4.1 Creating & Modifying Variables with Assignment Statements

4.1 Creating & Modifying Variables with Assignment Statements

Quick Answer: In SAS DATA step processing, an assignment statement evaluates an expression to the right of the equal sign and stores the resulting value in the variable specified to the left (variable = expression;). If a variable does not already exist in the input dataset or Program Data Vector (PDV), SAS creates it during the compilation phase and establishes its data type (numeric or character) and byte length based on its first appearance. To prevent character string truncation or unexpected default lengths, always use a LENGTH statement before assigning character values.

Anatomy of a SAS Assignment Statement

The assignment statement is the primary workhorse for data transformation in SAS. Its general syntax is:

variable = expression;

• variable: The name of a new or existing SAS variable. SAS variable names must follow standard naming rules: 1 to 32 characters long, begin with a letter or underscore, and contain only letters, numbers, or underscores.
• =: The assignment operator that copies the evaluated result of the expression into the variable's storage location in the PDV.
• expression: A combination of constants, variable names, mathematical operators, SAS functions, or logical operators that evaluates to a single numeric or character value.

Execution Flow in the DATA Step Loop

Unlike macro variables or compile-time directives, assignment statements execute sequentially during the execution phase for every observation processed by the DATA step iteration loop.

data work.employee_salaries;
   set sashelp.empdata;
   /* Arithmetic assignment statement */
   AnnualBonus = Salary * 0.15;
   /* Character assignment statement */
   BonusTier = 'Standard';
   /* Expression combining existing and newly created variables */
   TotalCompensation = Salary + AnnualBonus;
run;

When SAS compiles this DATA step:

1. It scans the statement tokens and identifies AnnualBonus, BonusTier, and TotalCompensation as new variables.
2. It adds these variables to the Program Data Vector (PDV).
3. It assigns default attributes (Numeric 8 bytes for numbers; Character length equal to the first string assigned for characters).
4. During execution, SAS reads a row from sashelp.empdata, computes the expressions row-by-row, updates the PDV, and writes the output row.

Variable Attributes & Compilation Mechanics

Every SAS variable possesses four core attributes: Name, Type (Numeric or Character), Length (bytes in memory), and optional Format/Informat / Label.

How SAS Determines Attributes Automatically

If you do not explicitly declare variable attributes using a LENGTH or ATTRIB statement, SAS inspects the first statement where the variable appears during compilation:

• Numeric Variables: Assigned a default storage length of 8 bytes (double-precision floating-point format), accommodating up to 16 significant digits.
• Character Variables: Assigned a length equal to the length of the string result in its very first reference or assignment in the DATA step code.

The Character Truncation Pitfall

One of the most frequent traps on the SAS Base Programming exam involves implicit character variable sizing. Consider the following DATA step:

data work.status_assignment;
   set work.orders;
   
   /* First reference to variable Status */
   if OrderAmount < 50 then Status = 'Low';
   else if OrderAmount < 200 then Status = 'Medium';
   else Status = 'High Priority';
run;

What Happens During Compilation?

1. SAS reads the first assignment: Status = 'Low';.
2. SAS determines Status is a character variable with length 3 (the byte length of 'Low').
3. SAS allocates exactly 3 bytes for Status in the PDV.

What Happens During Execution?

• When OrderAmount is 150, SAS attempts to assign 'Medium' (6 characters). Because Status is fixed at 3 bytes, SAS truncates the value to 'Med' without issuing any warning or error!
• When OrderAmount is 500, 'High Priority' (13 characters) is truncated to 'Hig'.

Resolving Truncation with the LENGTH Statement

To ensure proper byte allocation, place a LENGTH statement before the variable's first reference in the DATA step:

data work.status_fixed;
   set work.orders;
   
   /* Explicit length declaration MUST precede assignment */
   length Status $ 13;
   
   if OrderAmount < 50 then Status = 'Low';
   else if OrderAmount < 200 then Status = 'Medium';
   else Status = 'High Priority';
run;

Exam Rule: The LENGTH statement is a compile-time statement. Placing length Status $ 13; at the bottom of the DATA step after Status = 'Low'; will NOT work because compilation scans top-to-bottom and sets the PDV attributes on the first encounter.

Missing Values in Arithmetic Assignments

In SAS, numeric missing values are represented by a single period (.) or special missing characters (.A through .Z, ._). Character missing values are represented by a blank space (' ').

Propagation of Missing Values

When performing mathematical operations (+, -, *, /, **) in an assignment statement, if any operand in the expression is missing, the result of the entire expression evaluates to missing (.):

data work.calc_test;
   BaseSalary = 60000;
   Bonus = .; /* Missing numeric value */
   
   /* Standard arithmetic operator assignment */
   TotalComp1 = BaseSalary + Bonus; /* Evaluates to . */
   
   /* SAS Function assignment */
   TotalComp2 = sum(BaseSalary, Bonus); /* Evaluates to 60000 */
run;

When TotalComp1 = BaseSalary + Bonus; executes, SAS encounters the missing value for Bonus and performs the following actions:

1. Assigns TotalComp1 = . for that observation.
2. Writes a diagnostic note to the SAS log: NOTE: Missing values were generated as a result of performing an operation on missing values.

Contrast: Arithmetic Operators vs. SAS Functions

Formats vs. Stored Values in Assignments

A critical concept tested on the SAS Specialist exam is the difference between a variable's underlying stored value and its formatted display presentation.

data work.format_demo;
   format SalePrice dollar10.2;
   SalePrice = 1250.50;
   
   /* Derived assignment statement */
   Tax = SalePrice * 0.08;
   DiscountedPrice = SalePrice - 100;
run;

• In memory, SalePrice stores the exact unformatted floating-point number 1250.50.
• Tax is computed as 1250.50 * 0.08 = 100.04, NOT based on formatted text.
• Assigned variables do not automatically inherit formats from input variables. Tax and DiscountedPrice will store raw unformatted numbers (100.04 and 1150.50) until explicitly assigned a format.