Use a variable's value as part of another variable's name

[Solph]

I've a dataset with a variable death_year for year of death (2017 or before), and a bunch service use year variables (1 or 0 as yes or no) between 2010 and 2018 (service2010, service2011, service2012, .... service2018). I want to identify # of service use two years after death year to validate the value of death year.

num_use_post_death=sum(of serviceXXXX - service2017);
where XXXX is the individual person's value of death_year +2.

Thanks.

[Kurt_Bremser]

Please post example data (in a data step with datalines), and what you expect out of it.

[gamotte]

Hello,

It would be easier to answer if you post have and want datasets in the form of datasteps.

data want;
    set have;
    array services(*) service2010-service2018;

    startindex=death_year-2009;
    stopindex=dim(services);

    do i=startindex to stopindex;
        num_use_post_death=sum(num_use_post_death,services(i));
    end;
run;

[PGStats]

Like @gamotte, but specifying explicit array bounds:

data want;
set have;
array services{2010:2018} service2010-service2018;
num_use_post_death = 0;
do year = death_year+2 to hbound(services);
    num_use_post_death + services{year};
    end;
drop year;
run;

(untested)

[s_lassen]

Something like this?

data want;
  set have;
  array services(2010:2018) service2010-service2018;
  num_use_post_death=0;
  if death_year then do _N_=max(death_year+2,2010) to hbound(services);
    num_use_post_death+services(_N_);
    end;
run;

I put in the MAX function in case you have some that died before 2008.