What about putting a sequence number before the keys before sending them to
LOOKUP?

2009/1/8 Schuh, Richard <[email protected]>

> John,
>
> Not quite. I need something more like
>
> 1. Duplicate masters are not discarded.
> 2. Whenever a detail matches a master, one of the matching master
> records is deleted from the reference.
>
> That would result in a reference that contained unmatched instances of
> each master. For example, if there were 5 instances of a given key in
> the reference and only 3 instances of that key in the detail, then there
> would be two equal but unmatched references that would go to the
> unmatched master stream. If it were the other way around with 3
> duplicate masters and 5 instances in the detail stream, then there would
> be three matched and two unmatched details.
>
>
> Regards,
> Richard Schuh
>
>
>
> > -----Original Message-----
> > From: CMSTSO Pipelines Discussion List
> > [mailto:[email protected]] On Behalf Of John P. Hartmann
> > Sent: Thursday, January 08, 2009 2:23 AM
> > To: [email protected]
> > Subject: Re: Lookup
> >
> > Richard.
> >
> > Doesen't the tertiary output from Lookup contain what you desire?
> > [Master records that are not matched by a detail.]
> >
> >    j.
> >
> > 2009/1/8 Schuh, Richard <[email protected]>:
> > > I am trying to find a pipe solution to the following problem:
> > >
> > > I have 2 files that have tags in them. I want to identify
> > which tags
> > > are present in one of the files but not the other. A lookup
> > would fill
> > > the bill nicely, except for the fact that there may be more
> > than one
> > > occurrence of any given tag in either of the files. I want to do
> > > something akin to a lookup allmasters and delete one
> > occurrence from
> > > the reference for each detail that matches. After all
> > occurrences of a
> > > tag have been deleted from the reference, any remaining tags in the
> > > detail would be considered to be unmatched. Similarly, if not all
> > > masters were matched and deleted, those that remain would
> > be deemed unmatched.
> > >
> > > Is this something that can be done reasonably well. My brain has
> > > fogged over (too many meetings, I think) so I have not seen
> > the simple
> > > solution that probably exists.
> > >
> > > Regards,
> > > Richard Schuh
> > >
> >
>



--
Kris Buelens,
IBM Belgium, VM customer support

Reply via email to