Re: CHAIN not found and subsequent READ

Just for fun I looked back to a really early manual - V2R1 - for RPG/400 - here is the same paragraph -

When the CHAIN operation is successful, the file specified in factor 2 is positioned such that a subsequent read operation retrieves the next sequential record following the retrieved record. When the CHAIN operation does not complete successfully (for example, an error occurs or no record is found), the file specified in factor 2 must be repositioned (for example, by a CHAIN or SETLL operation) before a subsequent read operation can be done on that file.

Now as John suggests, maybe the behavior has changed - I hope not!!

To me, the wording suggests that if one tries a read operation after a failed chain, one should get an error.

I do admit that this is a situation that would rarely be done this way, but there are sometimes reasons to understand what COULD happen even when the craziest programmer is at work!!

Regards
Vern

On 12/18/2015 10:33 AM, Kurt Anderson wrote:

It's totally ok to disagree with my take. ;)

However, I disagree with this statement:
"The docs are absolutely clear and unequivocal, and the behavior clearly does not match the docs."

IMO it is not clear. The SetLL documentation is very clear about what happens when no record is found. The Chain documentation is not. It talks about reading a record after a chain. Now it's in example territory. Who cares about performing a Read after a Chain, I just want to know what happens in regard to the file positioning after a failed Chain. The documentation does not explicitly state what happens as it does with SetLL.

I say the documentation is in "Example" territory because what happens with a failed Chain is completely unrelated to any statement that comes after the Chain. It seems like the documentation is trying to be helpful in saying that if you Chain first, and then Read, and your Chain failed, you'll need to position the file before the read. When are you reading after a Chain? I suppose if you code your read loops to start with a Chain, then read at the end of the loop, that's a time you would do that. Personally I'm not a fan of that method and prefer to SetLL prior to a read (generically speaking).

Given what Vern is seeing, and that the documentation isn't stating explicitly what happens on a failed Chain, I extrapolated what does happen on a failed Chain. I could be wrong, but given the information presented I'm pretty confident in my guess.

Kurt Anderson
Sr. Programmer/Analyst - Application Development, Service Delivery Platform

-----Original Message-----
From: RPG400-L [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of John Yeung
Sent: Friday, December 18, 2015 10:12 AM
To: RPG programming on the IBM i (AS/400 and iSeries) <rpg400-l@xxxxxxxxxxxx>
Subject: Re: CHAIN not found and subsequent READ - docs wrong?

On Fri, Dec 18, 2015 at 9:36 AM, Vernon Hamberg <vhamberg@xxxxxxxxxxxxxxx> wrote:

So the docs are at least confusing, right?

Actually, I think the docs are just plain wrong, now.

I disagree with Kurt, because the manual in fact very carefully and explicitly goes over both the successful and unsuccessful cases, and spells them out very clearly, even mentioning specifically that "no record found" is in the "not completed successfully" case. I mean, it's a model of Chuck Pencian thoroughness and precision.

The docs are absolutely clear and unequivocal, and the behavior clearly does not match the docs. (I tested it just now, and was surprised by the result.)

The reason I use the word "now" when I say the docs are wrong is that I could have sworn, back in the day, that I encountered the behavior that the docs are describing. Namely, immediately following an unsuccessful chain, the next read was unsuccessful.

Either the behavior was changed sometime between V5R2 and 7.1, or I am just imagining that I encountered the doc-described behavior. Which is entirely possible - I have been a rule-follower for most of my life, as well as someone who reads all available instructions before doing anything. With the docs so utterly, unmistakably clear, I may well have just taken it as a given that the read would fail in that situation, and simply never allowed it to happen in my code.

John Y.
--
This is the RPG programming on the IBM i (AS/400 and iSeries) (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.