One of the biggest saving on
cpu time, and run time of batch jobs can be achieved by just doing a reorder on
the most used Super descriptor. But the catch is that the reorder has to be
done on the most used super and the super descriptor has to be under normal
compression and all the fields in the super should be always filled. Yes,
unfortunately this comes with the “disclaimer”.
The theory of such a change is
that since the records are loaded back in the sequence in which the super
descriptor exist, the records are fetched more faster and thereby using less
cpu. Please note: the fields that you identify for the super descriptor which
will be used for the re-order has to be always filled otherwise whatever fields
are having null won’t be loaded back into the environment once the reorder
completes. Confirm with your DBA before this is implemented in live production
environment.
