One must learn to use "PIPE AHELP stage" to see the author's help, and there AUTOADD etc are documented.
"HELP PIPE stage" and "PIPE HELP stage" both produce Endicott's help files, which are less good. Kris, IBM Belgium, VM customer support "Nix, Robert P." <[EMAIL PROTECTED]> Sent by: The IBM z/VM Operating System <[email protected]> 2006-06-28 14:53 Please respond to The IBM z/VM Operating System To [email protected] cc Subject Re: Finding DUPS in a CMS file Fanout your input. Using the lookup stage previously given, you could get a list of the duplicate keys. Feed this list of keys into a second lookup input stream 2 as the keys to look for, and feed the second output stream of the fanout into this lookup as input. Output only the records that match the keys found there. The online help for lookup doesn't show the autoadd or keyonly options, and says that input stream 2 MUST be connected, which I don't see in any of the examples that have been shown. However, assuming that these options exist and are just undocumented, the above pattern should be able to give you a complete list of the non-unique records, in the order they existed in the original file. -- Robert P. Nix Mayo Foundation RO-OC-1-13 200 First Street SW 507-284-0844 Rochester, MN 55905 ----- "In theory, theory and practice are the same, but in practice, theory and practice are different." From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Colin Allinson Sent: Wednesday, June 28, 2006 7:30 AM To: [email protected] Subject: Re: Finding DUPS in a CMS file That is fairly easy. PIPE the records through a UNIQUE stage into a hole. Label the UNIQUE stage and take the secondary output into a STEM (for instance). Then, if the stem has any entries you know there are duplicates (and have some idea what to look for). Regards Colin Allinson "Wakser, David" <[EMAIL PROTECTED]> Wrote: All: I appreciate the answers (and was relatively surprised at how easily this could be accomplished). HOWEVER, I need the capability of possibly merely issuing an error message, instead of "blindly" deleting one of the records, because the entire records are not duplicates, merely the first 20 bytes. And someone would need to look at the input data to see why the data is duplicated. For those that are curious and desire more details, I am putting together a method for our Help Desk to know who is on call at any given time. Considering the varied schedules that I and my co-workers have, this gets sort of complicated, and the Help Desk always seems to get it wrong! In general, one of us is on call for entire week, 24 hours a day. However, since I get up at 4:00 AM I would prefer that my co-workers are not called at that time, since I am up anyway! The identification of duplicates is to ensure that two people have not been assigned duty in the same time period! These would be "exceptions" (or overrides) to the default on-call person. For example, let's assume I am "on call" July 1 thru July 7. However, on July 3rd I have an appointment from 6:00 PM to 7:00 PM. I would set the "default" on-call record for the entire week with my name, and then have specific date/time records for the times that OTHER people would cover for me, which override the "default". I need to ensure that, except for the "default" record, no two people are assigned coverage for the same time period. Records consist of a date, a time (all times assume a thirty-minute interval), and the on-call person's name. Once I have sorted them by date and time, there should NOT be any duplicates. If there are, how can I identify them? I MIGHT possibly want to selectively delete one, but at this point in time I need to IDENTIFY that fact that there are dups and issue a message. I hope this explains things so that your solutions are more helpful! David Wakser InfoCrossing
