Re: On-break behavior

[John Milardovic <milardj@SX.COM>]

I think I figured it out - if you specify a print position in the on-break
line it writes to the buffer even if nothing is printed.

EX.

This produces a 'tmp' file :
begin-program
begin-select
invoice_no (,1) on-break print=never before=proc
from
invoice_header
order by invoice_no
end-select
end-program

begin-procedure proc
    move &invoice_no to $invoice_no
    new-report $invoice_no
    print 'Test' (+1,1)
    new-report 'tmp'
end-procedure

And this does not :

begin-program
begin-select
invoice_no () on-break print=never before=proc
from
invoice_header
order by invoice_no
end-select
end-program

begin-procedure proc
    move &invoice_no to $invoice_no
    new-report $invoice_no
    print 'Test' (+1,1)
    new-report 'tmp'
end-procedure

John Milardovic


> -----Original Message-----
> From: John Milardovic
> Sent: Tuesday, November 09, 1999 10:26 AM
> To:   'SQRUG'
> Subject:      On-break behavior
>
> Oracle 8
> SQR4.0.2
> AIX and NT
>
> I have a bunch of SQR reports that access a common procedure which
> determines how many instances of the report to produce and what format and
> sort options to use for each instance.  The common procedure creates a
> series of header files then returns control back to the calling program
> which produces the body (report output) files.  The common procedure is
> then called to concatenate the body and header files.
>
> The problem that I have is that in two places I use the command new-report
> 'tmp' to force SQR to flush the buffer and write the contents to a file
> this works great for most of my reports, but two of them (which use the
> same internal logic) create the expected files but also create a 'tmp'
> file which contains only a form-feed (\f) character.  At no time am I
> printing anything, or calling new-page, while the new-report 'tmp' command
> holds sway except for a col1 (,1) on-break print=never line.  This line
> seems to be the only difference between the reports which behave correctly
> and the two that do not.
>
> To make a long story as short as possible I was wondering whether the
> print=never works as I think or whether it might be printing a null or
> space or something to the buffer which would create the 'tmp' file.
>
> Program logic
>       1. Select rows - order by check no
>       2. Before break of 1 do driving routine
>       3. Driving routine calls common procedure to determine number of
> instances
>       4. Output file is now *.head - create header files
>       5. Output file is now 'tmp' - flush buffer
>       6. Set up loop for desired number of instances
>       7. Within loop create report (output file is now *.body)
>       8. Output file is now 'tmp' - flush buffer
>       9. Concatenate report with header file
>       10. Go to 7 if another instance to create else goto 1
>
> Thanks in advance.
> John Milardovic