Bug 4747 - SPEC - Verbiage about partition reducer is confusing
SPEC - Verbiage about partition reducer is confusing
Status: RESOLVED FIXED
Product: jbatch
Classification: Unclassified
Component: source
1
All All
: P5 trivial
: ---
Assigned To: cvignola
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2013-03-08 22:04 UTC by mminella
Modified: 2013-03-08 22:23 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description mminella 2013-03-08 22:04:41 UTC
Section 8.2.6.4 states: "Programmatic interception of the partition reducer lifecycle is possible by providing a partition reducer."  What does this actually mean?
Comment 1 cvignola 2013-03-08 22:15:43 UTC
It means we have an infinite loop in the spec :)
Comment 2 cvignola 2013-03-08 22:23:12 UTC
Thanks for catching that, Michael.  It should have said:

"Programmatic interception of the partitioned step's lifecycle is possible through the partition reducer."

I rewrote the entire paragraph to say: 

"A partitioned step may execute with an optional partition reducer.  A partition reducer provides a kind of unit of work demarcation around the processing of the partitions.  Programmatic interception of the partitioned step's lifecycle is possible through the partition reducer. The reducer element specifies a reference to a PartitionReducer batch artifact;  see section 9.5.2 for further information."

This change will appear in PFD v1.6.