data cleaning

[Amalik]

Hi,

I want to delete those observations from my dataset which are greater than or less than 50% from the previous observation. e.g.given the following dataset, I would like to drop the second and the last observation.Any suggestions on how to do this?

date	returns
1.1.2000	1%
1.2.2000	51%
1.3.2000	1.20%
1.4.2000	3%
1.5.2000	-50%

[Kurt_Bremser]

Use the lag() function:

data want;
set have;
if abs(lag(returns) - returns) >= .5 then delete;
run;

Note I used 0.5 because I assumed you use the standard percent format, where 1 = 100%.