[
https://issues.apache.org/jira/browse/OPENJPA-2179?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13280073#comment-13280073
]
Mark Struberg commented on OPENJPA-2179:
----------------------------------------
I digged deeper and found the reason for the sub-selects to be the DISTINCT.
Of course a CLOB or BLOB field cannot be used in a distinct by most DBs. But
most databases also optimize them away if they there is the primary key in the
result.
The corresponding code is in the HandlerFieldStrategy line 174:
http://svn.apache.org/viewvc/openjpa/trunk/openjpa-jdbc/src/main/java/org/apache/openjpa/jdbc/meta/strats/HandlerFieldStrategy.java?view=markup
> 'distinct' and 'join' combinations lead lots of unneccessary sub-queries for
> @Embedded and @Lob fields
> ------------------------------------------------------------------------------------------------------
>
> Key: OPENJPA-2179
> URL: https://issues.apache.org/jira/browse/OPENJPA-2179
> Project: OpenJPA
> Issue Type: Bug
> Components: kernel
> Affects Versions: 2.2.0
> Reporter: Mark Struberg
> Assignee: Mark Struberg
> Fix For: 2.3.0
>
> Attachments: OPENJPA-2179-test-1.patch
>
>
> I have an Entity (Course) with a simple @Embedded field and a @Lob. I do not
> use any LAZY attribution on them!
> If I do a normal em.find, the entity will be loaded as a whole (all the
> fields, including the embedded and the lob will be fetched immediately).
> Sidenote: the Lecturer referred in the select is defined as
> @OneToMany(mappedBy = "course",
> cascade = {CascadeType.PERSIST, CascadeType.REMOVE,
> CascadeType.MERGE},
> orphanRemoval = true, fetch = FetchType.EAGER)
> @OrderColumn(name = "POSITION")
> private List<Lecturer> lecturers;
> The following selects DO work
> * "select c from Course c join c.lecturers l "
> * "select distinct c from Course c"
> The following selects create tons of subqueries! 1 separate sub-query for
> each @Embedded field, and also for each @Lob
> * "select distinct c from Course c join c.lecturers l "
> * "select distinct c from Lecturer l join l.course c"
> * "select c from Lecturer l join l.course c"
> This happens ONLY if I run this stuff against Oracle. In MySQL it seems to
> work properly.
> I'll try to create a unit test for it.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
