Forward or backward fill missing values

[tomcmacdonald]

I'm trying to fill missing values in a sorted set.  So if the previous observation is 1 and the next is NULL then it should be 1.  Basically I'm trying to replicated the functionality of pandas.DataFrame.fillna(method='ffill').  

[novinosrin · Accepted solution]

here is a demo:

data have;
do n=1 to 10;
if n in (1,4,7) then k=1;
else k=.;
output;
end;
run;

data want;
set have;
retain _k;
if not missing(k) then _k=k;
else if missing(k) then k=_k;
drop _k;
run;