First.Variable Last.Variable to sum multiple observations

[mariamabotaleb]

Hello, 

I am trying to sum multiple observations with the First.Variable Last.Variable feature, however it never works in adding all the observations together, it either gives me the first or last variable values  

Data T_Total;
set T;
by m**; 
if first. m** then TOTAL_DAYS=0;
TOTAL_DAYS= sum (TOTAL_DAYS,SUPPLY_DAYS);
retain TOTAL_DAYS;
if last. m** then output;
run;

I also tried changing the order of sentences, or using the 

TOTAL_DAYS= TOTAL_DAYS+SUPPLY_DAYS;

[PaigeMiller]

There's no double-asterisk in this syntax.

And instead of TOTAL_DAYS=sum (TOTAL_DAYS,SUPPLY_DAYS); you probably want

total_days+supply_days;

Better yet, don't program this yourself in a DATA step. Use PROC SUMMARY, where this is a built-in feature.

[Astounding]

When you use first. and last. variables, there should not be a space after the dot.  Incorrect:

first. m

Correct:

first.m

And as was noted, asterisks are not a legitimate part of a variable name.  Clean up the obvious inaccuracies, and see if any problems remain.  Your syntax for summing should work, even if other choices exist.

[Tom]

Make sure that TOTAL_DAYS does not already exist as a variable in the input dataset T.  Because if it does then as you read observations from T the value you calculated after reading the pervious observation will be overwritten with what is read from T.