Return to Tutorial Index

Back to the home page of this site

morterboard

Java2 Certification
Tutorial

 

You can discuss this topic with others at http://www.jchq.net/discus
Read reviews and buy a Java Certification book at http://www.jchq.net/bookreviews/jcertbooks.htm


4) Language Fundamentals

Objective 5)

State the range of all primitive data types and declare literal values for String and all primitive types using all permitted formats, bases and representations.

Note on this objective

This is one of the slightly annoying but fairly easy to cover objectives. You can write a large amount of Java without knowing the range of primitive types but it should not take long to memorise these details. Beware of the requirement to be able to use all formats, don't overlook the octal form

The size of integral primitives

When this objective asks for the range of primitive data types I assume it is only required as representing the number 2 raised to the appropriate power rather than the actual number this represents. In my brain there are only three integral types to learn as the size of a byte is intuitively, in my PC based experience, 8 bits.

 

Range of Integral Primitives
Name Size Range
byte 8 bit -27 to 2 7-1
short 16 bit -215 to 215-1
int 32 bit -231 to 231-1
long 64 bit -2 63 to 2 63-1

Declaring integral literals

There are three ways to declare an integral literal. The default, as you might expect is decimal. Here are the options

Declaring 18 as an integral literal
Decimal 18
Octal 022 (Zero not letter O)
Hexidecimal 0x12

 

If you compile and run this little class you will get the value 18 output each time.

public class Lit{
public static void main(String[] argv){
        int i = 18;
        int j = 022;//Octal version: Two eights plus two 
        int k = 0x12;//Hex version: One sixteen plus two
        System.out.println(i);
        System.out.println(j);
        System.out.println(k);
        }
}

Roberts and Heller describe 6 ways of declaring integral literals, because unusually for Java letter X is not case sensitive, nor are the letters A through F for hexadecimal notation. I find it easier to remember the three ways and that it the letters are not case sensitive.

The size of floating point primitives

Floating point numbers are slightly strange beasts as calculations can have some unexpected results. Thus to quote Peter Van Der Linden "The exact accuracy depends on the number being represented". As compensation for this variable accuracy, you do get to play with numbers large almost beyond imagination. Thus the largest double can store a number that corresponds to 17 followed by 307 zeros. So you can even store the value of the financial paper worth of Bill Gates (until Linux reaches reaches total world domination, then an integer may do the job nicely).

Range of floating point types
float 32 bit
double 64 bit

Bear in mind that the default type for a literal with a decimal component is a double and not a float. This is slightly confusing as you might think that the default type for a "floating point number" would be a float. You may get questions in the exam in a similar form to the following.

Will the following compile?

float i = 1.0;

Intuition would tell you that this should compile cleanly. Unfortunately the exam is not designed to test your intuition. This will cause a compile time error because it attempts to assign a double to a float type. You can fix this code as follows

float i = 1.0F;

Or even

float i = (float) 1.0;

Indicating data type with a trailing letter

As demonstrated in the previous section you can tell Java that a numeric literal is of a particular type by giving it a trailing letter. These following are available

Suffix to indicate Data type
float F
long L
double D

boolean and char

The boolean and char primitives are a little odd. If you have a background in C/C++ pay attention particularly to the boolean and make sure you do not bring any "false friends" from these languages. A boolean can not be assigned any other value than true or false. The values true or false do not correspond to 0, -1 or any other number.

A boolean can only be true or false, it cannot be assigned a number such as -1 or 0

The char primitive is the only unsigned primitive in Java, and is 16 bits long. The char type can be used to denote a Unicode character. Unicode is an alternative to ASCII that stores characters in 2 bytes instead of the 1 byte of ASCII. This gives you 65K worth of characters, which although not enough to cover all world scripts, is an improvement of the 255 characters of ASCII. Internationalisation is a whole subject on its own, and just because you can represent characters from Chinese or Vietnamese, it does not mean that they will display correctly if you have a standard English style operating system.

A char literal can be created by enclosing the character in single quotes thus

char a = 'z';

Note that single quotes ' are used not double ".

This works fine in my English centered little world but as Java is a world system a char may contain any of the characters available in the Unicode system. This is done by using four hex digits preceded by \u, with the whole expression in single quotes.

Thus the space character can be represented as follows

char c = ‘\u0020’ 

If you assign a plain number to a char it can be output as a alphabetic character. Thus the following will print out the letter A (ASCII value 65) and a space.

public class MyChar{
public static void main(String argv[]){
        char i = 65;
        char c = '\u0020';
        System.out.println(i);
	System.out.println("This"+c+"Is a space");
        }
}

Declaring string literals

The String type is not a primitive but it is so important that in certain areas Java treats it like one. One of these features is the ability to declare String literals instead of using new to instantiate a copy of the class.

String literals are fairly straightforward. Make sure you remember that String literals are enclosed in double quotes whereas a char literal takes single quotes.

Thus

String name = "James Bond" 

See Objective 9.3 and 5.2 for more on the String class.


 

Question 1)


Which of the following will compile correctly?

1) float f=10f;
2) float f=10.1;
3) float f=10.1f;
4) byte b=10b;



Question 2)

Which of the following will compile correctly?

1) short myshort=99S;
2) String name='Excellent tutorial Mr Green';
3) char c=17c;
4) int z=015;



Question 3)

Which of the following will compile correctly?

1) boolean b=-1;
2) boolean b2=false;
3) int i=019;
4) char c=99;

Answers

Answer 1)

1) float f=10f;

3) float f=10.1f;

There is no such thing as a byte literal and option 2 will cause an error because the default type for a number with a decimal component is a double.


Answer 2)

4)int z=015;

The letters c and s do not exist as literal indicators and a String must be enclosed with double quotes, not single as in this example.


Answer 3)

2) boolean b2=false;
4) char c=99;

Option 1 should be fairly obvious as wrong, as a boolean can only be assigned the values true of false. Option 3 is slightly more tricky as this is the correct way to declare an octal literal but you cannot use the numeric 9 if you are in base 8 where you have numbers 0 through 7. A little tricky one there perhaps.

Other sources on this topic

This topic is covered in the Sun Tutorial at
http://java.sun.com/docs/books/tutorial/java/nutsandbolts/datatypes.html

The JLS
http://java.sun.com/docs/books/jls/second_edition/html/typesValues.doc.html#9151

Jyothi Krishnan on this topic at
http://www.geocities.com/SiliconValley/Network/3693/obj_sec4.html#obj13

Bruce Eckel's Thinking in Java
Chapter 2 "Special case: Primitive Types"
Chapter 3 "Literals"

Last updated
23 Aug 2001
copyright © Marcus Green 2001