Download Primitives - CIS @ UPenn

Introduction to Primitives
30-Apr-17
Overview

Today we will discuss:





The eight primitive types,
especially int and double
Declaring the types of variables
Operations on primitives
The assignment statement
How to print results
2
Primitives


Primitives are the “basic” data values
There are eight types of primitives:




boolean -- used for true and false values
char -- used for single characters (letters, etc.)
byte, short, int, long -- four different kinds of integer
(whole number) values
float, double -- two different kinds of decimal numbers
(numbers with a decimal point)
3
int

The most important integer type is int


Numbers occupy memory in the computer




An int is a “whole” number (no decimal point)
Larger numeric types require more memory
byte: 1 byte short: 2 bytes int: 4 bytes long: 8 bytes
An int can be between about two billion (two thousand million)
and negative two billion
If you just write a number, such as 25, Java assumes it is
an int
 Hence it is easier to work with int values than with the other
integer types (byte, short, and long)
Use int in preference to other integer types
4
byte and short



A byte can be between -128 and 127
A short can be -32768 to 32767
Why these numbers?


These are “round numbers” in binary; for example,
0111 1111 1111 1111 is binary for 32767
1000 0000 0000 0000 is binary for -32768
The first bit is the sign bit: a 1 means it’s a negative
number
Use byte or short only when


You know the numbers are all small
There are millions of numbers to remember
5
long

long integers are for when two billion isn’t large enough
for your needs





A long can be as long as about 19 digits
A long occupies twice as much space as an int
Arithmetic on long values is slower
Use long only when you need really big numbers
A long number is written like an int, but with an L suffix


Examples: 895000000000L, 25L, 5L, -123L
Even larger numbers are available in Java-but they are objects, not primitives
6
double

A double represents a “real” number



If you just write a real number, such as 1.37, Java assumes it is a
double



Hence it is easier to work with double values than with float values
doubles can also be written using scientific notation, where E
stands for “times ten to the power”


Also sometimes called “floating point”
These are numbers with a decimal point
Examples: 7.3E5 = 730000.0, 7.3E-5 = 0.000073
A double can represent numbers in the range 4.9*10-324 to
1.8*10308 and has 15 or 16 digits of accuracy
Use double in preference to float
7
float




float is the other kind of “real,” or “floating point” number
float has about 8 digits of accuracy
Arithmetic with float is not faster
Use float only to save space when there are millions of numbers
involved

A float constant is written like a double constant, but with an F
suffix


Examples: 1.23F, -9F, 6.7E14F
A float can represent numbers in the range 1.4*10-45 to 3.4*1038
and has 6 or 7 digits of accuracy
8
An aside: approximations

Integers are precise, but real numbers are always
approximate (inaccurate)


Computers always use the binary system internally
Many numbers that can be expressed precisely in decimal
cannot be represented precisely in binary



For example, the numbers 1.1, 1.2, 1.3, and 1.4 can only be
approximated in binary
Two numbers that look the same may actually be subtly
different
Never test floating point numbers for equality!


Only test for larger or smaller, or for “not larger” or “not
smaller”
This is not a Java rule—it’s a programming rule
9
Giving names to numbers

Sometimes you know what a number is






You have 10 fingers
There are 24 hours in a day
π is 3.141592653589793238
Numbers written like this are called literals
You can use literals anyplace in Java that you can use a number
It’s usually better to use names instead:


classSize, myBankBalance, myAge, speedometerReading
A literal used without explanation is called a magic number





Example: e = 8.987551787E16 * m;
Better: energy = speedOfLightSquared * mass;
Better yet: final double C = 2.99792458E8; e = m * C * C;
Style rule: Avoid magic numbers!
Sometimes names are simply more convenient, for example,
Math.PI instead of 3.141592653589793238
10
Variables

Before you use a variable, you must declare it (tell Java
what type it is: int, double, char, ...)

There are two reasons for this:



Different types require different amounts of space
So Java can prevent you from doing something meaningless (adding 5
and true, or multiplying two dates together)
Before you use a variable, you must also define it (tell Java
what value it has)


It makes no sense to print out your bankBalance, or to add 100.00
to your bankBalance, if you don’t have a meaningful, well-defined
initial value for bankBalance to start with
You might assign an initial value to your variable, or compute a
value, or read a value in; but you have to get one somehow
11
Declaring variables

You declare variables like this:
int classSize;
double myBankBalance;

When you declare a variable to be a primitive
type, Java automatically finds space for it


The amount of space Java needs to find depends on the
type of the variable
Think of a variable as a specially shaped “box,”
designed to hold a value of a particular type


An int variable is four bytes long and there’s a special place
for the sign bit
A float variable is also four bytes long, but the bits are used
differently--some are used to tell where the decimal point goes
12
Giving values to variables

A variable is just a name for some value



You have to supply the actual value somehow
Java tries to prevent you from using a variable that you
haven’t given a value
You can assign values like this:
classSize = 57;
myBankBalance = 123.01; // no "$"!
13
Number “width”

Numeric types are considered wider or narrower
than other numeric types




This is based partly on number of bytes occupied
Also based on how large a number it can hold
Java doesn’t mind if you assign a narrow value to a
wide variable: int n = 3;
Java is not happy if you assign a wide value to a
narrow variable: byte b = 7139946; // illegal
14
Widening and narrowing

byte
short
int
long
float
double

You can always widen
(assign a narrower type to a
wider type):
double wide;
int narrow;
wide = narrow;
But if you want to narrow
(assign a wider type to a
narrower type), you have to
cast it:
narrow = (int)wide;
15
Casts


You can convert (cast) one numeric type to another
When you widen, no explicit cast is necessary


When you narrow, an explicit cast is required



But it doesn’t hurt
This requirement is made to help avoid accidental loss of
precision
Casting tells Java that the value in the wider type
will fit in the narrower type
Java checks to make sure that the cast works, and
gives you an error if it didn’t
16
Example casts
short s = 0;
int i = 0;
double d = 0.0;
d = i; // legal
d = s; // legal
i = s; //legal
i = d; // illegal
s = d; // illegal
s = i; // illegal
i = (int) d;
// legal
s = (short) d; // legal
s = (short) i; // legal
d = 3.7E20;
i = 50000;
// The following give
// runtime errors:
s = (short) i;
i = (int) d;
17
The fifth integer type

The primitive type char refers to a single, two-byte Unicode
character



You can use characters in arithmetic (they will automatically be
converted to int)



There is no good reason this should be a numeric type...
...but characters were numbers in C
char ch = 'A';
char ch2 = (char) (ch + 1); // cast result back to char
System.out.println(ch + " " + ch2 + " " + (ch + 1));
A B 66
To assign a char to a byte, or a byte to a char, you must use a
cast
18
Mixed types

If you mix numeric types, the narrower type is
automatically promoted (widened) to the wider type


int narrow = 5;
double wide;
double anotherWide = wide + narrow;
Integer division is when you divide one integer type by
another



The fractional part is discarded
Example: narrow = 19 / 5; // result is 3
Example: narrow = -19 / 5; // result is -3
19
Math methods

Converting a double to an int just discards the fractional part:
(int)17.93 is 17
(int) –17.93 is -17

double Math.floor(double)


Given a double, returns (as a double) the largest integral value not
greater than the argument
Math.floor(17.93) returns 17.0
Math.floor(-17.93) returns –18.0
double Math.ceil(double)

Given a double, returns (as a double) the smallest integral value not
smaller than the argument
Math.ceil(17.93) returns 18.0
Math.ceil(-17.93) returns –17.0
20
Method parameters

When you send a message to an object with a numeric
parameter, and the object needs to promote the
parameter in order to use a method, it will do so


Example:
 double twice(double n) { return 2.0 * n; }
 twice(5) returns 10.0
This promotion will only occur if necessary

Example 2:
 double half(double n) { return n / 2; }
 int half(int n) { return n / 2; }
 half(25) returns 12
21
The End
“I think there is a world market for maybe five
computers.”
--Thomas Watson, chairman of IBM, 1943
22