Message authorized by: : /S=Abdelmoneim.H.Elnagheeb@AEXP.COM/OU=SMTP/O=1.UCN.GO.1/P=PROCTERGAMBLE/A The following codes would do what you want but still wouldn't avoid an internal sorting. The INDICAT table would be sorted by I.

data ONE; input Y Z X; cards; 3 4 5 ... ; run;

proc sql; create table INDICAT as select A.*, B.I from ONE as A left join (select distinct X, count(X) as I from ONE group by X) as B on A.X=B.X ; quit;

Min Fu fu.m@pg.com

______________________________ Reply Separator _________________________________ Subject: Stratification Index Author: (INTERNET)SAS-L@UGA.CC.UGA.EDU at external Date: 4/30/96 4:29 PM

Hi SAS-Lers: I have sent the message below few days back and hope to get responses this time. Thanks for any help. ====================================message======================== Hello SAS-Lers: I have a large data set with a variable, say X, which may have the same value for different OBSERVATIONS. I want to create an indicator that will have the number of times that X has the same value across observations. Here is a crude example:

Y Z X I ---------------------------------------------------------------- 3 4 5 3 2 7 3 Now I want to create Indicator, I: 2 3 9 5 3 9 6 1 1 6 1 3 2 5 4 5 3

Here 5 is repeated 3 times, 3 repeated 2 times, and 1 once.

Note: the data is big (with many observations) and if there is a way to avoid sorting that will be great. Any help will be appreciated. Thanks in advance.

Abdu