Re: Size of QRPLOBJ is large when empty

On 1/28/11 9:28 AM, Luis Rodriguez wrote:

Just curious...
Is that index saved with SAVLIB? What happens if the OP does a
SAVLIB/RSTLIB to QRPLOBJ?

The RSTLIB to a new library name causes an effective CRTLIB, thus a new library "index" implementation object; growing only with each object restored also as part of the RSTLIB request. However in this case, that means having first deleted QRPLOBJ using DLTLIB. Thus moot for QRPLOBJ, since that library might best be left alone or simply be allowed\forced to be re-created by some use of REPLACE(*YES); i.e. why even use RSTLIB QRPLOBJ? One might consider there is potential value in RSTLIB QRPLOBJ after DLTLIB QRPLOBJ, for example if the creation date of the library would best be maintained for some reason, but there is always the risk that the OS could delete and re-create for its own [recovery] reasons versus only by a user-requested DLTLIB. Regardless of how CRTLIB is effected after DLTLIB, the [unlikely case of customized] private authorities to the library is one attribute that would need to be considered, and if one chose to then a user-issued CRTLIB would best consider also the *PUBLIC authority.

And while the SAVLIB\DLTLIB\RSTLIB combination could reduce the index size for another [e.g. a user] library, except for a [near] empty library, I would probably almost never do that [nor by RNMOBJ, CRTLIB, MOVOBJ from new name to created\original name, then DLTLIB], simply per the amount of work involved on the part of the user and system, just to gain probably very little in storage. The size of the index would need to be very excessive in a library expected never to have as many objects again; so unless the library size was the side effect of a rogue process having created many objects incorrectly into that library, and the guilty application since corrected, I would probably just leave most "large" *LIB objects as is... just as with the *LIB QRPLOBJ.

Regards, Chuck