I was fairly certain there was a way to delay EOF from taking effect.  In
essence, once EOF is hit, wait nnn seconds for the potential for more
records to be added to the file.  If nnn seconds pass without any new
records added to the file, EOF is processed.

I looked at CRTPF.  Specifically the WAITFILE & WAITRCD parameters.  They
didn't look right.

Then I looked at OVRDBF and found the EOFDLY parameter.  A maximum of 99999
seconds (over 27 hours!) is allowed.  Have you tried this one?


> -----Original Message-----
> From: midrange-l-bounces@xxxxxxxxxxxx / Fiona Fitzgerald
> Sent: Tuesday, October 05, 2004 11:32 AM
> I have a CL which creates a file in QTEMP, populates it,
> then reads it in a loop. So far so good.
> Then it delays x seconds & repeats the process.
> This time, no processing occurs within the loop,
> because although the file is repopulated,
> the pointer is stuck at end-of-file at this stage.
> As you might guess, this was not the plan!
> I considered using OPNDBF with POSDBF.
> But from what I've read, RCVF won't share its ODP,
> so that's out.
> Do I have to resort to generating a new QTEMP file,
> suppressing level check, overriding to it
> & calling an RPG just to get the pgm to start again from the top
> every time I repopulate the file?
> I don't want to use RPG because most of the functions
> performed are CL functions, perfect for a CL pgm.
> Or should I delete & recreate the file each time - will that
> reposition it?
> All suggestions gratefully received, hope someone knows the theory
> better than I do : )
> Thanks,
> Fiona

This thread ...


Return to Archive home page | Return to MIDRANGE.COM home page