Date:         Fri, 20 Jun 1997 10:45:45 -0400
Reply-To:     Michael Stuart <mstuart@PRUDENTIAL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Michael Stuart <mstuart@PRUDENTIAL.COM>
Subject:      Getting a List of Distinct Values from an Array
Content-Type: multipart/mixed;
              Boundary="0__=oWgDKSmCCpPknM9omUpgwFrUsefLrDIHenAiKrczVeA6W3iFvmRafrNO"

The following solution was suggested by KLASSEN, and is pasted below.
 
       DO OVER COVTYP;
         V1 = COVTYP;
         V2 = COVUNT;
         V3 = COVPLN;
         V4 = COVTRM;
         V5 = COVBND;
         OUTPUT;
       END;
       DO OVER FNDNUM;
         W1 = FNDNUM;
         W2 =FNDVAL;
         W3 =FNDUNT;
         W4 =UNTVAL;
         W5 =FNDTYP;
         OUTPUT ;
       END;
I broke this into two separate datasteps, and retained only the variables
within each loop.  I then ran frequencies on these.
 
Many thanks to him and the others that responded.
 
---------------------- Forwarded by Michael Stuart/M&P/Prudential on
06/20/97 10:42 AM ---------------------------
Michael Stuart    Thursday June 19, 1997  03:15 PM
IIG Market Research    (201) 802-4125    Fax Number: (201) 622-3156
(Embedded image moved to file: PIC13038.PCX)
To:   SAS-L@UGA.CC.UGA.EDU
cc:
Subject:  Getting a List of Distinct Values from an Array
 
I've read data into SAS from one of our legacy systems.  The dataset had
multiplying occuring segments, and I used the following arrays to read in
these sections:
 
* SET UP ARRAYS FOR COVERAGE FIELDS ;
 
ARRAY COVTYP(20) $6 COVTYP01 - COVTYP20 ;
ARRAY COVUNT(20)    COVUNT01 - COVUNT20 ;
ARRAY COVPLN(20) $6 COVPLN01 - COVPLN20 ;
ARRAY COVTRM(20)    COVTRM01 - COVTRM20 ;
ARRAY COVBND(20) $1 COVBND01 - COVBND20 ;
 
* SET UP ARRAYS FOR FUND FIELDS ;
 
ARRAY FNDNUM(18)    FNDNUM01 - FNDNUM18 ;
ARRAY FNDVAL(18)    FNDVAL01 - FNDVAL18 ;
ARRAY FNDUNT(18)    FNDUNT01 - FNDUNT18 ;
ARRAY UNTVAL(18)    UNTVAL01 - UNTVAL18 ;
ARRAY FNDTYP(18) $1 FNDTYP01 - FNDTYP18 ;
 
 
 
So now I've got either 20 or 18 occurences of each of the above fields, on
every SAS record.  I want to get a list of the distinct values in each
array, without having to run frequencies on each field separately.  Is
there a 'compact' way I can code this?  I started to code a looping macro
that would rename and append to a new SAS dataset, but got a little bogged
down.  Anyone have creative ideas?
 
I would appreciate a personal response to mstuart@prudential.com, in
addition to the list, as my digests seem to arrive sporadically.
 
 
 
Much thanks in advance.
 
 
 
 
 
 
PIC13038.PCX [application/octet-stream]