Darren,
If I'm not mistaken, IBM enforces the "short circuit" Boolean logic within SQL, meaning that within a series of OR clauses, test for the MOST LIKELY term first. If you change your logic slightly, you MIGHT get slightly better performance...
/FREE
exec sql declare C1 cursor for
Select TWS
TUSER,
TTDTE
from TFILE
where (:Device = '*ANY' or :Device = TWS )
And (:User = '*ANY' or :User = TUser )
/END-FREE
The implication is that if the first term results in a TRUE, then the rest of the Boolean expression does not even get evaluated.
Hth,
-Eric DeLong
-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of darren@xxxxxxxxx
Sent: Thursday, August 11, 2011 12:28 PM
To: midrange-RPG RPG message board
Subject: Conditional SQL processing against host variables
Using something like the following statement, we take a pretty high
performance hit compared to writing multiple cursors for times when the
user has selected a wildcard (*ANY) as shown below. I suspect that the SQL
is actually looking for TUSER='*ANY' even though it will never find it, and
its not necessary. Is there a better way to condition this type of
operation? I'm aware of dynamic SQL, but its a bit of a pain for larger
statements.
/FREE
exec sql declare C1 cursor for
Select TWS
TUSER,
TTDTE
from TFILE
where (TWS = :Device or :Device = '*ANY')
And (TUser = :User or :User = '*ANY')
/END-FREE
--
This is the RPG programming on the IBM i / System i (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit:
http://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at
http://archive.midrange.com/rpg400-l.
midrange.com
