[hibernate-issues] [Hibernate-JIRA] Commented: (HHH-7060) Add support for mode that causes SQL Server dialect to treat application as semantically authoritative in regards to Java types and their value ranges

[Guenther Demetz (JIRA)]

    [ https://hibernate.onjira.com/browse/HHH-7060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=45512#comment-45512 ] 

Guenther Demetz commented on HHH-7060:
--------------------------------------

Following the procedure I used to convert tinyint columns into smallint without loosing data (SQLServer2008):

alter table MyTable
add  byteValueTrans smallint null

GO
update MyTable
set byteValueTrans = byteValue
where byteValue between 0 and 127

update MyTable
set byteValueTrans = byteValue - 256
where byteValue between 128 and 256

GO

alter table MyTable
drop column byteValue

GO

sp_rename 'MyTable.byteValueTrans', 'byteValue', 'COLUMN'

> Add support for mode that causes SQL Server dialect to treat application as semantically authoritative in regards to Java types and their value ranges
> ------------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: HHH-7060
>                 URL: https://hibernate.onjira.com/browse/HHH-7060
>             Project: Hibernate ORM
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 4.1.0
>            Reporter: John Verhaeg
>            Assignee: John Verhaeg
>
> In particular, there is a disconnect between the value ranges supported by bytes in Java vs. SQL Server: Java byte values are 1 byte, unsigned, while SQL Server values are 0-255.  Previously, negative byte values stored in the Java model would be treated as unsigned bytes by the driver, allowing the reading and writing of such values to "work" from the standpoint of the application, but also resulting in a completely different positive value being stored within the database.  HHH-6815 addressed this disconnect from the standpoint of ensuring values set in the Java model were kept semantically in tact within the database and forcing the use of SMALLINT columns instead of TINYINT columns for byte values.  However, this solution assumes it's important that the database values exactly match what's in the Java model.  It also does not allow for existing databases with existing TINYINT columns, which end up experiencing exceptions when negative byte values are persisted and later retrieved.  The proposed solution is to offer some type of mode that can be set by the user to indicate the Java model should be treated as authoritative wrt data values, which would cause Hibernate to perform as it did in previous releases.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira