Eclipse Community Forums: EclipseLink » Performance issues persisting large amounts of similar entities

Home

Home » Eclipse Projects » EclipseLink » Performance issues persisting large amounts of similar entities

Performance issues persisting large amounts of similar entities [message #655699]

Tue, 22 February 2011 05:59

Eclipse User

Hi everybody!

Our application needs to persist results of mathematical calculations. A result is represented by curves.
There is an entity for a result, an entity for a curve and an entity for a point of a curve (with doubles for x and y).
The result maintains a oneToMany relationship to the curve and the curve maintains a oneToMany relationship to the point.

Now it is not uncommon to have hundreds of curves and thousands of points per curve in one result making a total half a million points.

So, em.persist(result) either takes a couple of minutes to complete, which is unpleasent or a couple of minutes to throw an "OutOfMemoryError: Java heap space", which, of course, is inacceptable.

Are there JPA / EclipseLink means to speed things up?

Thanks

Re: Performance issues persisting large amounts of similar entities [message #655911 is a reply to message #655699]

Wed, 23 February 2011 04:27

Eclipse User

Hello

The EclipseLink performance guide contains some useful guidelines how to
tune the performance on JPA, JDBC and Database level. Especially, there
are some examples and hints to handle "batch" use cases which I think is
your case:

http://wiki.eclipse.org/Optimizing_the_EclipseLink_Applicati on_(ELUG)

Hope this helps
Adrian

Re: Performance issues persisting large amounts of similar entities [message #655912 is a reply to message #655699]

Wed, 23 February 2011 04:29

Eclipse User

Re: Performance issues persisting large amounts of similar entities [message #655913 is a reply to message #655699]

Wed, 23 February 2011 04:29

Eclipse User

Re: Performance issues persisting large amounts of similar entities [message #656344 is a reply to message #655699]

Fri, 25 February 2011 03:26

Eclipse User

Thanks Adrian.

I have tried eclipselink.jdbc.batch-writing = JDBC and it does reduce time consumption to maybe 75%.

But meanwhile, I tend to kick out the point entity and use a BLOB of the Java-serialized list of points. Time consumption is reduced to 5% this way.

Previous Topic:	[Resolved] Load a external jar with entities and persistence.xml
Next Topic:	Foreign Key Constraint Issue Involving One-To-Many and Many-To-One Relationship in EcliplseLink2.1.

Goto Forum:

-=] Back to Top [=-

Current Time: Sun Nov 09 11:21:43 EST 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter