Date:         Wed, 24 Jul 2002 09:03:54 -0700
Reply-To:     Stig Eide <stigeide@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Stig Eide <stigeide@YAHOO.COM>
Organization: http://groups.google.com/
Subject:      Re: how to get the mean of three month for each category
Content-Type: text/plain; charset=ISO-8859-1

If date is a SAS date value, you could sort the table
by category and date. Then you can use the lag function.
Here is some untested code:
*If the date variable is character, you should convert it
 to SAS date value first;
proc sort data=dt;
  by category date;
run;
data dt(drop=count);
  set dt;
  by category;
  retain count 0;
  if first.category then count=0;
  else count+1;
  if count > 2 then AVERAGE=mean(lag1(return),lag2(return),lag3(return));
run;

Good luck!
Stig Eide

gdx@CATS.UCSC.EDU (dongxu) wrote in message news:<5.1.0.14.0.20020724012924.00b9cd58@cats.ucsc.edu>...
> I have the following data set.How can i get the average of latest three
> month. for example: in apr78, i want to get the average return of
> jan78, feb78, mar78 for each category. in may78, i want to get the average
> return of feb78, mar78,apr78 for each category. And so on.
> Thank you for your reply.
>
>                                 Obs    date     category     return
>
>                                   1    JAN78      WLG       -0.06970
>                                   2    JAN78      WLV       -0.06270
> <snip>
>                                  46    AUG78      WMV        0.06320
>                                  47    AUG78      WSG        0.08700
>                                  48    AUG78      WSV        0.06230