Working with Floats

Nette\Utils\Floats is a static class containing useful functions for comparing floating-point numbers.

Installation:

composer require nette/utils

All examples assume the following class alias is defined:

use Nette\Utils\Floats;

Motivation

Are you wondering why we need a class for comparing floats? After all, you can use the operators <, >, ===, and you're done, right? Well, it's not entirely true. What do you think this code will output?

$a = 0.1 + 0.2;
$b = 0.3;
echo $a === $b ? 'same' : 'not same';

If you run the code, some of you might be surprised that the program outputs not same.

Mathematical operations with floating-point numbers can lead to precision errors due to the conversion between decimal and binary representations. For example, 0.1 + 0.2 results in something like 0.300000000000000044…. Therefore, when comparing floats, we need to tolerate a small difference, an epsilon.

And that's exactly what the Floats class does. The following comparison will now work as expected:

echo Floats::areEqual($a, $b) ? 'same' : 'not same'; // same

Trying to compare NAN throws a \LogicException.

The Floats class tolerates differences smaller than 1e-10. If you need to work with higher precision, consider using the BCMath library instead.

Float Comparison

areEqual (float $a, float $b): bool

Returns true if $a = $b.

Floats::areEqual(10, 10.0); // true

isLessThan (float $a, float $b): bool

Returns true if $a < $b.

Floats::isLessThan(9.5, 10.2); // true
Floats::isLessThan(INF, 10.2); // false

isLessThanOrEqualTo (float $a, float $b): bool

Returns true if $a <= $b.

Floats::isLessThanOrEqualTo(9.5, 10.2);    // true
Floats::isLessThanOrEqualTo(10.25, 10.25); // true

isGreaterThan (float $a, float $b): bool

Returns true if $a > $b.

Floats::isGreaterThan(9.5, -10.2); // true
Floats::isGreaterThan(9.5, 10.2);  // false

isGreaterThanOrEqualTo (float $a, float $b): bool

Returns true if $a >= $b.

Floats::isGreaterThanOrEqualTo(9.5, 10.2);  // false
Floats::isGreaterThanOrEqualTo(10.2, 10.2); // true

compare (float $a, float $b): int

Returns -1 if $a < $b, 0 if they are equal, and 1 if $a > $b.

It can be used, for example, with the usort() function.

$arr = [1, 5, 2, -3.5];
usort($arr, [Floats::class, 'compare']);
// $arr is now [-3.5, 1, 2, 5]

Helper Functions

isZero (float $value): bool

Returns true if the value is zero.

Floats::isZero(0.0); // true
Floats::isZero(0);   // true

isInteger (float $value): bool

Returns true if the value is an integer.

Floats::isInteger(0);    // true
Floats::isInteger(0.0);  // true
Floats::isInteger(-5.0); // true

Floats::isInteger(-5.1); // false
Floats::isInteger(INF);  // false
Floats::isInteger(NAN);  // false