using index to create a new variable

[epidemiologystu]

Hello,

This piece of code is not correctly assigning all breast cancer cases. Could someone identify the error? 

*create sample dataset;

data sample ;

  input curr_topog_desc $30. ;

  cards ;

Rectum NOS

Overlapping lesion of breast

Head of pancreas

Blood

Breast NOS

" " *meant to be blank;

;

run; 

Data sample;

Set sample;

IF index(curr_topog_desc, "breast") ne 0 then do;

breastcancer = 1;

othercancer = 0;

end;

*if blank space, the person was not linked to cancer registry;

IF curr_topog_desc = " " THEN do;

breastcancer = 0;

othercancer = 0;

end;

*meant to capture all other cancers;

ELSE do;

breastcancer = 0;

othercancer = 1;

end;

run;

[FreelanceReinh · Accepted solution]

Hello @epidemiologystu,

There are two issues:

1. The INDEX function is case-sensitive so that "breast" isn't found in "Breast NOS." Use the FIND function with the 'i' modifier instead:
   

   find(curr_topog_desc, "breast", 'i')

   or apply the LOWCASE function to the first argument:
   

   index(lowcase(curr_topog_desc), "breast")

2. Without an ELSE before the second IF the values assigned in the first THEN branch are overwritten by the last ELSE DO/END block.

---

@epidemiologystu wrote:

data sample ;

  input curr_topog_desc $30. ;

  cards ;

Rectum NOS

Overlapping lesion of breast

Head of pancreas

Blood

Breast NOS

" " *meant to be blank;

---

Use a single period (.) to enter a missing value in data lines, not a blank in double quotes.

[epidemiologystu]

Thank you!!