topic Re: selecting observations with value of interest appearing in more than one variable in New SAS User

selecting observations with value of interest appearing in more than one variable

femiajumobi1 — Tue, 18 Oct 2022 16:53:39 GMT

The G codes ('G11','G111','G1110') are potential values in each of three variables/columns: primary, secondary, and tertiary. The 3 columns (primary, secondary, tertiary) in csv file are separate. I want to create a new dataset by selecting observations with these G codes in variables: primary, secondary, and tertiary. The code below with the use "OR" does not work. How do I go about this? Thanks.

Data tricolumns;
set EDPh;
if Primary Diagnosis in ('G11','G111','G1110') OR secondary Diagnosis in ('G11','G111','G1110') OR tertiary Diagnosis in ('G11','G111','G1110') ;

run;

Re: selecting observations with value of interest appearing in more than one variable

Tom — Tue, 18 Oct 2022 16:58:25 GMT

That code would work if you used valid variable names.

So either use variable names that follow the rules (only contain digits letters or undescore and do not start with a digit).

if PrimaryDiagnosis in ('G11','G111','G1110')
 OR secondaryDiagnosis in ('G11','G111','G1110')
 OR tertiaryDiagnosis in ('G11','G111','G1110') 
;

Of if you have the VALIDVARNAME option set to ANY then use a name literal to reference those non-standard names.

if 'Primary Diagnosis'n in ('G11','G111','G1110')
 OR 'secondary Diagnosis'n in ('G11','G111','G1110')
 OR 'tertiary Diagnosis'n in ('G11','G111','G1110') 
;

Re: selecting observations with value of interest appearing in more than one variable

femiajumobi1 — Tue, 18 Oct 2022 17:35:11 GMT

Hi Tom, it worked, thank you.

Re: selecting observations with value of interest appearing in more than one variable

Reeza — Tue, 18 Oct 2022 17:40:42 GMT

If you require the G codes in any of the three diagnoses then the OR is correct.

If you require all three diagnosis to have a G value then the condition should be AND

Re: selecting observations with value of interest appearing in more than one variable

femiajumobi1 — Tue, 18 Oct 2022 18:12:57 GMT

Thanks, Reeza. I require the G codes in any of the three diagnoses.

Re: selecting observations with value of interest appearing in more than one variable

Reeza — Tue, 18 Oct 2022 18:58:56 GMT

You should mark Tom's solution as the correct one, not your own response. You can unselect yours and then select his I believe.