Bruce Eckel's Thinking in Java Contents | Prev | Next

Using Java operators

An operator takes one or more arguments and produces a new value. The arguments are in a different form than ordinary method calls, but the effect is the same. You should be reasonably comfortable with the general concept of operators from your previous programming experience. Addition ( +), subtraction and unary minus ( -), multiplication ( *), division ( /) and assignment ( =) all work much the same in any programming language.

All operators produce a value from their operands. In addition, an operator can change the value of an operand. This is called a side effect. The most common use for operators that modify their operands is to generate the side effect, but you should keep in mind that the value produced is available for your use just as in operators without side effects.

Almost all operators work only with primitives. The exceptions are ‘=’, ‘ ==’ and ‘ !=’, which work with all objects (and are a point of confusion for objects). In addition, the String class supports ‘ +’ and ‘ +=’.

Precedence

Operator precedence defines how an expression evaluates when several operators are present. Java has specific rules that determine the order of evaluation. The easiest one to remember is that multiplication and division happen before addition and subtraction. Programmers often forget the other precedence rules, so you should use parentheses to make the order of evaluation explicit. For example:

A = X + Y - 2/2 + Z;

has a very different meaning from the same statement with a particular grouping of parentheses:

A = X + (Y - 2)/(2 + Z);

Assignment

Assignment is performed with the operator =. It means “take the value of the right-hand side (often called the rvalue) and copy it into the left-hand side (often called the lvalue). An rvalue is any constant, variable or expression that can produce a value, but an lvalue must be a distinct, named variable. (That is, there must be a physical space to store a value.) For instance, you can assign a constant value to a variable ( A = 4; ), but you cannot assign anything to constant value – it cannot be an lvalue. (You can’t say 4 = A;. )

Assignment of primitives is quite straightforward. Since the primitive holds the actual value and not a handle to an object, when you assign primitives you copy the contents from one place to another. For example, if you say A = B for primitives, then the contents of B is copied into A. If you then go on to modify A, B is naturally unaffected by this modification. This is what you’ve come to expect as a programmer for most situations.

When you assign objects, however, things change. Whenever you manipulate an object, what you’re manipulating is the handle, so when you assign “from one object to another” you’re actually copying a handle from one place to another. This means that if you say C = D for objects, you end up with both C and D pointing to the object that, originally, only D pointed to. The following example will demonstrate this.

As an aside, the first thing you see is a package statement for package c03 , indicating this book’s Chapter 3. The first code listing of each chapter will contain a package statement like this to establish the chapter number for the remaining code listings in that chapter. In Chapter 17, you’ll see that as a result, all the listings in chapter 3 (except those that have different package names) will be automatically placed in a subdirectory called c03, Chapter 4’s listings will be in c04 and so on. All this will happen via the CodePackager.java program shown in Chapter 17, and in Chapter 5 the concept of packages will be fully explained. What you need to recognize at this point is that, for this book, lines of code of the form package c03 are used just to establish the chapter subdirectory for the listings in the chapter.

In order to run the program, you must ensure that the classpath contains the root directory where you installed the source code for this book. (From this directory, you’ll see the subdirectories c02, c03, c04, etc.)

For later versions of Java (1.1.4 and on), when your main( ) is inside a file with a package statement, you must give the full package name before the program name in order to run the program. In this case, the command line is:

java c03.Assignment

Keep this in mind any time you’re running a program that’s in a package.

Here’s the example:

//: Assignment.java
// Assignment with objects is a bit tricky
package c03;

class Number {
  int i;
}

public class Assignment {
  public static void main(String[] args) {
    Number n1 = new Number();
    Number n2 = new Number();
    n1.i = 9;
    n2.i = 47;
    System.out.println("1: n1.i: " + n1.i +
      ", n2.i: " + n2.i);
    n1 = n2;
    System.out.println("2: n1.i: " + n1.i +
      ", n2.i: " + n2.i);
    n1.i = 27;
    System.out.println("3: n1.i: " + n1.i +
      ", n2.i: " + n2.i);
  }
} ///:~ 

The Number class is simple, and two instances of it ( n1 and n2) are created within main( ). The i value within each Number is given a different value, and then n2 is assigned to n1, and n1 is changed. In many programming languages you would expect n1 and n2 to be independent at all times, but because you’ve assigned a handle here’s the output you’ll see:

1: n1.i: 9, n2.i: 47
2: n1.i: 47, n2.i: 47
3: n1.i: 27, n2.i: 27 

Changing the n1 object appears to change the n2 object as well! This is because both n1 and n2 contain the same handle, which is pointing to the same object. (The original handle that was in n1 that pointed to the object holding a value of 9 was overwritten during the assignment and effectively lost; its object will be cleaned up by the garbage collector.)

This phenomenon is often called aliasing and it’s a fundamental way that Java works with objects. But what if you don’t want aliasing to occur in this case? You could forego the assignment and say:

n1.i = n2.i;

This retains the two separate objects instead of tossing one and tying n1 and n2 to the same object, but you’ll soon realize that manipulating the fields within objects is messy and goes against good object-oriented design principles. This is a non-trivial topic, so it is left for Chapter 12, which is devoted to aliasing. In the meantime, you should keep in mind that assignment for objects can add surprises.

Aliasing during method calls

Aliasing will also occur when you pass an object into a method:

//: PassObject.java
// Passing objects to methods can be a bit tricky

class Letter {
  char c;
}

public class PassObject {
  static void f(Letter y) {
    y.c = 'z';
  }
  public static void main(String[] args) {
    Letter x = new Letter();
    x.c = 'a';
    System.out.println("1: x.c: " + x.c);
    f(x);
    System.out.println("2: x.c: " + x.c);
  }
} ///:~ 

In many programming languages, the method f( ) would appear to be making a copy of its argument Letter y inside the scope of the method. But once again a handle is being passed so the line

y.c = 'z';

is actually changing the object outside of f( ). The output shows this:

1: x.c: a
2: x.c: z 

Aliasing and its solution is a complex issue and, although you must wait until Chapter 12 for all the answers, you should be aware of it at this point so you can watch for pitfalls.

Mathematical operators

The basic mathematical operators are the same as the ones available in most programming languages: addition (+), subtraction ( -), division ( /), multiplication ( *) and modulus ( %, produces the remainder from integer division). Integer division truncates, rather than rounds, the result.

Java also uses a shorthand notation to perform an operation and an assignment at the same time. This is denoted by an operator followed by an equal sign, and is consistent with all the operators in the language (whenever it makes sense). For example, to add 4 to the variable x and assign the result to x, use: x += 4; .

This example shows the use of the mathematical operators:

//: MathOps.java
// Demonstrates the mathematical operators
import java.util.*;

public class MathOps {
  // Create a shorthand to save typing:
  static void prt(String s) {
    System.out.println(s);
  }
  // shorthand to print a string and an int:
  static void pInt(String s, int i) {
    prt(s + " = " + i);
  }
  // shorthand to print a string and a float:
  static void pFlt(String s, float f) {
    prt(s + " = " + f);
  }
  public static void main(String[] args) {
    // Create a random number generator,
    // seeds with current time by default:
    Random rand = new Random();
    int i, j, k;
    // '%' limits maximum value to 99:
    j = rand.nextInt() % 100;
    k = rand.nextInt() % 100;
    pInt("j",j);  pInt("k",k);
    i = j + k; pInt("j + k", i);
    i = j - k; pInt("j - k", i);
    i = k / j; pInt("k / j", i);
    i = k * j; pInt("k * j", i);
    i = k % j; pInt("k % j", i);
    j %= k; pInt("j %= k", j);
    // Floating-point number tests:
    float u,v,w;  // applies to doubles, too
    v = rand.nextFloat();
    w = rand.nextFloat();
    pFlt("v", v); pFlt("w", w);
    u = v + w; pFlt("v + w", u);
    u = v - w; pFlt("v - w", u);
    u = v * w; pFlt("v * w", u);
    u = v / w; pFlt("v / w", u);
    // the following also works for
    // char, byte, short, int, long,
    // and double:
    u += v; pFlt("u += v", u);
    u -= v; pFlt("u -= v", u);
    u *= v; pFlt("u *= v", u);
    u /= v; pFlt("u /= v", u);
  }
} ///:~ 

The first thing you will see are some shorthand methods for printing: the prt( ) method prints a String, the pInt( ) prints a String followed by an int and the pFlt( ) prints a String followed by a float. Of course, they all ultimately end up using System.out.println( ).

To generate numbers, the program first creates a Random object. Because no arguments are passed during creation, Java uses the current time as a seed for the random number generator. The program generates a number of different types of random numbers with the Random object simply by calling different methods: nextInt( ), nextLong( ), nextFloat( ) or nextDouble( ) .

The modulus operator, when used with the result of the random number generator, limits the result to an upper bound of the operand minus one (99 in this case).

Unary minus and plus operators

The unary minus (-) and unary plus (+) are the same operators as binary minus and plus. The compiler figures out which use is intended by the way you write the expression. For instance, the statement

x = -a;

has an obvious meaning. The compiler is able to figure out:

x = a * -b;

but the reader might get confused, so it is more clear to say:

x = a * (-b);

The unary minus produces the negative of the value. Unary plus provides symmetry with unary minus, although it doesn’t do much.

Auto increment and decrement

Java, like C, is full of shortcuts. Shortcuts can make code much easier to type, and either easier or harder to read.

Two of the nicer shortcuts are the increment and decrement operators (often referred to as the auto-increment and auto-decrement operators). The decrement operator is -- and means “decrease by one unit.” The increment operator is ++ and means “increase by one unit.” If A is an int, for example, the expression ++A is equivalent to ( A = A + 1 ). Increment and decrement operators produce the value of the variable as a result.

There are two versions of each type of operator, often called the prefix and postfix versions. Pre-increment means the ++ operator appears before the variable or expression, and post-increment means the ++ operator appears after the variable or expression. Similarly, pre-decrement means the -- operator appears before the variable or expression, and post-decrement means the -- operator appears after the variable or expression. For pre-increment and pre-decrement, (i.e., ++A or --A), the operation is performed and the value is produced. For post-increment and post-decrement (i.e. A++ or A--), the value is produced, then the operation is performed. As an example:

//: AutoInc.java
// Demonstrates the ++ and -- operators

public class AutoInc {
  public static void main(String[] args) {
    int i = 1;
    prt("i : " + i);
    prt("++i : " + ++i); // Pre-increment
    prt("i++ : " + i++); // Post-increment
    prt("i : " + i);
    prt("--i : " + --i); // Pre-decrement
    prt("i-- : " + i--); // Post-decrement
    prt("i : " + i);
  }
  static void prt(String s) {
    System.out.println(s);
  }
} ///:~ 

The output for this program is:

i : 1
++i : 2
i++ : 2
i : 3
--i : 2
i-- : 2
i : 1 

You can see that for the prefix form you get the value after the operation has been performed, but with the postfix form you get the value before the operation is performed. These are the only operators (other than those involving assignment) that have side effects. (That is, they change the operand rather than using just its value.)

The increment operator is one explanation for the name C++, implying “one step beyond C.” In an early Java speech, Bill Joy (one of the creators), said that “Java=C++--“ (C plus plus minus minus), suggesting that Java is C++ with the unnecessary hard parts removed and therefore a much simpler language. As you progress in this book you’ll see that many parts are simpler, and yet Java isn’t that much easier than C++.

Relational operators

Relational operators generate a boolean result. They evaluate the relationship between the values of the operands. A relational expression produces true if the relationship is true, and false if the relationship is untrue. The relational operators are less than (<), greater than (>), less than or equal to (<=), greater than or equal to (>=), equivalent (==) and not equivalent (!=). Equivalence and nonequivalence works with all built-in data types, but the other comparisons won’t work with type boolean.

Testing object equivalence

The relational operators == and != also work with all objects, but their meaning often confuses the first-time Java programmer. Here’s an example:

//: Equivalence.java

public class Equivalence {
  public static void main(String[] args) {
    Integer n1 = new Integer(47);
    Integer n2 = new Integer(47);
    System.out.println(n1 == n2);
    System.out.println(n1 != n2);
  }
} ///:~ 

The expression System.out.println(n1 == n2) will print out the result of the boolean comparison within it. Surely the output should be true and then false, since both Integer objects are the same. But while the contents of the objects are the same, the handles are not the same and the operators == and != compare object handles. So the output is actually false and then true. Naturally, this surprises people at first.

What if you want to compare the actual contents of an object for equivalence? You must use the special method equals( ) that exists for all objects (not primitives, which work fine with == and !=). Here’s how it’s used:

//: EqualsMethod.java

public class EqualsMethod {
  public static void main(String[] args) {
    Integer n1 = new Integer(47);
    Integer n2 = new Integer(47);
    System.out.println(n1.equals(n2));
  }
} ///:~ 

The result will be true, as you would expect. Ah, but it’s not as simple as that. If you create your own class, like this:

//: EqualsMethod2.java

class Value {
  int i;
}

public class EqualsMethod2 {
  public static void main(String[] args) {
    Value v1 = new Value();
    Value v2 = new Value();
    v1.i = v2.i = 100;
    System.out.println(v1.equals(v2));
  }
} ///:~ 

you’re back to square one: the result is false. This is because the default behavior of equals( ) is to compare handles. So unless you override equals( ) in your new class you won’t get the desired behavior. Unfortunately, you won’t learn about overriding until Chapter 7, but being aware of the way equals( ) behaves might save you some grief in the meantime.

Most of the Java library classes implement equals( ) so that it compares the contents of objects instead of their handles.

Logical operators

The logical operators AND (&&), OR (||) and NOT (!) produce a boolean value of true or false based on the logical relationship of its arguments. This example uses the relational and logical operators:

//: Bool.java
// Relational and logical operators
import java.util.*;

public class Bool {
  public static void main(String[] args) {
    Random rand = new Random();
    int i = rand.nextInt() % 100;
    int j = rand.nextInt() % 100;
    prt("i = " + i);
    prt("j = " + j);
    prt("i > j is " + (i > j));
    prt("i < j is " + (i < j));
    prt("i >= j is " + (i >= j));
    prt("i <= j is " + (i <= j));
    prt("i == j is " + (i == j));
    prt("i != j is " + (i != j));

    // Treating an int as a boolean is 
    // not legal Java
//! prt("i && j is " + (i && j));
//! prt("i || j is " + (i || j));
//! prt("!i is " + !i);

    prt("(i < 10) && (j < 10) is "
       + ((i < 10) && (j < 10)) );
    prt("(i < 10) || (j < 10) is "
       + ((i < 10) || (j < 10)) );
  }
  static void prt(String s) {
    System.out.println(s);
  }
} ///:~ 

You can apply AND, OR, or NOT to boolean values only. You can’t use a non- boolean as if it were a boolean in a logical expression as you can in C and C++. You can see the failed attempts at doing this commented out with a //! comment marker. The subsequent expressions, however, produce boolean values using relational comparisons, then use logical operations on the results.

One output listing looked like this:

i = 85
j = 4
i > j is true
i < j is false
i >= j is true
i <= j is false
i == j is false
i != j is true
(i < 10) && (j < 10) is false
(i < 10) || (j < 10) is true 

Note that a boolean value is automatically converted to an appropriate text form if it’s used where a String is expected.

You can replace the definition for int in the above program with any other primitive data type except boolean. Be aware, however, that the comparison of floating-point numbers is very strict. A number that is the tiniest fraction different from another number is still “not equal.” A number that is the tiniest bit above zero is still nonzero.

Short-circuiting

When dealing with logical operators you run into a phenomenon called “short circuiting.” This means that the expression will be evaluated only until the truth or falsehood of the entire expression can be unambiguously determined. As a result, all the parts of a logical expression might not be evaluated. Here’s an example that demonstrates short-circuiting:

//: ShortCircuit.java
// Demonstrates short-circuiting behavior
// with logical operators.

public class ShortCircuit {
  static boolean test1(int val) {
    System.out.println("test1(" + val + ")");
    System.out.println("result: " + (val < 1));
    return val < 1;
  }
  static boolean test2(int val) {
    System.out.println("test2(" + val + ")");
    System.out.println("result: " + (val < 2));
    return val < 2;
  }
  static boolean test3(int val) {
    System.out.println("test3(" + val + ")");
    System.out.println("result: " + (val < 3));
    return val < 3;
  }
  public static void main(String[] args) {
    if(test1(0) && test2(2) && test3(2))
      System.out.println("expression is true");
    else
      System.out.println("expression is false");
  }
} ///:~ 

Each test performs a comparison against the argument and returns true or false. It also prints information to show you that it’s being called. The tests are used in the expression:

if(test1(0) && test2(2) && test3(2))

You might naturally think that all three tests would be executed, but the output shows otherwise:

test1(0)
result: true
test2(2)
result: false
expression is false 

The first test produced a true result, so the expression evaluation continues. However, the second test produced a false result. Since this means that the whole expression must be false, why continue evaluating the rest of the expression? It could be expensive. The reason for short-circuiting, in fact, is precisely that; you can get a potential performance increase if all the parts of a logical expression do not need to be evaluated.

Bitwise operators

The bitwise operators allow you to manipulate individual bits in an integral primitive data type. Bitwise operators perform boolean algebra on the corresponding bits in the two arguments to produce the result.

The bitwise operators come from C’s low-level orientation; you were often manipulating hardware directly and had to set the bits in hardware registers. Java was originally designed to be embedded in TV set-top boxes, so this low-level orientation still made sense. However, you probably won’t use the bitwise operators much.

The bitwise AND operator (&) produces a one in the output bit if both input bits are one; otherwise it produces a zero. The bitwise OR operator (|) produces a one in the output bit if either input bit is a one and produces a zero only if both input bits are zero. The bitwise, EXCLUSIVE OR, or XOR (^), produces a one in the output bit if one or the other input bit is a one, but not both. The bitwise NOT (~, also called the ones complement operator) is a unary operator; it takes only one argument. (All other bitwise operators are binary operators.) Bitwise NOT produces the opposite of the input bit – a one if the input bit is zero, a zero if the input bit is one.

The bitwise operators and logical operators use the same characters, so it is helpful to have a mnemonic device to help you remember the meanings: since bits are “small,” there is only one character in the bitwise operators.

Bitwise operators can be combined with the = sign to unite the operation and assignment: &=, |= and ^= are all legitimate. (Since ~ is a unary operator it cannot be combined with the = sign.)

The boolean type is treated as a one-bit value so it is somewhat different. You can perform a bitwise AND, OR and XOR, but you can’t perform a bitwise NOT (presumably to prevent confusion with the logical NOT). For booleans the bitwise operators have the same effect as the logical operators except that they do not short circuit. Also, the bitwise operators on booleans gives you a XOR logical operator that is not included under the list of “logical” operators. You’re prevented from using booleans in shift expressions, which is described next.

Shift operators

The shift operators also manipulate bits. They can be used solely with primitive, integral types. The left-shift operator ( <<) produces the operand to the left of the operator shifted to the left by the number of bits specified after the operator (inserting zeroes at the lower-order bits). The signed right-shift operator ( >>) produces the operand to the left of the operator shifted to the right by the number of bits specified after the operator. The signed right shift >> uses sign extension : if the value is positive, zeroes are inserted at the higher-order bits; if the value is negative, ones are inserted at the higher-order bits. Java has also added the unsigned right shift >>>, which uses zero extension : regardless of the sign, zeroes are inserted at the higher-order bits. This operator does not exist in C or C++.

If you shift a char, byte, or short, it will be promoted to int before the shift takes place, and the result will be an int. Only the five low-order bits of the right-hand side will be used. This prevents you from shifting more than the number of bits in an int. If you’re operating on a long, long will be the result. Only the six low-order bits of the right-hand side will be used so you can’t shift more than the number of bits in a long. There is a problem, however, with the unsigned right shift. If you use it with byte or short you might not get the correct results. (It’s broken in Java 1.0 and Java 1.1.) These are promoted to int and right shifted, but the zero extension does not occur, so you get -1 in those cases. The following example can be used to test your implementation:

//: URShift.java
// Test of unsigned right shift

public class URShift {
  public static void main(String[] args) {
    int i = -1;
    i >>>= 10;
    System.out.println(i);
    long l = -1;
    l >>>= 10;
    System.out.println(l);
    short s = -1;
    s >>>= 10;
    System.out.println(s);
    byte b = -1;
    b >>>= 10;
    System.out.println(b);
  }
} ///:~ 

Shifts can be combined with the equal sign ( <<= or >>= or >>>=). The lvalue is replaced by the lvalue shifted by the rvalue.

Here’s an example that demonstrates the use of all the operators involving bits:

//: BitManipulation.java
// Using the bitwise operators
import java.util.*;

public class BitManipulation {
  public static void main(String[] args) {
    Random rand = new Random();
    int i = rand.nextInt();
    int j = rand.nextInt();
    pBinInt("-1", -1);
    pBinInt("+1", +1);
    int maxpos = 2147483647;
    pBinInt("maxpos", maxpos);
    int maxneg = -2147483648;
    pBinInt("maxneg", maxneg);
    pBinInt("i", i);
    pBinInt("~i", ~i);
    pBinInt("-i", -i);
    pBinInt("j", j);
    pBinInt("i & j", i & j);
    pBinInt("i | j", i | j);
    pBinInt("i ^ j", i ^ j);
    pBinInt("i << 5", i << 5);
    pBinInt("i >> 5", i >> 5);
    pBinInt("(~i) >> 5", (~i) >> 5);
    pBinInt("i >>> 5", i >>> 5);
    pBinInt("(~i) >>> 5", (~i) >>> 5);

    long l = rand.nextLong();
    long m = rand.nextLong();
    pBinLong("-1L", -1L);
    pBinLong("+1L", +1L);
    long ll = 9223372036854775807L;
    pBinLong("maxpos", ll);
    long lln = -9223372036854775808L;
    pBinLong("maxneg", lln);
    pBinLong("l", l);
    pBinLong("~l", ~l);
    pBinLong("-l", -l);
    pBinLong("m", m);
    pBinLong("l & m", l & m);
    pBinLong("l | m", l | m);
    pBinLong("l ^ m", l ^ m);
    pBinLong("l << 5", l << 5);
    pBinLong("l >> 5", l >> 5);
    pBinLong("(~l) >> 5", (~l) >> 5);
    pBinLong("l >>> 5", l >>> 5);
    pBinLong("(~l) >>> 5", (~l) >>> 5);
  }
  static void pBinInt(String s, int i) {
    System.out.println(
      s + ", int: " + i + ", binary: ");
    System.out.print("   ");
    for(int j = 31; j >=0; j--)
      if(((1 << j) &  i) != 0)
        System.out.print("1");
      else
        System.out.print("0");
    System.out.println();
  }
  static void pBinLong(String s, long l) {
    System.out.println(
      s + ", long: " + l + ", binary: ");
    System.out.print("   ");
    for(int i = 63; i >=0; i--)
      if(((1L << i) & l) != 0)
        System.out.print("1");
      else
        System.out.print("0");
    System.out.println();
  }
} ///:~ 

The two methods at the end, pBinInt( ) and pBinLong( ) take an int or a long, respectively, and print it out in binary format along with a descriptive string. You can ignore the implementation of these for now.

You’ll note the use of System.out.print( ) instead of System.out.println( ). The print( ) method does not put out a new line, so it allows you to output a line in pieces.

As well as demonstrating the effect of all the bitwise operators for int and long, this example also shows the minimum, maximum, +1 and -1 values for int and long so you can see what they look like. Note that the high bit represents the sign: 0 means positive and 1 means negative. The output for the int portion looks like this:

-1, int: -1, binary: 
   11111111111111111111111111111111
+1, int: 1, binary: 
   00000000000000000000000000000001
maxpos, int: 2147483647, binary: 
   01111111111111111111111111111111
maxneg, int: -2147483648, binary: 
   10000000000000000000000000000000
i, int: 59081716, binary: 
   00000011100001011000001111110100
~i, int: -59081717, binary: 
   11111100011110100111110000001011
-i, int: -59081716, binary: 
   11111100011110100111110000001100
j, int: 198850956, binary: 
   00001011110110100011100110001100
i & j, int: 58720644, binary: 
   00000011100000000000000110000100
i | j, int: 199212028, binary: 
   00001011110111111011101111111100
i ^ j, int: 140491384, binary: 
   00001000010111111011101001111000
i << 5, int: 1890614912, binary: 
   01110000101100000111111010000000
i >> 5, int: 1846303, binary: 
   00000000000111000010110000011111
(~i) >> 5, int: -1846304, binary: 
   11111111111000111101001111100000
i >>> 5, int: 1846303, binary: 
   00000000000111000010110000011111
(~i) >>> 5, int: 132371424, binary: 
   00000111111000111101001111100000 

The binary representation of the numbers is referred to as signed two’s complement .

Ternary if-else operator

This operator is unusual because it has three operands. It is truly an operator because it produces a value, unlike the ordinary if-else statement that you’ll see in the next section of this chapter. The expression is of the form

boolean-exp ? value0 : value1

If boolean-exp evaluates to true, value0 is evaluated and its result becomes the value produced by the operator. If boolean-exp is false, value1 is evaluated and its result becomes the value produced by the operator.

Of course, you could use an ordinary if-else statement (described later), but the ternary operator is much terser. Although C prides itself on being a terse language, and the ternary operator might have been introduced partly for efficiency, you should be somewhat wary of using it on an everyday basis – it’s easy to produce unreadable code.

The conditional operator can be used for its side effects or for the value it produces, but in general you want the value since that’s what makes the operator distinct from the if-else. Here’s an example:

static int ternary(int i) {
  return i < 10 ? i * 100 : i * 10;
} 

You can see that this code is more compact than what you’d need to write without the ternary operator:

static int alternative(int i) {
  if (i < 10)
    return i * 100;
  return i * 10;
}

The second form is easier to understand, and doesn’t require a lot more typing. So be sure to ponder your reasons when choosing the ternary operator.

The comma operator

The comma is used in C and C++ not only as a separator in function argument lists, but also as an operator for sequential evaluation. The sole place that the comma operator is used in Java is in for loops, which will be described later in this chapter.

String operator +

There’s one special usage of an operator in Java: the + operator can be used to concatenate strings, as you’ve already seen. It seems a natural use of the + even though it doesn’t fit with the traditional way that + is used. This capability seemed like a good idea in C++, so operator overloading was added to C++ to allow the C++ programmer to add meanings to almost any operator. Unfortunately, operator overloading combined with some of the other restrictions in C++ turns out to be a fairly complicated feature for programmers to design into their classes. Although operator overloading would have been much simpler to implement in Java than it was in C++, this feature was still considered too complex, so Java programmers cannot implement their own overloaded operators as C++ programmers can.

The use of the String + has some interesting behavior. If an expression begins with a String, then all operands that follow must be Strings:

int x = 0, y = 1, z = 2;

String sString = "x, y, z ";

System.out.println(sString + x + y + z);

Here, the Java compiler will convert x, y, and z into their String representations instead of adding them together first. However, if you say:

System.out.println(x + sString);

earlier versions of Java will signal an error. (Later versions, however, will turn x into a String.) So if you’re putting together a String (using an earlier version of Java) with addition, make sure the first element is a String (or a quoted sequence of characters, which the compiler recognizes as a String).

Common pitfalls when using operators

One of the pitfalls when using operators is trying to get away without parentheses when you are even the least bit uncertain about how an expression will evaluate. This is still true in Java.

An extremely common error in C and C++ looks like this:

while(x = y) {
    // ....
}

The programmer was trying to test for equivalence ( ==) rather than do an assignment. In C and C++ the result of this assignment will always be true if y is nonzero, and you’ll probably get an infinite loop. In Java, the result of this expression is not a boolean, and the compiler expects a boolean and won’t convert from an int, so it will conveniently give you a compile-time error and catch the problem before you ever try to run the program. So the pitfall never happens in Java. (The only time you won’t get a compile-time error is when x and y are boolean, in which case x = y is a legal expression, and in the above case, probably an error.)

A similar problem in C and C++ is using bitwise AND and OR instead of logical. Bitwise AND and OR use one of the characters ( & or |) while logical AND and OR use two ( && and ||). Just as with = and ==, it’s easy to type just one character instead of two. In Java, the compiler again prevents this because it won’t let you cavalierly use one type where it doesn’t belong.

Casting operators

The word cast is used in the sense of “casting into a mold.” Java will automatically change one type of data into another when appropriate. For instance, if you assign an integral value to a floating-point variable, the compiler will automatically convert the int to a float. Casting allows you to make this type conversion explicit, or to force it when it wouldn’t normally happen.

To perform a cast, put the desired data type (including all modifiers) inside parentheses to the left of any value. Here’s an example:

void casts() {
  int i = 200;
  long l = (long)i;
  long l2 = (long)200;
}

As you can see, it’s possible to perform a cast on a numeric value as well as on a variable. In both casts shown here, however, the cast is superfluous, since the compiler will automatically promote an int value to a long when necessary. You can still put a cast in to make a point or to make your code more clear. In other situations, a cast is essential just to get the code to compile.

In C and C++, casting can cause some headaches. In Java, casting is safe, with the exception that when you perform a so-called narrowing conversion (that is, when you go from a data type that can hold more information to one that doesn’t hold as much) you run the risk of losing information. Here the compiler forces you to do a cast, in effect saying “this can be a dangerous thing to do – if you want me to do it anyway you must make the cast explicit.” With a widening conversion an explicit cast is not needed because the new type will more than hold the information from the old type so that no information is ever lost.

Java allows you to cast any primitive type to any other primitive type, except for boolean, which doesn’t allow any casting at all. Class types do not allow casting. To convert one to the other there must be special methods. ( String is a special case, and you’ll find out later in the book that objects can be cast within a family of types; an Oak can be cast to a Tree and vice-versa, but not to a foreign type such as a Rock.)

Literals

Ordinarily when you insert a literal value into a program the compiler knows exactly what type to make it. Sometimes, however, the type is ambiguous. When this happens you must guide the compiler by adding some extra information in the form of characters associated with the literal value. The following code shows these characters:

//: Literals.java

class Literals {
  char c = 0xffff; // max char hex value
  byte b = 0x7f; // max byte hex value
  short s = 0x7fff; // max short hex value
  int i1 = 0x2f; // Hexadecimal (lowercase)
  int i2 = 0X2F; // Hexadecimal (uppercase)
  int i3 = 0177; // Octal (leading zero)
  // Hex and Oct also work with long.
  long n1 = 200L; // long suffix
  long n2 = 200l; // long suffix
  long n3 = 200;
  //! long l6(200); // not allowed
  float f1 = 1;
  float f2 = 1F; // float suffix
  float f3 = 1f; // float suffix
  float f4 = 1e-45f; // 10 to the power
  float f5 = 1e+9f; // float suffix
  double d1 = 1d; // double suffix
  double d2 = 1D; // double suffix
  double d3 = 47e47d; // 10 to the power
} ///:~ 

Hexadecimal (base 16), which works with all the integral data types, is denoted by a leading 0x or 0X followed by 0–9 and a–f either in upper or lower case. If you try to initialize a variable with a value bigger than it can hold (regardless of the numerical form of the value), the compiler will give you an error message. Notice in the above code the maximum possible hexadecimal values for char, byte, and short. If you exceed these, the compiler will automatically make the value an int and tell you that you need a narrowing cast for the assignment. You’ll know you’ve stepped over the line.

Octal (base 8) is denoted by a leading zero in the number and digits from 0-7. There is no literal representation for binary numbers in C, C++ or Java.

A trailing character after a literal value establishes its type. Upper or lowercase L means long, upper or lowercase F means float and upper or lowercase D means double.

Exponents use a notation that I’ve always found rather dismaying: 1.39 e-47f . In science and engineering, ‘e’ refers to the base of natural logarithms, approximately 2.718. (A more precise double value is available in Java as Math.E.) This is used in exponentiation expressions such as 1.39 x e -47, which means 1.39 x 2.718 -47. However, when FORTRAN was invented they decided that e would naturally mean “ten to the power,” which is an odd decision because FORTRAN was designed for science and engineering and one would think its designers would be sensitive about introducing such an ambiguity. [16] At any rate, this custom was followed in C, C++ and now Java. So if you’re used to thinking in terms of e as the base of natural logarithms, you must do a mental translation when you see an expression such as 1.39 e-47f in Java; it means 1.39 x 10 -47.

Note that you don’t need to use the trailing character when the compiler can figure out the appropriate type. With

long n3 = 200;

there’s no ambiguity, so an L after the 200 would be superfluous. However, with

float f4 = 1e-47f; // 10 to the power

the compiler normally takes exponential numbers as doubles, so without the trailing f it will give you an error telling you that you must use a cast to convert double to float.

Promotion

You’ll discover that if you perform any mathematical or bitwise operations on primitive data types that are smaller than an int (that is, char, byte, or short), those values will be promoted to int before performing the operations, and the resulting value will be of type int. So if you want to assign back into the smaller type, you must use a cast. (And, since you’re assigning back into a smaller type, you might be losing information.) In general, the largest data type in an expression is the one that determines the size of the result of that expression; if you multiply a float and a double, the result will be double; if you add an int and a long, the result will be long.

Java has no “sizeof”

In C and C++, the sizeof( ) operator satisfies a specific need: it tells you the number of bytes allocated for data items. The most compelling need for sizeof( ) in C and C++ is portability. Different data types might be different sizes on different machines, so the programmer must find out how big those types are when performing operations that are sensitive to size. For example, one computer might store integers in 32 bits, whereas another might store integers as 16 bits. Programs could store larger values in integers on the first machine. As you might imagine, portability is a huge headache for C and C++ programmers.

Java does not need a sizeof( ) operator for this purpose because all the data types are the same size on all machines. You do not need to think about portability on this level – it is designed into the language.

Precedence revisited

Upon hearing me complain about the complexity of remembering operator precedence during one of my seminars, a student suggested a mnemonic that is simultaneously a commentary: “Ulcer Addicts Really Like C A lot.”

Mnemonic

Operator type

Operators

Ulcer

Unary

+ - ++ – [[ rest...]]

Addicts

Arithmetic (and shift)

* / % + - << >>

Really

Relational

> < >= <= == !=

Like

Logical (and bitwise)

&& || & | ^

C

Conditional (ternary)

A > B ? X : Y

A Lot

Assignment

= (and compound assignment like *=)

Of course, with the shift and bitwise operators distributed around the table it is not a perfect mnemonic, but for non-bit operations it works.

A compendium of operators

The following example shows which primitive data types can be used with particular operators. Basically, it is the same example repeated over and over, but using different primitive data types. The file will compile without error because the lines that would cause errors are commented out with a //!.

//: AllOps.java
// Tests all the operators on all the
// primitive data types to show which
// ones are accepted by the Java compiler.

class AllOps {
  // To accept the results of a boolean test:
  void f(boolean b) {}
  void boolTest(boolean x, boolean y) {
    // Arithmetic operators:
    //! x = x * y;
    //! x = x / y;
    //! x = x % y;
    //! x = x + y;
    //! x = x - y;
    //! x++;
    //! x--;
    //! x = +y;
    //! x = -y;
    // Relational and logical:
    //! f(x > y);
    //! f(x >= y);
    //! f(x < y);
    //! f(x <= y);
    f(x == y);
    f(x != y);
    f(!y);
    x = x && y;
    x = x || y;
    // Bitwise operators:
    //! x = ~y;
    x = x & y;
    x = x | y;
    x = x ^ y;
    //! x = x << 1;
    //! x = x >> 1;
    //! x = x >>> 1;
    // Compound assignment:
    //! x += y;
    //! x -= y;
    //! x *= y;
    //! x /= y;
    //! x %= y;
    //! x <<= 1;
    //! x >>= 1;
    //! x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! char c = (char)x;
    //! byte B = (byte)x;
    //! short s = (short)x;
    //! int i = (int)x;
    //! long l = (long)x;
    //! float f = (float)x;
    //! double d = (double)x;
  }
  void charTest(char x, char y) {
    // Arithmetic operators:
    x = (char)(x * y);
    x = (char)(x / y);
    x = (char)(x % y);
    x = (char)(x + y);
    x = (char)(x - y);
    x++;
    x--;
    x = (char)+y;
    x = (char)-y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    x= (char)~y;
    x = (char)(x & y);
    x  = (char)(x | y);
    x = (char)(x ^ y);
    x = (char)(x << 1);
    x = (char)(x >> 1);
    x = (char)(x >>> 1);
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    x <<= 1;
    x >>= 1;
    x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    byte B = (byte)x;
    short s = (short)x;
    int i = (int)x;
    long l = (long)x;
    float f = (float)x;
    double d = (double)x;
  }
  void byteTest(byte x, byte y) {
    // Arithmetic operators:
    x = (byte)(x* y);
    x = (byte)(x / y);
    x = (byte)(x % y);
    x = (byte)(x + y);
    x = (byte)(x - y);
    x++;
    x--;
    x = (byte)+ y;
    x = (byte)- y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    x = (byte)~y;
    x = (byte)(x & y);
    x = (byte)(x | y);
    x = (byte)(x ^ y);
    x = (byte)(x << 1);
    x = (byte)(x >> 1);
    x = (byte)(x >>> 1);
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    x <<= 1;
    x >>= 1;
    x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    short s = (short)x;
    int i = (int)x;
    long l = (long)x;
    float f = (float)x;
    double d = (double)x;
  }
  void shortTest(short x, short y) {
    // Arithmetic operators:
    x = (short)(x * y);
    x = (short)(x / y);
    x = (short)(x % y);
    x = (short)(x + y);
    x = (short)(x - y);
    x++;
    x--;
    x = (short)+y;
    x = (short)-y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    x = (short)~y;
    x = (short)(x & y);
    x = (short)(x | y);
    x = (short)(x ^ y);
    x = (short)(x << 1);
    x = (short)(x >> 1);
    x = (short)(x >>> 1);
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    x <<= 1;
    x >>= 1;
    x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    byte B = (byte)x;
    int i = (int)x;
    long l = (long)x;
    float f = (float)x;
    double d = (double)x;
  }
  void intTest(int x, int y) {
    // Arithmetic operators:
    x = x * y;
    x = x / y;
    x = x % y;
    x = x + y;
    x = x - y;
    x++;
    x--;
    x = +y;
    x = -y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    x = ~y;
    x = x & y;
    x = x | y;
    x = x ^ y;
    x = x << 1;
    x = x >> 1;
    x = x >>> 1;
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    x <<= 1;
    x >>= 1;
    x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    byte B = (byte)x;
    short s = (short)x;
    long l = (long)x;
    float f = (float)x;
    double d = (double)x;
  }
  void longTest(long x, long y) {
    // Arithmetic operators:
    x = x * y;
    x = x / y;
    x = x % y;
    x = x + y;
    x = x - y;
    x++;
    x--;
    x = +y;
    x = -y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    x = ~y;
    x = x & y;
    x = x | y;
    x = x ^ y;
    x = x << 1;
    x = x >> 1;
    x = x >>> 1;
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    x <<= 1;
    x >>= 1;
    x >>>= 1;
    x &= y;
    x ^= y;
    x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    byte B = (byte)x;
    short s = (short)x;
    int i = (int)x;
    float f = (float)x;
    double d = (double)x;
  }
  void floatTest(float x, float y) {
    // Arithmetic operators:
    x = x * y;
    x = x / y;
    x = x % y;
    x = x + y;
    x = x - y;
    x++;
    x--;
    x = +y;
    x = -y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    //! x = ~y;
    //! x = x & y;
    //! x = x | y;
    //! x = x ^ y;
    //! x = x << 1;
    //! x = x >> 1;
    //! x = x >>> 1;
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    //! x <<= 1;
    //! x >>= 1;
    //! x >>>= 1;
    //! x &= y;
    //! x ^= y;
    //! x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    byte B = (byte)x;
    short s = (short)x;
    int i = (int)x;
    long l = (long)x;
    double d = (double)x;
  }
  void doubleTest(double x, double y) {
    // Arithmetic operators:
    x = x * y;
    x = x / y;
    x = x % y;
    x = x + y;
    x = x - y;
    x++;
    x--;
    x = +y;
    x = -y;
    // Relational and logical:
    f(x > y);
    f(x >= y);
    f(x < y);
    f(x <= y);
    f(x == y);
    f(x != y);
    //! f(!x);
    //! f(x && y);
    //! f(x || y);
    // Bitwise operators:
    //! x = ~y;
    //! x = x & y;
    //! x = x | y;
    //! x = x ^ y;
    //! x = x << 1;
    //! x = x >> 1;
    //! x = x >>> 1;
    // Compound assignment:
    x += y;
    x -= y;
    x *= y;
    x /= y;
    x %= y;
    //! x <<= 1;
    //! x >>= 1;
    //! x >>>= 1;
    //! x &= y;
    //! x ^= y;
    //! x |= y;
    // Casting:
    //! boolean b = (boolean)x;
    char c = (char)x;
    byte B = (byte)x;
    short s = (short)x;
    int i = (int)x;
    long l = (long)x;
    float f = (float)x;
  }
} ///:~ 

Note that boolean is quite limited. You can assign to it the values true and false, and you can test it for truth or falsehood, but you cannot add booleans or perform any other type of operation on them.

In char, byte, and short you can see the effect of promotion with the arithmetic operators. Each arithmetic operation on any of those types results in an int result, which must be explicitly cast back to the original type (a narrowing conversion that might lose information) to assign back to that type. With int values, however, you do not need to cast, because everything is already an int. Don’t be lulled into thinking everything is safe, though. If you multiply two ints that are big enough, you’ll overflow the result. The following example demonstrates this:

//: Overflow.java
// Surprise! Java lets you overflow.

public class Overflow {
  public static void main(String[] args) {
    int big = 0x7fffffff; // max int value
    prt("big = " + big);
    int bigger = big * 4;
    prt("bigger = " + bigger);
  }
  static void prt(String s) {
    System.out.println(s);
  }
} ///:~ 

The output of this is:

big = 2147483647
bigger = -4 

and you get no errors or warnings from the compiler, and no exceptions at run-time. Java is good, but it’s not that good.

Compound assignments do not require casts for char, byte, or short, even though they are performing promotions that have the same results as the direct arithmetic operations. On the other hand, the lack of the cast certainly simplifies the code.

You can see that, with the exception of boolean, any primitive type can be cast to any other primitive type. Again, you must be aware of the effect of a narrowing conversion when casting to a smaller type, otherwise you might unknowingly lose information during the cast.


[16] John Kirkham writes, “I started computing in 1962 using FORTRAN II on an IBM 1620. At that time, and throughout the 1960s and into the 1970s, FORTRAN was an all uppercase language. This probably started because many of the early input devices were old teletype units that used 5 bit Baudot code, which had no lowercase capability. The ‘E’ in the exponential notation was also always upper case and was never confused with the natural logarithm base ‘e’, which is always lower case. The ‘E’ simply stood for exponential, which was for the base of the number system used – usually 10. At the time octal was also widely used by programmers. Although I never saw it used, if I had seen an octal number in exponential notation I would have considered it to be base 8. The first time I remember seeing an exponential using a lower case ‘e’ was in the late 1970s and I also found it confusing. The problem arose as lowercase crept into FORTRAN, not at its beginning. We actually had functions to use if you really wanted to use the natural logarithm base, but they were all uppercase.”

Contents | Prev | Next