Re: Update primary filre

[Robert Nesiba <bob@xxxxxxxxxxx>]

That's another good idea which we're considering.
As you say, simply output the key fields & the field to be updated
to a a temporary file, that when the main program completes,
simply update from the temporary file back into the primary.

Thanks.
No one has disputed the fact that match record processing simply does not
allow updating back into the primary file at total time, in a simple 
straight-forward manner.   (Perhaps with InfDS or look-ahead fields 
there is a way - but not so easy.)


Booth Martin wrote:
> Just a small comment:  This is already a bad enough program, right?  The 
> less you add to it, the better.  Would you consider  just adding a work 
> file and input the data into it, and then after this monster runs, runs 
> another single-purpose program to update the fields based on the data 
> captured by the monster?
>
> . 
>
> Robert Nesiba wrote:
>   
> > Probably the infDS, in lieu of look-ahead, might work
> > but  our program has a bunch of total time L1 calculations,
> > and I don't want to make any major changes to the overall
> > program logic. 
> >
> > We're probably going to have to leave our primary file as input,
> > and then add another file spec for update, either updating a new logical 
> > file
> > or using OVRDBF to point to the same file.   
> > Any reason that won't work?
> > Thanks.
> >   
> >     
>
>