# Operators and Control Flow

In this chapter, you learn about operators, control flow statements, and the C# preprocessor. Operators provide syntax for performing different calculations or actions appropriate for the operands within the calculation. Control flow statements provide the means for conditional logic within a program or looping over a section of code multiple times. After introducing the if control flow statement, the chapter looks at the concept of Boolean expressions, which are embedded within many control flow statements. Included is mention of how integers cannot be converted (even explicitly) to bool and the advantages of this restriction. The chapter ends with a discussion of the C# preprocessor directives.

Now that you have been introduced to the predefined data types (refer to Chapter 2), you can begin to learn how to use these data types in combination with operators to perform calculations. For example, you can make calculations on variables that you have declared.

Operators are used to perform mathematical or logical operations on values (or variables) called operands to produce a new value, called the result. For example, in Listing 4.1 the subtraction operator, -, is used to subtract two operands, the numbers 4 and 2. The result of the subtraction is stored in the variable difference.

There are three operator categories—unary, binary, and ternary—corresponding to the number of operands (one, two, and three, respectively). Furthermore, while some operators are represented with symbols like +, -, ?., and ??, other operators take the form of keywords, like default and is. This section covers some of the most basic unary and binary operators. The ternary operators appear later in the chapter.

Sometimes you may want to change the sign of a numeric value. In these cases, the unary minus operator (-) comes in handy. For example, Listing 4.2 changes the total current U.S. debt to a negative value to indicate that it is an amount owed.

Using the minus operator is equivalent to subtracting the operand from zero.

The unary plus operator (+) rarely1 has any effect on a value. It is a superfluous addition to the C# language and was included for the sake of symmetry.

Binary operators require two operands. C# uses infix notation for binary operators: The operator appears between the left and right operands. The result of every binary operator other than assignment must be used somehow—for example, by using it as an operand in another expression such as an assignment.

In contrast to the rule mentioned previously, C++ allows a single binary expression to form the entirety of a statement, such as 4+5;, and compile. In C#, only assignment, call, increment, decrement, await, and object creation expressions are allowed to be the entirety of a statement.

The subtraction example in Listing 4.3 with Output 4.1 illustrates the use of a binary operator—more specifically, an arithmetic binary operator. The operands appear on each side of the arithmetic operator, and then the calculated value is assigned. The other arithmetic binary operators are addition (+), division (/), multiplication (*), and remainder (% [sometimes called the mod operator]).

In the highlighted assignment statements, the division and remainder operations are executed before the assignments. The order in which operators are executed is determined by their precedence and associativity. The precedence for the operators used so far is as follows:

Therefore, you can assume that the statement behaves as expected, with the division and remainder operators executing before the assignment.

If you forget to assign the result of one of these binary operators, you will receive the compile error shown in Output 4.2.

When an expression contains multiple operators, it can be unclear precisely what the operands of each operator are. For example, in the expression x+y*z, clearly the expression x is an operand of the addition, and z is an operand of the multiplication. But is y an operand of the addition or the multiplication?

Parentheses allow you to unambiguously associate an operand with its operator. If you wish y to be a summand, you can write the expression as (x+y)*z; if you want it to be a multiplicand, you can write x+(y*z).

However, C# does not require you to parenthesize every expression containing more than one operator; instead, the compiler can use associativity and precedence to figure out from the context which parentheses you have omitted. Associativity determines how similar operators are parenthesized; precedence determines how dissimilar operators are parenthesized.

A binary operator may be left-associative or right-associative, depending on whether the expression “in the middle” belongs to the operator on the left or the right. For example, a-b-c is assumed to mean (a-b)-c, and not a-(b-c); subtraction is therefore said to be left-associative. Most operators in C# are left-associative; the assignment operators are right-associative.

When the operators are dissimilar, the precedence for those operators is used to determine the side to which the operand in the middle belongs. For example, multiplication has higher precedence than addition, so the expression x+y*z is evaluated as x+(y*z) rather than (x+y)*z.

It is often good practice to use parentheses to make the code more readable, even when the use of parentheses does not change the meaning of the expression. For example, when performing a Celsius-to-Fahrenheit temperature conversion, (c*9.0/5.0)+32.0 is easier to read than c*9.0/5.0+32.0, even though the parentheses are completely unnecessary.

Clearly, operators of higher precedence must execute before adjoining operators of lower precedence: in x+y*z, the multiplication must be executed before the addition because the result of the multiplication is the right-hand operand of the addition. However, it is important to realize that precedence and associativity affect only the order in which the operators themselves are executed; they do not in any way affect the order in which the operands are evaluated.

Operands are always evaluated from left to right in C#. In an expression with three method calls, such as A()+B()*C(), first A() is evaluated, then B(), then C(); then the multiplication operator determines the product; and finally the addition operator determines the sum. Just because C() is involved in a multiplication and A() is involved in a lower-precedence addition, that does not imply that method invocation C() happens before method invocation A().

In contrast to the rule mentioned here, the C++ specification allows an implementation broad latitude to decide the evaluation order of operands. When given an expression such as A()+B()*C(), a C++ compiler can choose to evaluate the function calls in any order, as long as the product is one of the summands. For example, a legal compiler could evaluate B(), then A(), then C(); then the product; and finally the sum.

Operators can also work with non-numeric operands. For example, it is possible to use the addition operator to concatenate two or more strings, as shown in Listing 4.4 with Output 4.3.

Because sentence structure varies among languages in different cultures, developers should be careful not to use the addition operator with strings that possibly will require localization. Similarly, although we can embed expressions within a string using string interpolation, localization to other languages still requires moving the string to a resource file, neutralizing the string interpolation. For this reason, you should use the addition operator sparingly, favoring composite formatting when localization is a possibility.

When introducing the char type in Chapter 2, we mentioned that even though it stores characters and not numbers, the char type is an integral type (integral means it is based on an integer). It can participate in arithmetic operations with other integer types. However, interpretation of the value of the char type is not based on the character stored within it but rather on its underlying value. The digit 3, for example, is represented by the Unicode value 0x33 (hexadecimal), which in base 10 is 51. The digit 4 is represented by the Unicode value 0x34, or 52 in base 10. Adding 3 and 4 in Listing 4.5 results in a hexadecimal value of 0x67, or 103 in base 10, which is the Unicode value for the letter g, as shown in Output 4.4.

You can use this trait of character types to determine how far two characters are from each other. For example, the letter f is three characters away from the letter c. You can determine this value by subtracting the letter c from the letter f, as Listing 4.6 with Output 4.5 demonstrates.

The binary floating-point types, float and double, have some special characteristics, such as the way they manage precision. This section looks at some specific examples, as well as some unique floating-point type characteristics.

A float, with seven decimal digits of precision, can hold the value 1,234,567 and the value 0.1234567. However, if you add these two floats together, the result will be rounded to 1,234,567, because the exact result requires more precision than the seven significant digits that a float can hold. The error introduced by rounding off to seven digits can become large compared to the value computed, especially with repeated calculations. (See also “Advanced Topic: Unexpected Inequality with Floating-Point Types” later in this section.)

Internally, the binary floating-point types store a binary fraction, not a decimal fraction. Consequently, “representation error” inaccuracies can occur with a simple assignment, such as double number = 140.6F. The exact value of 140.6 is the fraction 703/5, but the denominator of that fraction is not a power of 2, so a binary floating-point number cannot represent it exactly. Instead, the value represented is the closest fraction with a power of 2 in the denominator that fits into the 32 bits of a float.

Since the double can hold a more accurate value than the float can store, the C# compiler actually evaluates this expression to double number = 140.600006103516 because 140.600006103516 is the closest binary fraction to 140.6 as a float. This fraction is slightly larger than 140.6 when represented as a double.

Because floating-point numbers can be unexpectedly rounded off to non-decimal fractions, comparing floating-point values for equality can be quite confusing. Consider Listing 4.7 with Output 4.6.

The Assert() methods alert the developer whenever arguments evaluate to false. However, of all the Assert() calls in this code listing, only half have arguments that evaluate to true. Despite the apparent equality of the values in the code listing, they are not actually equivalent due to the inaccuracies associated with float values.

You should be aware of some additional unique floating-point characteristics as well. For instance, you would expect that dividing an integer by zero would result in an error—and it does with data types such as int and decimal. The float and double types, however, allow for certain special values. Consider Listing 4.8, and its resultant output, Output 4.7.

In mathematics, certain mathematical operations are undefined, including dividing zero by itself. In C#, the result of dividing the float zero by zero results in a special Not a Number (NaN) value; all attempts to print the output of such a number will result in NaN. Similarly, taking the square root of a negative number with System.Math.Sqrt(-1) will result in NaN.

A floating-point number could overflow its bounds as well. For example, the upper bound of the float type is approximately 3.4 × 1038. Should the number overflow that bound, the result would be stored as positive infinity, and the output of printing the number would be Infinity. Similarly, the lower bound of a float type is −3.4 × 1038, and computing a value below that bound would result in negative infinity, which would be represented by the string -Infinity. Listing 4.9 produces negative and positive infinity, respectively, and Output 4.8 shows the results.

Further examination of the floating-point number reveals that it can contain a value very close to zero without actually containing zero. If the value exceeds the lower threshold for the float or double type, the value of the number can be represented as negative zero or positive zero, depending on whether the number is negative or positive, and is represented in output as -0 or 0.

Chapter 1 discussed the simple assignment operator, which places the value of the right-hand side of the operator into the variable on the left-hand side. Compound mathematical assignment operators combine common binary operator calculations with the assignment operator. For example, consider Listing 4.10.

In this assignment, first you calculate the value of x + 2, and then you assign the calculated value back to x. Since this type of operation is performed relatively frequently, an assignment operator exists to handle both the calculation and the assignment with one operator. The += operator increments the variable on the left-hand side of the operator with the value on the right-hand side of the operator, as shown in Listing 4.11.

This code, therefore, is equivalent to Listing 4.10.

Numerous other compound assignment operators exist to provide similar functionality. You can also use the assignment operator with subtraction, multiplication, division, and remainder operators (as demonstrated in Listing 4.12).

C# includes special unary operators for incrementing and decrementing counters. The increment operator, ++, increments a variable by one each time it is used. In other words, all of the code lines shown in Listing 4.13 are equivalent.

Similarly, you can decrement a variable by 1 using the decrement operator, --. Therefore, all the code lines shown in Listing 4.14 are also equivalent.

The increment and decrement operators are especially prevalent in loops, such as the while loop described later in the chapter. For example, Listing 4.15 with Output 4.9 uses the decrement operator to iterate backward through each letter in the alphabet.

Listing 4.15 uses the increment and decrement operators to control how many times a particular operation is performed. In this example, notice that the decrement operator acts on a character (char) data type. You can use increment and decrement operators on various data types given some meaning is programmed to the concept of the “next” or “previous” value for that data type. See the section “Operator Overloading” in Chapter 10.

We saw that the assignment operator first computes the value and then performs the assignment. The result of the assignment operator is the value that was assigned. The increment and decrement operators are similar: They compute the value, perform the assignment, and return a value. It is therefore possible to use the assignment operator with the increment or decrement operator, though doing so carelessly can be extremely confusing. See Listing 4.16 and Output 4.10 for an example.

You might be surprised that result was assigned the value that was count before count was incremented. Where you place the increment or decrement operator determines whether the assigned value should be the value of the operand before or after the calculation. If you want the value of result to be the value assigned to count, you need to place the operator before the variable being incremented, as shown in Listing 4.17 with Output 4.11.

In this example, the increment operator appears before the operand, so the result of the expression is the value assigned to the variable after the increment. If count is 123, ++count assigns 124 to count and produces the result 124. By contrast, the postfix increment operator count++ assigns 124 to count and produces the value that count held before the increment: 123. Regardless of whether the operator is postfix or prefix, the variable count is incremented before the value is produced; the only difference is which value is produced. The difference between prefix and postfix behavior is illustrated in Listing 4.18. The resultant output is shown in Output 4.12.

As Listing 4.18 demonstrates, where the increment and decrement operators appear relative to the operand can affect the result produced by the expression. The result of the prefix operators is the value that the variable had before incrementing or decrementing. The result of the postfix operators is the value that the variable had after incrementing or decrementing. Use caution when embedding these operators in the middle of a statement. When in doubt as to what will happen, use these operators independently, placing them within their own statements. This way, the code is also more readable and there is no mistaking the intention.

Earlier we discussed how the operands in an expression can be evaluated in any order in C++, whereas they are always evaluated from left to right in C#. Similarly, in C++ an implementation may legally perform the side effects of increments and decrements in any order. For example, in C++ a call of the form M(x++, x++), where x begins as 1, can legally call either M(1,2) or M(2,1) at the whim of the compiler. In contrast, C# always calls M(1,2) because C# makes two guarantees: (1) The arguments to a call are always computed from left to right, and (2) the assignment of the incremented value to the variable always happens before the value of the expression is used. C++ makes neither guarantee.

Despite the brevity of the increment and decrement operators, these operators are not atomic. A thread context switch can occur during the execution of the operator and can cause a race condition. You could use a lock statement to prevent the race condition. However, for simple increments and decrements, a less expensive alternative is to use the thread-safe Increment() and Decrement() methods from the System.Threading.Interlocked class. These methods rely on processor functions for performing fast, thread-safe increments and decrements. See Chapter 19 for more details.

Chapter 3 discussed literal values, or values embedded directly into the code. It is possible to combine multiple literal values in a constant expression using operators. By definition, a constant expression is one that the C# compiler can evaluate at compile time (instead of evaluating it when the program runs) because it is composed entirely of constant operands. Constant expressions can then be used to initialize constant locals, which allow you to give a name to a constant value (similar to the way local variables allow you to give a name to a storage location). For example, the computation of the number of seconds in a day can be a constant expression that is then used in other expressions by name.

The const keyword in Listing 4.19 declares two constant locals: secondsPerDay and secondsPerWeek. Since a constant local is, by definition, the opposite of a variable—constant means “not able to vary”—any attempt to modify the value later in the code would result in a compile-time error.

In Listing 4.19, 60 * 60 * 34 and secondsPerDay * 7 are both constant expressions. Note that the expression assigned to secondsPerWeek is a constant expression because all the operands in the expression are also constants.

C# 10 added support for constant interpolated strings whereby you can define a constant string with interpolation if it is comprised solely of other constant strings and, as such, it can be evaluated at compile time (see Listing 4.20).

Even though announcement in Listing 4.20 is an interpolated string, the values within the code portions of the string are also constants, enabling the compiler to evaluate them at compile time rather than waiting until execution time. Also note, however, that windspeed is a string, not an integer. Constant string interpolation works when formed solely of other constant strings, not of other data types even if those types are constant. The restriction is to allow conversion to a string to account for culture variation at execution time. For example, converting the golden ratio to a string could be 1.6180339887 or 1,6180339887 depending on the culture.

________________________________________