RE: Chain vs SetLL/ReadE

Here is what I have been told, I believe from an IBM contactor ( its been a while)...

The primary reason for using Setll/Reade instead of a Chain is merely from an IO standpoint.

If you know that the record you are searching for definitely exists a Chain is fine.

If however, you are not definitely sure the record exists, the following is what is recommend:

SetLL (KeyFields) FIleName;
If %EQUAL(FIleName);
ReadE (KeyFields FIleName;
DoW Not %EOF(FIleName);
do my stuff
ReadE (KeyFields) FileName;
EndDo;
EndIf;
The true IO impact happens on the ReadE and Chain. The above example will make your process more efficient.






From: Chandra Krieg
Sent: Tue 9/2/2008 8:34 AM
To: 'rpg400-l@xxxxxxxxxxxx'
Subject: Chain vs SetLL/ReadE


Hello all,

I have been told that the preferred method of random record retrieval is to use the SetLL, then do a ReadE and check for the %EOF indicator.

Why is this preferred over using a Chain with %found or even checking for %Equal on the SetLL before reading the file? I have been told that the Chain doesn't do what I'm expecting it to but I've never had issues in the past.

I have read the archive posts and found opinions on the subject but no concrete explenation as to why. As with some of the posts I don't understand why you would want to readE the file if the SetLL didn't find an %equal match or why if you plan to read the file a chain isn't appropriate.

I'm being told to use the following code instead of a Chain.

SetLL (KeyFields) FIleName;
ReadE (KeyFields FIleName;
DoW Not %EOF(FIleName);
do my stuff
ReadE (KeyFields) FileName;
EndDo;

Any explanations would be helpful!

TIA!

Chandra Krieg
i5 Programmer/Analyst
RateWatch
(P) 1.800.348.1831 ext 311
(F) 1.920.568.1403
www.rate-watch.com