Date:         Fri, 2 Feb 1996 09:48:21 -0500
Reply-To:     Jay Goode <JGOODE@QUINTILES.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Jay Goode <JGOODE@QUINTILES.COM>
Subject:      Fwd: Error in delivering message.

Unknown user at destination host: darnold@qrtp
 
 
 
Returned message follows:
 
 
 
Error #101 has occurred attempting to deliver a message to
 
darnold@qrtp
 
 
 
Unknown user at this host, or no path to specified workgroup.
 
 
 
Original Message Follows:
 
=========================
 
SMF-70
 
240Send-to: darnold@qrtp
 
To: mail@ih (Multiple recipients of list SAS-L) {SAS-L@uga.cc.uga.edu}
 
202C2-ID: 886C113181C60200
 
200MCB-options: NNYOANA
 
Date-posted: 1-Feb-96 19:21:33
 
220Hop-count: 1
 
200Message-id: 896C113101C60200
 
200O-C2-Routed-By: IH; received from -Local-; at 1-Feb-96 19:21:33
 
X-smtp-file: qfa07783
 
Date-received: 01-Feb-96 19:20:39 +0500
 
X-received: uga.cc.uga.edu [128.192.1.5]
 
Sender: MAIL@IH {owner-sas-l@UGA.CC.UGA.EDU}
 
Date: 01-Feb-96 14:43:22 -0700
 
Subject: Re: Stumped with Duplicate Records
 
Importance: LOW
 
From: mail@ih (John Beeman) {John_Beeman@NBS.GOV}
 
Reply-to: mail@ih (John Beeman) {John_Beeman@NBS.GOV}
 
200MHS-id: 886C113181C60200
 
200Via-host: IH.IH
 
200O-C2-Serial-Host: IH
 
200O-C2-Serial-No: IAANAAACMG
 
 
 
 
 
     newlin@pcmail.schev.edu asked how to get rid of some duplicate
records
 
     in a student database.  The goal is to omit observations with
 
     undeclared majors and then keep only those social security numbers
 
     that have the largest number of credits when there are duplicate ssn
 
     observations.
 
 
 
     I think the following code will meet your needs.  First, keep only
 
     those with declared majors (you already did), then sort by ssn and
the
 
     number of credits, and sort again by ssn with the nodupkey option.
 
     This will only keep the first observation of ssn , which will be the
 
     one with the largest number of credits.  Be sure to sort the credits
 
     variable DESCENDING, because the NODUPKEY option of PROC SORT only
 
     writes the first duplicate observation to the output dataset.
 
 
 
     Data tryme; set students ;
 
        where major not in ('und') ;
 
 
 
     proc sort ;
 
        by ssn descending credits ;
 
 
 
     proc sort nodupkey ;
 
       by ssn ;
 
     run ;