Re: Passing a Null Date to a Procedure

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

To me, that's more of a question of "whose job is it", and less of a 
question of performance. Determine which routine SHOULD be responsible 
for supplying the date...  should the caller even know or care that 
there's a date involved?  That sort of question is how I'd decide which 
routine retrieves the record.

Sorry, I don't know enough of your environment and application to give 
you a more direct answer.

But... options(*NULLIND) does let you pass the null indicator. :)


Doug Palme wrote:
> Hey Scott,
> A thought just hit me, what if I passed in the keys for the data in the 
> original file, did a chain and then tested for the *NULLIND locally in the 
> procedure? 
>
> The only difference is I am adding some overhead for the file lookup local 
> to the procedure.
>
> Thoughts?
>
>
>
>
> From:   Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>
> To:     RPG programming on the IBM i / System i <rpg400-l@xxxxxxxxxxxx>
> Date:   12/31/2009 12:36 PM
> Subject:        Re: Passing a Null Date to a Procedure
> Sent by:        rpg400-l-bounces@xxxxxxxxxxxx
>
>
>
> Hi Doug,
>
> You need to declare options(*NULLIND) on the prototype.  When you do 
> that, RPG will automatically pass the null indicator with the parameter.
>
> However, bear in mind that it's not possible to declare a field on 
> D-spec with a null indicator.  That means parameters passed to this 
> subprocedure have to be externally defined variables coming from 
> null-capable PFs declared on an F-spec.   That severely limits the 
> reusability of your subprocedure.  Personally I would give that a lot of 
> thought before using options(*NULLIIND)
>
>
>
> Doug Palme wrote:
> > Is there any way to define a date value in a procedure interface as 
> being 
> > null capable? 
> >
> > I have checked the reference manual and so far have come up empty handed 
>
> > on this.
> >
> > We have a file that allows for null capable dates and I have to pass 
> this 
> > value to a procedure for some processing, but it does not like the usage 
>
> > of the %NULLIND parm.
> >
> >  *RNF0342 20      3 The parameter for %NULLIND is not valid; %NULLIND is 
>
> > ignored
> >
> > *RNF0342 20    259 008812  Parameter WENDDATE for %NULLIND is not null 
> > capable
> >                            %NULLIND is ignored. 
> >
> > Suggestions?
> >
> > Douglas