Java Data Types - Numeric Data
Written by Alex Armstrong   
Wednesday, 06 February 2013
Article Index
Java Data Types - Numeric Data
Literals and Expressions
Types of Assignment and Casting
Numeric Output

Literals

In most cases we get numeric values into variables by initializing them using a literal - i.e. data that you write out. For example in:

int a=12345;

12345 is a numeric literal.

To be able to specify things correctly we need to specify the type of the literal

If you just type an integer it is assumed to be an int

If you end the an integer with an L then it assumed to be a long.

You can also specify integers in a number of different number bases. If you start a number off with a 0x then it is taken to be in hex and if you start it with 0b then it is binary. For example 0xA is ten in hex and 0b11 is three in binary.

For floating point numbers we have two type of literal. If you just type a value with a decimal point, 123.4 then is assumed to be a double. To specify a float you have to add a trailing f as in 123.4f.or to explicitly specify a double you can use a trailing d as in 123.4d

You can also specify floating point numbers using scientific notation. For example, 1E17.

Summary

  • There are two integer literals - int the default and long indicated by a trailing L as in 123L.
  • You can enter literals using decimal, the default, or hex, indicated by a leading 0x or binary, indicated by a leading 0b.
  • There are two floating point literals - the default double indicated by a trailing d e.g. 123,4d and float indicated by a trailing f e.g. 123.4f.
  • You can use scientific notation for floating point numbers using E to indicate the exponent.

Using Literals

When you initialize a variable you don't have to bother with setting the type of the literal - it is assumed to be the same type as the variable. Initialization always works unless you specify a value that is too big.

For example

byte myByte=42;

works even though the literal is by default an int. However if you do specify a type then you will see an error message:

byte myByte=42L;

For a long you can write either

long myLong=123;

or

long myLong=123L;

What this means is that the only time you need to use a long literal is when the value is too big to represent as an int.

When it comes to floating point the situation is a little strange. If you use:

float myFloat=123;

then it all works - the int is converted to a float. However if you use:

float myFloat=123.0;

you will see an error that says you are losing precision due to the conversion of a double to a float.

Any literal with a decimal point is assumed to be a double.

float myFloat=123.0f;

works.

Summary

  • When initializing a variable you generally only have to worry about long integers (when the value is too big for an int) and float.

Expressions

Of course usually the whole point of numbers is to do arithmetic. There isn't much to say about arithmetic expressions in general. The operators that you can use should be familiar to you from the chapter Java - Command Line Programs

  • + Additive operator
  • - Subtraction operator
  • * Multiplication operator
  • / Division operator

Notice that there is no raise to a power operator. If you want to raise a value to a power you have to use the Math class and its pow method which takes two doubles and raises the first to the second. For example:

double x=10.2;
double result=Math.pow(x,2.0);

will square the contents of x.

It is also worth noting that the division operator works at the level of the most precise type in the expression. That is

int myInt=3;
int result=myint/2;

stores 1 in result because this is integer division and any fractional parts are simply ignored. However

int myInt=3;
double result=myint/2.0;

attempts to do a floating point division and automatically converts myint to double and produces the answer 1.5.

Arithmetic with integers can become more complicated with respect to type.  The problem for integer arithmetic is that all arithmetic is by default done as int unless there is a long involved int he expression when it is done as long.

Let's look at an example and see how this can complicate things.

If you write:

byte myByte=10;
myByte=10+10;

Then everything works as you would expect and 20 is stored in myByte. However if you try the superficially similar:

byte myByte=10;
myByte=myByte+10;

The result is an error about possible loss of precision because the right hand side is an int.

The rule is that any expression that involves a variable is converted to an int and the problem occurs when you attempt to store the int back in the byte.

if you are writing arithmetic that involves only ints and are assigning to either an int or a long this is no problem as there is no loss of precision. If you are assigning to a byte or a short then you get an error, even if the expression involves only bytes or shorts. For example:

short myShort=10;
myShort=myShort+myShort;

generats and error because you are about to store an int in a short.

This loss of precision problem doesn't occur with floating point values because arithmetic is done either as float or double depending on the most precise type that occurs in the expression.



Last Updated ( Wednesday, 06 February 2013 )
 
 

   
RSS feed of all content
I Programmer - full contents
Copyright © 2014 i-programmer.info. All Rights Reserved.
Joomla! is Free Software released under the GNU/GPL License.