topic MISSING INDICATOR TRUE/FALSE in SAS Programming

MISSING INDICATOR TRUE/FALSE

himalayancat — Fri, 22 Jun 2018 00:36:58 GMT

hi all ,

i saw someone try to flag ,from my presumtion on data step.

not the original code ,but i have tried to create ,what i saw.highlighted in yellow is what am specifically looking for.

proc sql noprint;

select count(*) into:cnt from sashelp.class;

quit;

data have;

set sashelp.class(keep=name sex age);

if _n_=&cnt then do;

a=1;

b='c';

end;

age_ind=age=age;

a_ind_1=a=a;

a_ind_2=a=.;

b_ind_1=b=b;

b_ind_2=b=" ";

run;

Question 1: why does age_ind=age=age flags all 1?

Question 2: why does a_ind_1=a=a flags all 1.(please see in xlsx attached)?(same as question 1)

Question 3: why does a_ind_2=a=. flags all 1 but the last( which is not missing) 0?

is it some TRUE/FALSE logic going on here??

Re: MISSING INDICATOR TRUE/FALSE

Patrick — Fri, 22 Jun 2018 00:55:38 GMT

@himalayancat

is it some TRUE/FALSE logic going on here?

Yes, there is.

What code like age_ind=age=age; does:

1. Logical expression age=age returns value of 1 if TRUE and 0 if False

2. Assigns value to variable age_ind on the left.

....and because age=age is always TRUE age_ind will always be set to 1.

Unlike to other languages a single equal sign is used for both logical expressions and value assignments to variables so the code can look a bit confusing.

For clarity it would may be better to formulate such logic like:

age_ind = (age EQ age);

Re: MISSING INDICATOR TRUE/FALSE

ballardw — Fri, 22 Jun 2018 14:36:29 GMT

You might want to compare:

a_ind_2=a=.;

with

a_ind_2alt = missing(a);

and

b_ind_2=b=" ";

with

b_ind_2al= missing(b);

SAS by default returns 1 for "true" and 0 for "false" when evaluating and comparison on the right side of an assignment.

So for non-trivial, your a=a for example, this can be a slightly faster execution (and shorter code)than:

if a=. then a_ind_2=1;

else a_ind_2=0.

Something along those lines is very common when creating indicator variables or flags or other dichotomous results.