Fay, thanks for your reply! "openjpa.jdbc.EagerFetchMode" is currently set to parallel. "openjpa.jdbc.SubclassFetchMode" is set to parallel as well. The description of parallel mode seems to be exactly what I want to happen, but not what seems to be happening.
Is there any other setting, or anything in the object graph that would prevent this behavior? > > > -----Original Message----- > From: Fay Wang [mailto:fyw...@yahoo.com] > Sent: Thursday, July 02, 2009 9:19 AM > To: users@openjpa.apache.org > Subject: Re: SQL generation for OneToMany object graph > > > Hi, it seems that you are using the Join strategy for your eager fetch > mode. Can you try parallel? > <property name="openjpa.jdbc.EagerFetchMode" value="parallel"/> > > Fay > > > > ----- Original Message ---- > From: David Minor <davemi...@gmail.com> > To: users@openjpa.apache.org > Sent: Monday, June 29, 2009 5:23:25 PM > Subject: SQL generation for OneToMany object graph > > Hi, > > If I have a deep object graph with a root element that has OneToMany > relationships with other elements, and those have OneToMany > relationships (and so on), and all the objects are included via a fetch > group, how does openjpa decide how to generate the SQL for a statement > like: > > "SELECT r FROM RootElement r WHERE r.uid IN (:list)" > > Looking at the SQL logging (this is using mysql BTW), the initial > statements use an OR to select among rows in the root element's table, > and some of the other initial tables, but then it seems to start issuing > individual SELECT statements as it gets deeper into the graph. > Is there any way to make it use OR combined with JOIN statements > instead? > > Thanks, > Dave > > -- > _____________ > David Minor > > > > > > > > -- _____________ David Minor