Hi @jaiganesh
The Lag function returns the value from a previous input record. Lag1 (or just Lag) returns the value from the record preceding the current record, Lag2 from the one before that and so on. There is no limit on the number of Lags other than the available memory, so if you have 1000 records, you can use Lag1 - Lag999.
But you cannot use Lag on variables created in the data step, so you cannot accumulate values using Lag. Here is an example using Lag on your data to give the wanted result, so you can see what happens:
data kk;
input id names$ marks;
cards;
1 jai 10
2 sai 20
3 ram 30
4 shyam 40
5 hari 50
;
run;
data want; set kk;
lag1marks = lag1(marks);
lag2marks = lag2(marks);
lag3marks = lag3(marks);
lag4marks = lag4(marks);
New_marks = sum(marks,lag1marks,lag2marks,lag3marks,lag4marks);
run;
This works with your 5 observations, because the code uses Lag1 - Lag4, but it cannot be made dynamic, at least not without a lot of macro coding. If there were more observations, it would give a rolling sum over the last 5 observarions, which might be useful in some cases, but to use Lag to make a total summation over a column in a data set is - politely expressed - not advisable.
... View more
