[hibernate-issues] [Hibernate-JIRA] Commented: (HHH-5633) AbstractEmptinessExpression producing incorrect sql for innerselect

[minh huon (JIRA)]

    [ http://opensource.atlassian.com/projects/hibernate/browse/HHH-5633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=38621#action_38621 ] 

minh huon commented on HHH-5633:
--------------------------------

Further examination of the produced query lends me to believe the construction is wrong or I'm creating an incorrect Criteria object.  In my test I want a list of Shipment objects where its dnas list is not empty.  In reality, you'd never have a Shipment object without an associated DNA type object but I still think the query production is incorrect and even with adding an alias to the inner query, the query produced doesn't make sense because its not filtering on the Shipment id of the outer query so in essence, if it was possible to have a Shipment object without any dnas, then the query would return false positives because of the *{color:red}exists (select 1 ...){color:red}* doesn't check for a match but only checks for existence if that makes any sense.



> AbstractEmptinessExpression producing incorrect sql for innerselect
> -------------------------------------------------------------------
>
>                 Key: HHH-5633
>                 URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-5633
>             Project: Hibernate Core
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 3.5.0-Final
>         Environment: eclipse gallileo windows 7 32 bit mysql
>            Reporter: minh huon
>         Attachments: hibernate bug.zip
>
>
> Attached are the domain objects that are used in question for the query.  I cannot see a way to explicitly give an alias to the Restrictions.isNotEmpty when the query produce on the inside of the isNotEmpty is a derived query table.
> Looking at the code for version 3.5.0 AbstractEmptinessExpression like number 68:
> {code}
> String innerSelect = "(select 1 from " + collectionPersister.getTableName()
> 		        + " where "
> 		        + new ConditionFragment().setTableAlias( sqlAlias ).setCondition( ownerKeys, collectionKeys ).toFragmentString()
> 		        + ")";
> {code}
> The string creation seems to need an alias at the end of the creation to be correct my mysql syntax.
> The test case "shipmentQuery" is the one failing to produce the right query.
> I'd create a patch but I'm not sure if that is exactly the problem or an I doing something wrong.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira