.valueOf(_allocate * _increment)});
> to build the select next value sql. The format is pased 2 arguments "name" 
> and "_allocate * _increment"
> but the second is not honored by the nextSequenceQuery.
> So when the allocation size is set to 1 the hsqldb nextSequenceQuery 
> statement will return 2 next values
> anyway. That is why always one next value will be skipped... 1, 3, 5, 7, and 
> so on.
> The nextSequenceQuery should be set with an limit. E.g.
> "SELECT NEXT VALUE FOR {0} FROM INFORMATION_SCHEMA.SYSTEM_SEQUENCES LIMIT {1}"
> I will attatch an unit test that uses in memory hsqldb to show the problem 
> soon.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to