A year ago this past April, I was able, in my UDF experiments, to
achieve proof-of-concept on a UDF, based on an ILE RPG SRVPGM, to censor
rows returned in an SQL view.
And now, I've about forgotten most of what I did.
I was able to find the RPGLE source member for the proof-of-concept
censorship function. And I've found the docs on how to execute a CREATE
FUNCTION statement for an external UDF.
But any general advice on the subject would be appreciated.
As would the answer to a couple of specific questions:
(1) If I update the service program, do I then have to DROP the UDF and
CREATE it again?
(2) My censorship function's return value ("C" for censor, "P" for pass)
is derived from the parameters, and from the user profile under which
the function is executed. Can I assume that cached results of
DETERMINISTIC functions aren't shared across job boundaries? That this
function that's deterministic for any given user is DETERMINISTIC for
SQL purposes?
--
JHHL
