Repeat a value by group

[AmirSari]

Hello,

I would like to have a value repeated until the next line is zero and if the next line is greater than zero then value be added to the previous value . here is a sample of my dataset and my desired output.

My data looks like this

id	date	gwil
1	2014	0
1	2015	0
1	2016	100
1	2017	0
1	2018	0
1	2019	20
1	2020	0
1	2021	0
2	2015	0
2	2016	80
2	2017	0
2	2018	0
2	2019	15
2	2020	0

My desired output will look like this

id	date	gwill	gwill1
1	2014	0	0
1	2015	0	0
1	2016	100	100
1	2017	0	100
1	2018	0	100
1	2019	20	120
1	2020	0	120
1	2021	0	120
2	2015	0	0
2	2016	80	80
2	2017	0	80
2	2018	0	80
2	2019	15	95
2	2020	0	95

I used the following code and was able to repeat until the next value is zero but can't figure out how to add the next value for the rest of the obseravations.

data want;
set have;
by id notsorted;
retain gwlim;
if first.id or gwill > 0 then gwlim = gwill;
rename gwlim = gwill1;
run;

Any help is greatly appreciated!

[mkeintz · Accepted solution]

For each ID, you apparently want GWILL1 to contain the running sum of all values of GWILL:

data want;
  set have;
  by id notsorted;
  gwill1+gwill;
  if first.id then gwill1=gwill;
run;

Untested, in the absence of a working DATA step with the sample data.

And slightly more compact, but a bit more obscure:

data want;
  set have;
  by id notsorted;
  gwill2+ifn(first.id,gwill-gwill2,gwill);
run;

[AmirSari]

Thank you! This is exactly what I wanted.

[japelin]

try this code

proc sort data=have;
  by id date;
run;

data want;
  set have;
  by id date;
  retain gwill1;
  if first.id then gwill1=0;
  gwill1=gwill+gwill1;/* If you think of "add if greater equal to zero" as "always add," then you don't need the if statement. */
run;