sql function - allow to define the scope of deterministic

Hello,
if I have understand correctly, setting deterministic a function when I create it means that sql can cache the results to improve performance.

the sql cache is purged by the IPL. so this means that the scope of deterministic for a sql function is the IPL.

The sql registry, the sql global variable and some sql function provided by IBM have a different deterministic scope, and this is well documented in the sql reference.
for example :
"The VERIFY_GROUP_FOR_USER function is deterministic within a connection"

the ability to define the scope of deterministic of a function is an excellent method to permit a function to give the correct result, and to allow sql to use the sql cache as often as possible.

today, when the deterministic scope of a user function is not the IPL, then the function should be created as not deterministic, and cannot be used in a mask or a permission.

being able to scope deterministic allow a better tuning, then better performance.

I suggest these levels of scope :
- IPL (the current level)
- job
- connection (such as VERIFY_GROUP_FOR_USER)
- statement (such as registry TIMESTAMP)
- UserDemand
This last one is a way for an application to signal to the sql cache it should selectively purge the cache
for example
when I do a CHGUSRPRF USRPRF(MYPRF) GRPPRF(xxx , then the cache of
VERIFY_GROUP_FOR_USER('MYPRF'
is purged, then the next call to VERIFY_GROUP_FOR_USER('MYPRF'
returns the correct result.

allow a not deterministic function to be a few deterministic