NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/bit/listserv/sasl/4002

Text File | 1992-09-02 | 1.9 KB | 36 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!paladin.american.edu!auvm!CPMAIL-NZ.CIS.COLUMBIA.EDU!PASSRAY
X-Mailer:     Pegasus Mail v2.3 (R3).
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.sas-l
Date:         Wed, 2 Sep 1992 18:45:32 EDT
Reply-To:     Ray Pass <[e-mail]>
Sender:       "SAS(r) Discussion" <[e-mail]>
From:         Ray Pass <[e-mail]>
Subject:      Finding non-dups via SQL
Lines: 23

OK Howard, Kernon, Andy, et al.  Here's an easy one for you.  I can do it in
non-SQL code, and I could figure it out in SQL, but this time I need a quick
answer.  I've got two files;  I preach and always use meaningful names, but
for now lets call them FILE1 and FILE2.  They've both got a variable CASENO
and 4 other vars (ADMITDT DISDATE MRN HOSPITAL).  They've both got about
50,000 obs, give or take, but there are some obs in FILE1 that aren't in FILE2
and vice versa.  I need two listings, one showing the cases in FILE1 that
aren't in FILE2, and the other, obviously showing those in FILE2 that aren't
in FILE1.  I don't want to see the dups.  The only var that I care about in
matching is CASENO.  In other words I only want to match on the key CASENO.
I can plow through it by creating a few tables in SQL and then a few
subqueries, but is there a neat way with JOINs or UNIONs or some such SQL
esoterica.  I await your collective wisdom.

Ray


 *--------------------------------------------------------------------------*
 | Ray Pass, Ph.D.                                    Voice: [phone] |
 | Columbia-Presbyterian Medical Center                 Fax: [phone] |
 | 161 Ft. Washington Ave.       AP1310                                     |
 | New York, NY 10032-3784               [e-mail] |
 *--------------------------------------------------------------------------*