INDEX/counting

[stancemcgraw]

Trying to code data so I can count the number of times the phrase "Medicare" appears in the insurance type listed. I used the index only to see that it counts where in the phrase it occurs for example, Secondary Medicare, was counted as 11. Is there any way to eliminate the characters that appear before Medicare so I can create a variable that properly counts the number of patients with any type of Medicare insurance?

This was the original code I used:

data test;

set insurance;

Medicare = index (Insurancetype,'Medicare');

run;

data have:

Studyid     Insurancetype                                        Medicare

3548           Medicare                                                   1

6789            Managed Medicare                                  9

4567             Auto Generic/Secondary Medicare        25

data want:

Studyid     Insurancetype                                        Medicare

3548           Medicare                                                   1

6789            Managed Medicare                                  1

4567             Auto Generic/Secondary Medicare         1

[ballardw · Accepted solution]

Try 

data test;
   set insurance;
   Medicare = (index (Insurancetype,'Medicare')>0);
run;

When SAS does a comparison such as >0 then a true result will return 1 and false 0.

Note that I pasted the code into a code box opened using the forum's {I} icon. The code box will maintain format of text from the program editor or Log window while the forum main message windows will reformat the text, sometimes quite drastically.

[PeterClemmensen]

data have;
input Studyid Insurancetype :$50.;
infile datalines dlm=',';
datalines;
3548,Medicare
6789,Managed Medicare
4567,Auto Generic/Secondary Medicare
;

data want(drop=i);
   set have;
   Medicare=0;
   do i=1 to countw(Insurancetype);;
      if upcase(scan(Insurancetype, i, ' /'))=upcase('Medicare') then Medicare+1;
   end;
run;

[stancemcgraw]

That also works! Thanks!