I have a colleague who adamantly insist that running statistic only without ever doing REORG;s is perfectly sufficient for DB2 optimizer to construct an efficient access plan. Can this be true in any way? What happens to statistics if this tables are frequently refreshed - resume & also replaced?
Can you give me 5 to 10 reasons for having a regular REORD for DB2?
And finally, what are the critical DB2 columns for DB2 access path development? I also adamantly insist that without REORG, the underlying DB2 dataset may remain fragmented, clustering out of wack if it is in that state already, wasted space not reclaimed where the DB is fragmented, etc.
Can someone help me expand on this iwith solid verifiable arguments for having a consistent and regular REORG?