Rob,
My apologies beforehand for any misconceptions I may have on this issue. I
reckon that I focused at first on the ability of using a select/omit within
the index (my bad).
That said I wonder if this could be a case of "working as designed". Before
I go further please have in mind a couple of points: 1) I don't have access
to a 7.1 system. My comments come solely of reading a lot of PDFs every
time a new release (or TR) comes out (I really love the platform, and try
to keep current) and, 2) Although I do read a lot in English, it is not my
first language and sometimes it is really easy to arrive to the wrong
presumption...
Googling a little (the issue is intriguing...), I found this APAR :
http://www-912.ibm.com/n_dir/nas4apar.nsf/ALLAPARS/SE45098
that states (in part):
**********
"When a user creates an index from a generated SQL CREATE INDEX
statement AND the user wishes to preserve the same format level
ID, the user must specify the same date and time format in the
environment that were specified when the index was originally
created. The same applies to tables and views.
However, if an index shares the based on table (physical file)
format, we will enhance QSQGNDDL to omit generating the RCDFMT
clause. This will ensure that in this one case, the format level
ID will be the same if the underlying table is the same."
**********
The second paragraph seems to imply that this could be a "normal" behaviour
for this? Maybe someone in the community can have a better explanation for
this (or yes, maybe it is time for a PMR :-)
Best Regards,
Luis
Luis Rodriguez
IBM Certified Systems Expert — eServer i5 iSeries
midrange.com
