topic Re: Sum Function in SAS Programming

Sum Function

_maldini_ — Sat, 20 Aug 2022 01:29:22 GMT

I am trying to calculate a sum of several variables using the SUM function. I get different totals when I use a list versus specifying each individual variable in the paraenthesis. What's going on here?

/* 	8. Female, Black, Untreated, Non-smoker, Non-diabetic */
/***********************************************************/
	else if 	
	gender = 2 and
	race_all = 4 and 
	htn_meds = 2 and
	smk_status ne 1 and
	ever_told_dm2 ne 1
	
	then do; 
	ascvd_age = log(age_yrs)*(17.114);
	ascvd_tchol = log(tchol_mg)*0.940;
	ascvd_hdl = log(hdl_mg)*(-18.920);
	
/* 	Interaction - age and HDL */
	ascvd_age_x_hdl = log(age_yrs)*log(hdl_mg)*(4.475);
	
/* 	Untreated SBP */
	ascvd_sbp_notx = log(mean_sbp)*27.820;
	
/* 	Interaction - age and UNTREATED SBP */
	ascvd_age_x_sbp_notx = log(age_yrs)*log(mean_sbp)*(-6.087);
	
/* 	Non-smoker */
	ascvd_smk = 0*(0.691);
	
/* 	Non-diabetic */
	ascvd_ever_told_dm2 = 0*(0.874);

/***********************************************************/
	ind_sum_blk_f_wtf = sum(of ascvd_:);
	ind_sum_blk_f = sum(ascvd_age, ascvd_tchol, ascvd_hdl, ascvd_age_x_hdl, ascvd_sbp_notx, ascvd_age_x_sbp_notx, ascvd_smk, ascvd_ever_told_dm2);
		
	mean_blk_f = 86.61;
	
	bline_surv_blk_f = 0.9533;
	
	ascvd_score = (1 - bline_surv_blk_f**(exp(ind_sum_blk_f - mean_blk_f)))*100;
	
	end;

ind_sum_blk_f_wtf=102.55228166

ind_sum_blk_f=87.81412573 (this is the correct/desired value)

Re: Sum Function

WarrenKuhfeld — Sat, 20 Aug 2022 02:43:20 GMT

The variable

ascvd_score

is in your first list (of ascdv_) but not the second.

Re: Sum Function

_maldini_ — Sat, 20 Aug 2022 22:03:01 GMT

@WarrenKuhfeld

Are you suggesting that ascvd_score is being included in the calculation of the sum? How could this be if the SUM function precedes the creation of that variable?

Re: Sum Function

Kurt_Bremser — Sat, 20 Aug 2022 22:09:18 GMT

What if the variable is already in the incoming dataset?

Re: Sum Function

Quentin — Sun, 21 Aug 2022 00:27:13 GMT

There must be some other variable that starts with ascvd_ which you are not expecting.

To see it, suggest you add:

put (ascvd_:) (=);

Before the assignment statement.

That should show you all of the variables in the ascvd_: variable list.

Re: Sum Function

WarrenKuhfeld — Sun, 21 Aug 2022 01:13:08 GMT

You did not provide full working code, but you asked why the two function calls were different. I pointed out the obvious difference that they used two different variable lists. Armed with that insight, you should be able to run your actual code and see what is going on.