You can use SQL and do a self-join to generate both leads and lags.

proc sql;

create table want as select distinct have.*, sum(case when have.year-1=cross.year then cross.x1 else . end) as lag1x1, sum(case when have.year-2=cross.year then cross.x1 else . end) as lag2x1, sum(case when have.year-3=cross.year then cross.x1 else . end) as lag3x1, sum(case when have.year+1=cross.year then cross.x1 else . end) as lead1x1, sum(case when have.year+2=cross.year then cross.x1 else . end) as lead2x1, sum(case when have.year+3=cross.year then cross.x1 else . end) as lead3x1, sum(case when have.year-1=cross.year then cross.x2 else . end) as lag1x2, sum(case when have.year-2=cross.year then cross.x2 else . end) as lag2x2, sum(case when have.year-3=cross.year then cross.x2 else . end) as lag3x2, sum(case when have.year+1=cross.year then cross.x2 else . end) as lead1x2, sum(case when have.year+2=cross.year then cross.x2 else . end) as lead2x2, sum(case when have.year+3=cross.year then cross.x2 else . end) as lead3x2 from have inner join have as cross on have.id=cross.id group by have.id, have.year ;

quit;

On Tue, 12 Apr 2005 04:04:13 -0400, Thomas <tythong@YAHOO.COM> wrote:

>Hi all, > >I know the "lag" function is available to lag the variable for N-period. >But, I have a problem on how to take the "forward/leading" of the variable >using SAS. I hereby illustrate my problem with the follwoing example. I >would very much appreciate it if you could kindly advise me how to achieve >it (A generic case). > >Thank you very much! >Thomas > >*Input; >ID YEAR X1 X2 >1 1990 0.3 10 >1 1991 0.45 15 >1 1992 0.4 14.5 >1 1993 0.25 15 >1 1994 0.5 20 >3 1999 1.1 50 >3 2000 1.3 100 >3 2001 2.45 150 >3 2002 3.4 145 >3 2003 2.5 75 >3 2004 4.5 90 > >*Expected output (lag & forward of X1); >ID YEAR X1 X2 LAG1X1 LAG2X1 LAG3X1 LEAD1X1 LEAD2X1 LEAD3X1 >1 1990 0.3 10 . . . 0.45 0.4 0.25 >1 1991 0.45 15 0.3 . . 0.4 0.25 0.5 >1 1992 0.4 14.5 0.45 0.3 . 0.25 0.5 . >1 1993 0.25 15 0.4 0.45 0.3 0.5 . . >1 1994 0.5 20 0.25 0.4 0.45 . . . >3 1999 1.1 50 . . . 1.3 2.45 3.4 >3 2000 1.3 100 1.1 . . 2.45 3.4 2.5 >3 2001 2.45 150 1.3 1.1 . 3.4 2.5 4.5 >3 2002 3.4 145 2.45 1.3 1.1 2.5 4.5 . >3 2003 2.5 75 3.4 2.45 1.3 4.5 . . >3 2004 4.5 90 2.5 3.4 2.45 . . . > >/*The output only show for the "lag and lead" of one variable. may I know >how to generalize to N variable case and the N "lag and lead"?*/