Re: Using CHAIN instead of SETLL AND READE to begin a loop which reads a group

[Gary Guthrie <garyguthrie@xxxxxxxxxxx>]

Yes. Don't count on a keyed access path to return things in the same
sequence except for the keyed portion of course!

I know of one location where V5R1 gave them much grief because their
applications were written with the presumption that within a key,
duplicates would be retrieved in arrival sequence. Well, that worked
great until they upgraded to V5R1.

Even a simple, single PTF could change the behavior.

Bottom line - if you need a particular sequence, provide the key fields
that will yield that sequence.

Gary Guthrie
Senior Technical Editor, iSeries NEWS



Jim Langston wrote:
>
> In my experience it has been FIFO, but you have to realize what
> "unpredictable" means, same thing as "undefined".  It means that in V5R2 it
> may be the last record if it makes the compiler easier to make.  Means
> whatever it returns now doesn't' necessarily mean it will return the same
> thing tomorrow, so don't count on it.
>
> Regards,
>
> Jim Langston
>
> -----Original Message-----
> From: Frank.Kolmann@revlon.com [mailto:Frank.Kolmann@revlon.com]
> Sent: Sunday, May 12, 2002 8:55 PM
> To: rpg400-l@midrange.com
> Subject: Using CHAIN instead of SETLL AND READE to begin a loop which
> reads a group
>
> Looks to me like the issue here is
> 'What does CHAIN retreive when used on a file with duplicate keys.'
> Long time ago I was told by IBM that the record retreived by the
> CHAIN is unpredictable.
> Experience tells that the retreival is FIFO.
> Has anyone noticed anything other than the above.
>
> Frank Kolmann
>
> >To: rpg400-l@midrange.com
> >From: "Mark Adkins" <adkinsm@amerch.com>
> >
> >I can't find any documentation that addresses this.
> >
> >If there is a table of order detail which has an index on Order #, and
> UPC.
> >Will these two pieces of code execute identically.
> >
> >1.   MYORD#    CHAIN     ORDDTL    99
> >     *IN99          DOWEQ*OFF
> >               <SOMECODE>
> >     MYORD#    READE     ORDDTL          99
> >               ENDDO
> >
> >2.   MYORD#    SETLL     ORDDTL
> >     MYORD#    READE     ORDDTL          99
> >     *IN99          DOWEQ*OFF
> >               <SOMECODE>
> >     MYORD#    READE     ORDDTL          99
> >               ENDDO
> >
> >What if the index is not unique?  What if I set my pointer past the end of
> >file before execution?
> >
> >Human logic dictates that a CHAIN should return the first occurrence of
> the
> >index which would be the smallest upc.  Computer implementation dictates
> >that the CHAIN should return the first occurrence of MYORD# found via an
> >algorithm.
> >
> >Mark
>
> _______________________________________________
> This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
> To post a message email: RPG400-L@midrange.com
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/cgi-bin/listinfo/rpg400-l
> or email: RPG400-L-request@midrange.com
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/rpg400-l.
> _______________________________________________
> This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
> To post a message email: RPG400-L@midrange.com
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/cgi-bin/listinfo/rpg400-l
> or email: RPG400-L-request@midrange.com
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/rpg400-l.