Still More, Re: Questions about RPG service program as an SQL UDF

Ok.
After attempting to read all 240k records made Squirrel collapse with a cry of <fly = "David Hedison">HELLLLLLLLPPPPPPP MEEEEEEEEEEE!!!!</fly>, I tried again, limiting it to 1% of the total number of records.

Reading a censored view from a connection that did *not* have the date format set to ISO (and thus, throwing exceptions for all zeroed-out and invalid date conversions) took 6 minutes, 26 seconds.

Reading an uncensored view of the same file from a connection that did *not* have the date format set to ISO took 5 minutes, 19 seconds.

Reading the same censored view from a connection that *did* have the date format set to ISO (eliminating the date exceptions) took 2 minutes, 43 seconds.

Reading the same uncensored view from a connection that *did* have the date format set to ISO took under 25 seconds: it was done while I was still trying to figure out why the I/O count and RRN weren't climbing in step with each other like the other three tests. I'm guessing that this one, because it's not throwing date exceptions, and isn't censored, is getting better blocking.

*****

Regarding the "FOR READ ONLY" and "WITH NC" clauses:

I'm somewhat familiar with the WITH NC clause, since my SQL-CLI-based search program has been using it for so long, I'd forgotten what it was for, until I just looked it up. And wouldn't the fact that the VIEW is full of fields processed by UDFs tend to imply "FOR READ ONLY"? At any rate, I can certainly experiment with them.

--
JHHL