| Re: [eclipselink-users] Batch reading - poor performance |
There is a bug/enhancement request for this issue. Please vote for this bug.
You should be able to join-fetch the relationship in the query for now.
mkaluzny wrote:
>
> Hi,
>
> I've got problem with batch reading. The current solution implemented in
> EclipseLink offers unacceptable performance with tables that contains a
> lot of rows. Here is an example:
>
> @Entity Container {
> @ManyToOne Content lastContent;
> }
>
> @Entity Content {
> @ManyToOne Container container;
> }
>
> Let's assume that I've got one million containers, and each of my
> containers contains about one million of content entities. I want to query
> for a list of containers:
>
> SELECT c Container c ORDER BY c.modifiedDate [firstResult=0,maxResults=20]
>
> If I set batch reading query hint for Container.lastContent relationship,
> the batch reading query will ignore the order by, firstResult and
> maxResults and will fetch one million of content entities (lastMessage for
> each container entity in the database).
>
> Are there any plans to add to EclipseLink batch reading solution similar
> to hibernate's IN selects? Current solution is unacceptable for scenarios
> similar to described.
>
> P.S. I've found a ticket for this:
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=211699
> But it is open since december 2007, and has no target milestone set.
>
>
-----
---
http://wiki.eclipse.org/User:James.sutherland.oracle.com James Sutherland
http://www.eclipse.org/eclipselink/
EclipseLink , http://www.oracle.com/technology/products/ias/toplink/
TopLink
Wiki: http://wiki.eclipse.org/EclipseLink EclipseLink ,
http://wiki.oracle.com/page/TopLink TopLink
Forums: http://forums.oracle.com/forums/forum.jspa?forumID=48 TopLink ,
http://www.nabble.com/EclipseLink-f26430.html EclipseLink
Book: http://en.wikibooks.org/wiki/Java_Persistence Java Persistence
--
View this message in context: http://www.nabble.com/Batch-reading---poor-performance-tp23354914p23369937.html
Sent from the EclipseLink - Users mailing list archive at Nabble.com.