4.3 Conditional Processing: IF-THEN/ELSE Statements & WHERE Subsetting

4.3 Conditional Processing: IF-THEN/ELSE Statements & WHERE Subsetting

Quick Answer: Conditional processing allows SAS programs to execute specific statements or filter observations based on whether a logical expression evaluates to true (non-zero and non-missing) or false (zero or missing). SAS provides two major mechanisms for conditional handling: IF-THEN/ELSE statements (executable statements used inside DATA steps) and WHERE statements (subsetting directives valid in both DATA and PROC steps).

IF-THEN / ELSE Statement Mechanics

An IF-THEN statement evaluates a logical condition for the current observation in the Program Data Vector (PDV). If the condition is true, SAS executes the statement following THEN. If false, SAS skips the THEN clause and proceeds to the next statement.

Basic Syntax and Comparison Operators

if condition then statement;
else if condition then statement;
else statement;

SAS supports both symbolic and mnemonic comparison operators:

Exam Trap: <> is not a not-equal operator in SAS. <> is the MAX operator (a <> b returns the larger of the two values) and >< is the MIN operator. The valid not-equal operators are NE, ^=, and ~=. Writing if Status <> 'Closed' therefore compares nothing — it evaluates the larger of the two operands, which is almost never what the programmer intended.

Mutually Exclusive Logic with ELSE IF

Using ELSE IF creates efficient, mutually exclusive decision trees. When an IF condition succeeds, SAS skips all remaining ELSE IF and ELSE clauses for that observation.

/* EFFICIENT: SAS stops checking conditions once a match is found */
if CreditScore >= 750 then Rating = 'Excellent';
else if CreditScore >= 700 then Rating = 'Good';
else if CreditScore >= 600 then Rating = 'Fair';
else Rating = 'Poor';

/* INEFFICIENT & BUG-PRONE: SAS evaluates ALL four IF statements for EVERY row */
if CreditScore >= 750 then Rating = 'Excellent';
if CreditScore >= 700 then Rating = 'Good';
if CreditScore >= 600 then Rating = 'Fair';
if CreditScore < 600 then Rating = 'Poor';

In the inefficient example, an observation with CreditScore = 780 first sets Rating = 'Excellent', then gets overwritten by Rating = 'Good', and finally gets overwritten by Rating = 'Fair'! Using ELSE IF prevents overwrites and reduces CPU execution time.

Conditional Execution of Multiple Statements: DO Groups

By default, a THEN or ELSE clause executes only a single SAS statement. To execute multiple statements conditionally, enclose them in a DO ... END block:

data work.bonus_processing;
   set sashelp.empdata;
   
   if Sales > 100000 then do;
      Bonus = Sales * 0.10;
      Tier = 'Top Performer';
      ExecutiveReview = 'Approved';
   end;
   else do;
      Bonus = Sales * 0.02;
      Tier = 'Standard';
      ExecutiveReview = 'Not Required';
   end;
run;

Subsetting IF vs. WHERE Statements

Subsetting is the process of selecting a specific subset of observations from an input dataset. SAS offers two ways to subset data in a DATA step: the subsetting IF statement (if condition;) and the WHERE statement (where condition;).

While they often produce identical output datasets, their underlying operational mechanics in the SAS engine are vastly different.

Subsetting IF Statement (if condition;)

• Operates during the execution phase.
• Reads the row from the input dataset into the PDV first, computes any new variables, and then evaluates the condition.
• If the condition is false, SAS immediately stops processing the current observation, discards it from the PDV, and returns to the top of the DATA step loop.

WHERE Statement (where condition;)

• Operates during the compilation/input phase.
• Acts as an index/input filter at the data engine level before observations are brought into the PDV.
• If an observation does not meet the WHERE condition, it is never read into memory, saving I/O and processing time.
• Can be used in PROC steps (e.g., proc print data=sashelp.class; where Age > 12; run;).

Detailed Comparison: IF vs. WHERE

Special WHERE Operators

The WHERE statement supports specialized operators not available in standard IF statements:

/* Range checking */
where Salary between 50000 and 80000;

/* Substring matching */
where LastName contains 'SON';

/* Pattern matching with wildcard (% for multiple chars, _ for single char) */
where Phone like '919-%';

/* Missing value test */
where Commission is missing; /* Or IS NULL */

/* Combining multiple WHERE conditions incrementally */
where State = 'NC';
where same-and Sales > 5000; /* Combines as State='NC' AND Sales>5000 */

Code Demonstration: Subsetting IF vs. WHERE

/* EXAMPLE 1: Valid WHERE statement in PROC PRINT */
proc print data=sashelp.shoes;
   where Region = 'Canada' and Sales > 10000;
run;

/* EXAMPLE 2: Valid Subsetting IF evaluating a newly created variable */
data work.heavy_returns;
   set sashelp.shoes;
   NetSales = Sales - Returns;
   /* MUST use IF here because NetSales does not exist in input dataset sashelp.shoes */
   if NetSales < 1000;
run;

/* EXAMPLE 3: INVALID WHERE statement attempt (WILL ERROR) */
data work.invalid_where;
   set sashelp.shoes;
   NetSales = Sales - Returns;
   /* ERROR: Variable NetSales is not on file SASHELP.SHOES */
   where NetSales < 1000;
run;