IF ELSE IF CONDITION

[singhsahab]

Hello,

I've a dataset test1 where i want to write expertise column based on some condition . when i gave the condition for Medium (else if level = 2 or 3 then) then else condition is not working, however I have a value for else. Can anybody please explain why Else condition is not working and how SAS considering given condition (else if level = 2 or 3).

data test1;
input NAME $ LEVEL;
if level = . then
expertise = 'Unknown';
else if level = 1 then
expertise = 'Low';
else if level = 2 or 3 then
expertise = 'Medium';
else
expertise = 'High';

CARDS;
Frank 1
Joan 2
Sui 2
Jose 3
Burt 4
Kelly .
Juan 1
;
run;

Thank you all in advance !!

[RW9]

You need:

else if level in (2,3) then

Or:

else if level=2 or level=3 then

You can't use 2 or 3 like that.

[FreelanceReinh · Accepted solution]

Hello @singhsahab,

The condition level = 2 or 3 is true regardless of the value of variable LEVEL because SAS interprets 3 (and numeric values in general) as a Boolean value of its own, like (level = 2) or (3). Hence, the subsequent ELSE statement will not take effect. Numeric values (as Boolean expressions) are evaluated as TRUE if they are nonmissing and not zero.

[Kurt_Bremser]

Logical operators are resolved after arithmetic operators and expressions. So

if level = 2 or 3

resolves to

if (level = 2) or (3)

3 (which is not zero or missing) is considered "true", so the condition will always be true, and the "else" never be executed.

[ballardw]

If you want to display a given text for values or ranges of values for a single variable consider using FORMATS instead.

You can apply a format at the time of use so adding a variable is usually not needed and the formats will create groups honored by analysis or graphing procedures.

Example:

proc format library=work;
value expertise
.='Unknown'
1='Low'
2,3='Medium'
other='High'
;
run;
data test1;
input NAME $ LEVEL;
CARDS;
Frank 1
Joan 2
Sui 2
Jose 3
Burt 4
Kelly .
Juan 1
;
run;

proc freq data=test1;
   /* missing is needed to include missing 
      values of the variable in proc freq 
      output
   */
   tables level/missing;
   format level expertise.;
run;

Another advantage of formats is the code for assigning values to ranges can be much more concise then a slew of If/then/else because you can use end of interval indicators:  32.1 - 34.5 includes the ends, 32.1 <- 34.5 excludes 32.1 but includes 32.10004 for example, and 32.1 <-< 34.5 excludes both end points. Additionally you have keywords LOW and HIGH to get from the smallest value or largest value (for numerics) that your SAS session can use.