3.5 Integers and Floats

As we discussed in Chapter 2, PHP supports both integer and floating-point numbers. PHP stores integers as a 32-bit signed word, providing a range of integers from -2147483647 to +2147483647. PHP automatically converts numbers that overflow out of this range to floats. You can see this behavior by adding one to the largest integer value:

// Largest positive integer (for a 32 bit signed integer)

$variable =  2147483647;



// prints int(2147483647)

var_dump($variable);



$variable++;



// prints float(2147483648)

var_dump($variable);

Floating-point numbers can store a wide range of values, both very large and very small, by storing a mantissa and an exponent. However a floating-point number can't precisely represent all numbers—for example, the fraction 2/3—and some precision can be lost.

Integers can be represented in a decimal, hexadecimal, or octal notation:

$var = 42;        // a positive integer

$var = -186;      // a negative integer

$var = 0654;      // 428 expressed as an octal number

$var = 0xf7;      // 247 expressed as a hexadecimal number

Floating-point numbers can represented in a decimal or exponential notation:

$var = 42.0;      // a positive float

$var = -186.123;  // a negative float

$var = 1.2e65;    // a very big number

$var = 10e-75;    // a very small number

Apart from the basic operators +, -, /, *, and %, PHP provides the usual array of mathematical library functions. In this section, we present some of the library functions that are used with integer and float numbers.

3.5.1 Absolute Value

The absolute value of an integer or a float can be found with the abs( ) function:

integer abs(integer number)

float abs(float number)

The following examples show the result of abs( ) on integers and floats:

print abs(-1);       // prints 1

print abs(1);        // prints 1

print abs(-145.89);  // prints 145.89

print abs(145.89);   // prints 145.89

3.5.2 Ceiling and Floor

The ceil( ) and floor( ) functions return the integer value above and below a fractional value, respectively:

float ceil(float value)

float floor(float value)

The return type is a float because an integer may not be able to represent the result when a large value is passed as an argument. Consider the following examples:

print ceil(27.3);   // prints 28

print floor(27.3);  // prints 27

3.5.3 Rounding

The round( ) function uses 4/5 rounding rules to round up or down a value to a given precision:

float round(float value [, integer precision])

By default, rounding is to zero decimal places, but the precision can be specified with the optional precision argument. The 4/5 rounding rules determine if a number is rounded up or down based on the digits that are lost due to the rounding precision. For example, 10.4 rounds down to 10, and 10.5 rounds up to 11. Specifying a negative precision rounds a value to a magnitude greater than zero, for example a precision of -3 rounds a value to the nearest thousand. The following examples show rounding at various precisions:

print round(10.4);           // prints 10

print round(10.5);           // prints 11

print round(2.40964, 3);     // prints 2.410

print round(567234.56, -3);  // prints 567000

print round(567234.56, -4);  // prints 570000

3.5.4 Number Systems

PHP provides the following functions that convert numbers between integer decimal and the commonly used number systems, binary, octal, and hexadecimal:

string decbin(integer number)

integer bindec (string binarystring)

string dechex(integer number)

integer hexdec(string hexstring)

string decoct(integer number)

integer octdec(string octalstring)

The decimal numbers are always treated as integers, and the numbers in the other systems are treated as strings. Here are some examples:

print decbin(45);        // prints "101101"

print bindec("1001011"); // prints 75

print dechex(45);        // prints "2D"

print hexdec("5a7b");    // prints 23163

print decoct(45);        // prints "55"

print octdec("777");     // prints 511

It is possible to represent binary, octal, and hexadecimal numbers that are bigger than can be held in a 32-bit integer. The results of such conversions automatically overflow to a float value. For example:

// $a is an integer assigned the largest possible value

$a = hexdec("7fffffff");



// $a is a float

$a = hexdec("80000000");

3.5.5 Basic Trigonometry Functions

PHP supports the basic set of trigonometry functions listed in Table 3-5.

Table 3-5. Trigonometry functions supported by PHP

Function	Description
float sin(float arg)	Sine of arg in radians
float cos(float arg)	Cosine of arg in radians
float tan(float arg)	Tangent of arg in radians
float asin(float arg)	Arc sine of arg in radians
float acos(float arg)	Arc cosine of arg in radians
float atan(float arg)	Arc tangent of arg in radians
float atan2(float y, float x)	Arc tangent of x/y where the sign of both arguments determines the quadrant of the result
float pi( )	Returns the value 3.1415926535898
float deg2rad(float arg)	Converts arg degrees to radians
float rad2deg(float arg)	Converts arg radians to degrees

3.5.6 Powers and Logs

The PHP mathematical library includes the exponential and logarithmic functions listed in Table 3-6.

Table 3-6. Exponential and logarithmic functions

Function	Description
float exp(float arg)	e to the power of arg
float pow(float base, number exp)	Exponential expression base to the power of exp
float sqrt(float arg)	Square root of arg
float log(float arg [, float base])	Natural logarithm of arg, unless a base is specified, in which case the function returns log(arg)/log(base)
float log10(float arg)	Base-10 logarithm of arg

3.5.7 Testing Number Results

Many of the functions described in this section can return values that are undefined, or are too big or small to hold in a floating point number. PHP provides three functions that can be used to test numeric results before they cause problems later in a script:

boolean is_nan(float val)

boolean is_infinite(float val)

boolean is_finite(float val)

The function is_nan( ) tests the expression val and returns true if val is not a number. For example, the square root of a negative number is not a real number. is_finite( ) returns true if the number val can be represented as a valid float, and is_infinite( ) returns true if val can't. Here are some examples:

// square root of a negative number

$a = -1;

$result = sqrt($a);



print $result;    // prints -1.#IND



// Test if not a number

if (is_nan($result))

    print "Result not defined"; // prints

else

    print "Square root of {$a} = {$result}";

3.5.8 Random Number Generation

PHP provides the function rand( ), which returns values from a generated sequence of pseudo-random numbers. The sequence generated by rand( ) is pseudo random because the algorithm used appears to have random behavior but isn't truly random. The function rand( ) can be called in one of two ways:

integer rand( )

integer rand(integer min, integer max)

When called with no arguments, rand( ) returns a random number between 0 and the value returned by getrandmax( ) . When rand( ) is called with two arguments, the min and max values, the returned number is a random number between min and max. Consider an example:

// Generate some random numbers

print rand( );      // between 0 and getmaxrand( )

print rand(1, 6);  // between 1 and 6 (inclusive)

Prior to PHP 4.2.0, you needed to seed the random number generator with a call to srand( ) before the first use of rand( ), otherwise the function returns the same numbers each time a script is called. Since 4.2.0, the call to srand( ) is not required, however you can reliably reproduce a random sequence by calling srand( ) with the same argument at the start of the script. The following example reliably prints the same sequence of numbers each time it is called:

srand(123456);



// Prints six random numbers

for ($i=0; $i<6; $i++)

    print rand( ) . "   ";