Post-Processing Logic

The product supports executing one or more algorithms after all the threads of a given batch job have completed. This allows for some special processing to occur at the end of a batch job. Algorithms for this plug-in spot receive the batch control, batch run number, batch processing business date, number of threads and the list of the ad hoc parameters of the batch job.

Note: This plug-in spot is available for all Java based batch programs. For programs of type Java (converted), the individual batch programs need to explicitly support this plug-in spot. Unless otherwise noted, assume that a Java (converted) program does not support it.

The following are some examples of functionality that may be executed at the end of a batch job:

Another dependent batch job can be kicked off. Note that this use case is only needed when the multiple dependent jobs are not part of a scheduler (which can also detect the successful end of one batch job so as to submit the next job).
Statistics for the batch run may be analyzed and based on results, a To Do Entry may be sent to an administrator.
If the current batch job is processing a large number of child records in multiple threads, a parent record could be updated to a different status or with some other audit information.

Note that the units of work for all threads are committed prior to executing the post-processing logic. The algorithm should perform standard error handling. If an error occurs in one of the post-processing algorithms, the overall batch job’s status is set to Error so that it can be re-submitted to retry the logic in the finalize step.