Re: easier SQL statements in CL

[DrFranken <midrange@xxxxxxxxxxxx>]

Scott,

    In practice I find a combination of *BASED and *DEFINED is the 
right answer. For an API that returns a list and each element in the 
list is a number of fields I use *DEFINED to build the 'record layout' 
and then use *BASED to roll through the list. Maybe this is what you 
meant, but here is parts of an example from a program I wrote recently:

/* Define the 'record' and needed pointers. */

DCL &Cptr  *PTR                                      /* Pointer for 
&Cinf    */
DCL &Cofs  *INT    4                                 /* Pointer Offset 
&Cinf */
DCL &Cinf  *CHAR 112 STG(*BASED) BASPTR(&Cptr)       /* Cartridge 
Information*/
DCL &CID   *CHAR   6 STG(*DEFINED) DEFVAR(&Cinf 1  ) /* Cartridge 
ID         */
DCL &VID   *CHAR   6 STG(*DEFINED) DEFVAR(&Cinf 7  ) /* Volume 
ID            */

/* Call the actual API Here */

CHGVAR &Cptr %ADDRESS(&RCTG)  /* Point to beginning of cartridge 
information */
CHGVAR &Cofs %OFFSET(&Cptr)                      /* Initialize the 
offset */
CHGVAR &Cofs (&Cofs + &Offs)                     /* Initialize the 
offset */
/* This places &Cinf in the same memory as the first row returned in 
&RCTG */
/* All *DEFINED variables are now automatically updated with 
corresponding values. */

DOFOR &Cnum FROM(1) TO(&NCat) /* Loop through all available tapes */
 CHGVAR %OFFSET(&Cptr) &Cofs

/* Process each entry here using &CID, &VID etc. */

 CHGVAR &Cofs (&Cofs + &LCat)  /* Point to next cartridge */
/* This places &Cinf in the same memory as the NEXT row returned in &RCTG */
/* All *DEFINED variables are now automatically updated with 
corresponding values. */
ENDDO

Larry "DrFranken" Bolhuis


On 5/14/2012 1:26 PM, Scott Klement wrote:
> Hiya Folks,
>
> Consider using STG(*DEFINED) fields instead of *BASED for this type of
> logic.  No need to do pointer math that way, and aside from the pointer
> stuff, they work the same as *BASED.  Save *BASED for times when you
> need to be able to change the address in memory where the variables point.
>
> As for Joel's initial question, a *MSGF/RTVMSG might be an easier
> solution.  But, ultimately, the problem is that CL just isn't that good
> at string manipulation. Using a better language like RPG or REXX would
> almost certainly work better. Why force a square peg through a round hole?
>
> -SK
>
>
> On 5/14/2012 9:49 AM, DrFranken wrote:
> > I'm going to guess no, but in some cases you might be able to simplify
> > by declaring&P_CLIENT,&P_SEQ etc as *BASED fields of a longer
> > character variable thus building the VALUES( clause there.  It could
> > look as simple as VALUES(&VALUES) avoiding all those *cats in your program.