Be sure *INLR is not on. If *INLR = *On, the select statement will not return any records.

Not true, WoodEngineer

*inlr only affects the “cycle” operations. (most modern programs run in one cycle, just looping around in the calcs until finished.)

I routinely set *inlr = *on as the first calc in my program, and then READ, CHAIN, exec sql, etc. That way, whenever I encounter a situation that says my program is finished, I can just code a RETURN and everything is closed up and ends.

Hello BigKat,

I’m speaking from experience here, as in this week. As soon as I moved the seton LR to occur after the select statement, data was returned. That was the only change I made to the program.

Perhaps you are setting on LR after your select statement.

I was very suprised because I also thought *INLR only effected files accessed in the RPG cycle.

By the say, the program seton LR in the *INZSR subroutine. Maybe that is the cause.

In any case, I’d like to fully understand this.

If you set *INLRon in *INZSR, the system will immediately end the program (due to the cycle) after initialization, and BEFORE any of the calculations (including the SQL) are run. Setting on *INLR in *INZSR is like aborting the program because something failed to initialize.

this is part of a program I am working on now and it works just fine

 * Main Program                                 
c                   eval      *inlr = *on       
c                   dou       #exit             
c                   exsr      openCsr           
c                   exsr      clearSfl          
c                   exsr      loadSfl           
c                   dou       #reload or #exit  
c                   exsr      displaySfl        
c                   enddo                       
c                   exsr      closeCsr          
c                   enddo                       

Hi,

I find it really strange to see the *INLR at the start of the program, but if it works for you, why not? I guess eventually the *INLR will disappear and get replaced with a less cycle-oriented opcode. Maybe there’ll even be a CYCLE or NOCYCLE keyword for the H-SPEC…

Regards,

Martin Gilbert.

It really is just a matter of being able to issue a return anywhere in the program that it is necessary, without having to have eval *INLR= *ON all over the place. But yeah, it is just a matter of coding style. The important thing to this thread, is that it is NOT in *INZSR, as that makes the cycle think you are aborting the program.

Thanks for the clarification, Big Kat.
I should have known that.
It makes a lot of sense.
It also clears up the cause of my problem. Well . . . this one at least.

With 6.1 of IBM i (a.k.a., V6R1 of OS/400), a callable program can be created without the cycle, by using the MAIN keyword and specifying which procedure will be the entry point of the program.
http://www.mcpressonline.com/programming/rpg/v6r1-rpg-enhancements.html
I’d start using, but we’re not at 6.1 yet – hopefully, one day soon.

I also use *INLR at the start of many of my programs. I always comment it to explain why, in case someone not familiar with the effects of that on the cycle reviews the code for some reason.

I also do it so I can use return by itself in the program I suppose I feel warm/fuzzy not having to repetitively type *INLR = *off before every return in my program. ^_^