Re: CHAIN + DS as result field

["M. Lazarus" <mlazarus@xxxxxxxx>]

Alan,

 Although I usually prefer the explicit key 
field method you mention, I can see the value in 
externalizing the key, the same way that the DS is externalized.

 -mark


At 2/17/09 11:06 AM, you wrote:
> Just curious. Why would you ever want to use the form of
>
> Chain %Kds(WRecordKey) FILE1 wRecord1;
>
> instead of
>
> Chain (Value1:Value2) FILE1 wRecord1;
>
> I don't use file I/O anymore but it puzzles me why you would want to
> use the %Kds when you can just list the fields you want in the key. If
> you use the %Kds then you have to go look up the key values and you
> have to load the key values.
>
> Just never understood the value.
>
> On Tue, Feb 17, 2009 at 8:46 AM, Kurt Anderson
> <kurt.anderson@xxxxxxxxxxxxxx> wrote:
> > Without testing, doesn't the chain need to be like this:
> >       CHAIN %kds(WrecordKey) FILE1 Wrecord1;
> >
> > Also, don't forget, David, that you'll still 
> have to load the key values, and they'll be qualified.
> >
> > e.g.   wRecordKey.key1 = myvalue;
> >
> > -----Original Message-----
> > From: rpg400-l-bounces@xxxxxxxxxxxx 
> [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of M. Lazarus
> > Sent: Tuesday, February 17, 2009 9:13 AM
> > To: RPG programming on the IBM i / System i
> > Subject: RE: CHAIN + DS as result field
> >
> > David,
> >
> >  You're missing the key on the CHAIN!  Assuming
> > that the key is identical for both files, try this:
> >
> > D WrecordKey      DS                  LIKEREC ( RECORD1 : *Key )
> > D Wrecord1        DS                  LIKEREC ( RECORD1 : *INPUT )
> > D Wrecord2        DS                  LIKEREC ( RECORD2 : *INPUT )
> >
> > When condition1
> >         CHAIN WrecordKey FILE1 Wrecord1;
> >       EVAL-CORR Globalrecord= Wrecord1;
> >
> >
> > When condition2
> >         CHAIN WrecordKey FILE2 Wrecord2;
> >       EVAL-CORR Globalrecord= Wrecord2;
> >
> >
> >  -mark
> >
> > At 2/17/09 04:16 AM, you wrote:
> >>Thanks, Alan
> >>
> >>I already had the global definitions as you suggest.
> >>
> >>In my subprocedure, I now have :
> >>
> >>D Wrecord1        DS                  LIKEREC ( RECORD1 : *INPUT )
> >>D Wrecord2        DS                  LIKEREC ( RECORD2 : *INPUT )
> >>
> >>When condition1
> >>         CHAIN FILE1 Wrecord1;
> >>       EVAL-CORR Globalrecord= Wrecord1;
> >>
> >>
> >>When condition2
> >>         CHAIN FILE2 Wrecord2;
> >>       EVAL-CORR Globalrecord= Wrecord2;
> >>
> >>
> >>So now I can put the result into one common DS, thanks again.
> >>
> >>I don't find the compile message very clear. Just to test it, I tried :
> >>         CHAIN FILE1 Wrecord2;
> >>
> >>File1 is a PF with RECORD1, File2 is an LF with RECORD1 renamed RECORD2.
> >>
> >>This is an extract of the compiled listing : the
> >>gravity level of the message is 20 although in
> >>the summary  there is just one message at level
> >>30. If I look at the definitions of Wrecord1 
> and Wrecord2 they are identical.
> >>
> >>*RNF7701 20      1 La structure de données n'est pas admise pour l'opér
> >>* * * * *  F I N   D U   R E C A P I T U L A T I F   D E S   M E S S A
> >>  5722WDS V5R4M0  060210 RN        IBM ILE RPG             D126169V03/CL
> >>R é c a p i t u l a t i f   f i n a l
> >>Nombre total de messages :
> >>Information  (00) . . . . . . . :     4041
> >>Avertissement (10)  . . . . . . :       11
> >>Erreur       (20) . . . . . . . :        0
> >>Erreur grave (30+)  . . . . . . :        1
> >>
> >>
> >>-----Message d'origine-----
> >>De : rpg400-l-bounces@xxxxxxxxxxxx
> >>[mailto:rpg400-l-bounces@xxxxxxxxxxxx] De la part de Alan Campin
> >>Envoyé : lundi 16 février 2009 18:58
> >>À : RPG programming on the IBM i / System i
> >>Objet : Re: CHAIN + DS as result field
> >>
> >>What release level are you at? The new ability
> >>to pass a file to a sub-procedure looks perfect for you. (V6R1).
> >>
> >>The other one I could think of is why not just
> >>use an override? Have one dummy input file overridden to the actual.
> >>
> >>The other one I could think of is that you do
> >>not need a separate data structure. Couldn't you
> >>declare a templete (Type defintion) in the header and use EVAL-CORR?
> >>
> >>d TD_FORMAT...
> >>d                           e  ds                   ExtName(LF1)
> >>d                                                     Based(StdNulPtr)
> >>
> >>d FileData...
> >>d                               ds                  LikeDs(TD_Format)
> >>
> >>   FileData = GetFileData(values);
> >>
> >>In Subprocedure
> >>
> >>d FileData...
> >>d                          ds              LikeDs(TD_Format)
> >>
> >>Chain LFILE1 DS1;
> >>
> >>Eval-Corr FileData = DS1;
> >>
> >>Return FileData;
> >>
> >>This could be especially handy if you don't want all the data.
> >>
> >>Not sure I am really answering your question.
> >>
> >>On Mon, Feb 16, 2009 at 10:09 AM, David FOXWELL
> >><David.FOXWELL@xxxxxxxxx> wrote:
> >> > I don't understand why I can't do this :
> >> >
> >> >
> >> > I have a procedure that selects which LF to
> >> CHAIN to depending upon its input parameters.
> >> All the LF's compiled on the same PF.
> >> >
> >> > Trouble is I seem to need a different DS
> >> receptor for each LF. When I try to compile I
> >> get RNF7595 "Position 1of the DS doesn't
> >> contain a subzone for a record format".
> >> >
> >> > If I look at the compile listing of the DS
> >> definitions DS1 and DS2, they seem to be identical.
> >> >
> >> > D DS1  LIKEREC(LFORMAT1 *INPUT)
> >> > D DS2  LIKEREC(LFORMAT2 *INPUT)
> >> >
> >> > CHAIN LFILE1 DS1;
> >> >
> >> > CHAIN LFILE2 DS2;
> >> >
> >> > How can I get the result of the CHAIN into one common DS?
> >> >
> >> > Thanks.