Apr 11th, 2008, 12:47 PM
Use Cases Supported in 1.0.0-Final
Referring to the official Spring Batch website, nicely presented use cases are defined.
May I know which of them can be realized using the features in 1.0.0-Final?
I am particularly interested in the following two use cases. If 1.0.0-Final supports these use cases, any samples can be provided to illustrate how to implement them?
* Massively Parallel Batch Processing
* Scheduled Processing
Apr 11th, 2008, 02:21 PM
The scheduling use case as described is supported, to the extent that it ever will be (Spring Batch needs something external to trigger a JobLauncher, like a shell script). The 'massively parallel' use case is possible to implement with 1.0, but isn't directly supported out of the box yet. (Actually there is more than one way to approach it. The easiest is to flush to some durable middleware on a chunk boundary.) I have some proof of concept code that I will be publishing shortly. I chose to implement it with Spring Integration, which makes it much easier, but any kind of durable middleware could be adapted to the same pattern natively as well.
Apr 12th, 2008, 11:17 AM
In the Massively Parallel Batch Processing use case, it mentioned a class org.springframework.core.batch.support.provider.Pa rtioningInterceptor . But I cannot find this in 1.0.0-Final api. May I know if it is achieved by another ways?
Besides, I found that the use cases documentation at: http://static.springframework.org/sp...ses/index.html are very useful. But it seems contains some out-of-date details, which are still referring to mx versions things like ItemProvider. Could you help to fix the use case documentation to avoid the confusion?
Apr 12th, 2008, 09:13 PM
Thanks for pointing out that the inaccuracy. I've created an issue for tracking:
Regarding the partitioning, the parallel sample job should serve as a good example.