PHP 8.4.0 RC2 available for testing

round

(PHP 4, PHP 5, PHP 7, PHP 8)

round浮動小数点数を丸める

説明

round(int|float $num, int $precision = 0, int $mode = PHP_ROUND_HALF_UP): float

num を、指定した precision(小数点以下の桁数)に丸めた値を 返します。precision を負またはゼロ(デフォルト) とすることも可能です。

パラメータ

num

丸める値。

precision

オプションで指定する、丸める桁数。

precision が正の値の場合、 num は 小数点以下、有効桁数 precision 桁に丸められます。

precision が負の場合、 num の 丸め処理は小数点より前の precision 桁に対して行われます。 つまり、一番近い pow(10, -$precision) の倍数に丸められます。 たとえば、precision が -1 の場合、 num は 一番近い10の倍数に、 precision が -2 の場合は 一番近い100の倍数に丸められる、という具合です。

mode

次の定数のいずれかを使って、丸めのモードを指定します。

定数 説明
PHP_ROUND_HALF_UP 端数が5(半分)の場合、 num をゼロから離れる方向に丸めます。 1.5 は 2 に、そして -1.5 は -2 になります。
PHP_ROUND_HALF_DOWN 端数が5(半分)の場合、 num をゼロに近づく方向に丸めます。 1.5 は 1 に、そして -1.5 は -1 になります。
PHP_ROUND_HALF_EVEN 端数が5(半分)の場合、 num をもっとも近い偶数に丸めます。 1.5 や 2.5 は 2 になります。
PHP_ROUND_HALF_ODD 端数が5(半分)の場合、 num をもっとも近い奇数に丸めます。 1.5 は 1 に、そして 2.5 は 3 になります。

戻り値

precision で与えられた精度に丸められた値を float で返します。

変更履歴

バージョン 説明
8.0.0 num は、 数値への変換をサポートした内部オブジェクトを受け入れなくなりました。

例1 round() の例

<?php
var_dump
(round(3.4));
var_dump(round(3.5));
var_dump(round(3.6));
var_dump(round(3.6, 0));
var_dump(round(5.045, 2));
var_dump(round(5.055, 2));
var_dump(round(345, -2));
var_dump(round(345, -3));
var_dump(round(678, -2));
var_dump(round(678, -3));
?>

上の例の出力は以下となります。

float(3)
float(4)
float(4)
float(4)
float(5.05)
float(5.06)
float(300)
float(0)
float(700)
float(1000)

例2 precision が、どうfloatの値に影響するか

<?php
$number
= 135.79;

var_dump(round($number, 3));
var_dump(round($number, 2));
var_dump(round($number, 1));
var_dump(round($number, 0));
var_dump(round($number, -1));
var_dump(round($number, -2));
var_dump(round($number, -3));
?>

上の例の出力は以下となります。

float(135.79)
float(135.79)
float(135.8)
float(136)
float(140)
float(100)
float(0)

例3 mode の例

<?php
echo 'Rounding modes with 9.5' . PHP_EOL;
var_dump(round(9.5, 0, PHP_ROUND_HALF_UP));
var_dump(round(9.5, 0, PHP_ROUND_HALF_DOWN));
var_dump(round(9.5, 0, PHP_ROUND_HALF_EVEN));
var_dump(round(9.5, 0, PHP_ROUND_HALF_ODD));

echo
PHP_EOL;
echo
'Rounding modes with 8.5' . PHP_EOL;
var_dump(round(8.5, 0, PHP_ROUND_HALF_UP));
var_dump(round(8.5, 0, PHP_ROUND_HALF_DOWN));
var_dump(round(8.5, 0, PHP_ROUND_HALF_EVEN));
var_dump(round(8.5, 0, PHP_ROUND_HALF_ODD));
?>

上の例の出力は以下となります。

Rounding modes with 9.5
float(10)
float(9)
float(10)
float(9)

Rounding modes with 8.5
float(9)
float(8)
float(8)
float(9)

例4 桁数を指定した mode の例

<?php
echo 'Using PHP_ROUND_HALF_UP with 1 decimal digit precision' . PHP_EOL;
var_dump(round( 1.55, 1, PHP_ROUND_HALF_UP));
var_dump(round(-1.55, 1, PHP_ROUND_HALF_UP));

echo
PHP_EOL;
echo
'Using PHP_ROUND_HALF_DOWN with 1 decimal digit precision' . PHP_EOL;
var_dump(round( 1.55, 1, PHP_ROUND_HALF_DOWN));
var_dump(round(-1.55, 1, PHP_ROUND_HALF_DOWN));

echo
PHP_EOL;
echo
'Using PHP_ROUND_HALF_EVEN with 1 decimal digit precision' . PHP_EOL;
var_dump(round( 1.55, 1, PHP_ROUND_HALF_EVEN));
var_dump(round(-1.55, 1, PHP_ROUND_HALF_EVEN));

echo
PHP_EOL;
echo
'Using PHP_ROUND_HALF_ODD with 1 decimal digit precision' . PHP_EOL;
var_dump(round( 1.55, 1, PHP_ROUND_HALF_ODD));
var_dump(round(-1.55, 1, PHP_ROUND_HALF_ODD));
?>

上の例の出力は以下となります。

Using PHP_ROUND_HALF_UP with 1 decimal digit precision
float(1.6)
float(-1.6)

Using PHP_ROUND_HALF_DOWN with 1 decimal digit precision
float(1.5)
float(-1.5)

Using PHP_ROUND_HALF_EVEN with 1 decimal digit precision
float(1.6)
float(-1.6)

Using PHP_ROUND_HALF_ODD with 1 decimal digit precision
float(1.5)
float(-1.5)

参考

  • ceil() - 端数の切り上げ
  • floor() - 端数の切り捨て
  • number_format() - 数字を千の位毎にグループ化してフォーマットする

add a note

User Contributed Notes 29 notes

up
319
takingsides at gmail dot com
10 years ago
In my opinion this function lacks two flags:

- PHP_ROUND_UP - Always round up.
- PHP_ROUND_DOWN - Always round down.

In accounting, it's often necessary to always round up, or down to a precision of thousandths.

<?php
function round_up($number, $precision = 2)
{
$fig = (int) str_pad('1', $precision, '0');
return (
ceil($number * $fig) / $fig);
}

function
round_down($number, $precision = 2)
{
$fig = (int) str_pad('1', $precision, '0');
return (
floor($number * $fig) / $fig);
}
?>
up
33
depaula at unilogica dot com
7 years ago
As PHP doesn't have a a native number truncate function, this is my solution - a function that can be usefull if you need truncate instead round a number.

<?php
/**
* Truncate a float number, example: <code>truncate(-1.49999, 2); // returns -1.49
* truncate(.49999, 3); // returns 0.499
* </code>
* @param float $val Float number to be truncate
* @param int f Number of precision
* @return float
*/
function truncate($val, $f="0")
{
if((
$p = strpos($val, '.')) !== false) {
$val = floatval(substr($val, 0, $p + 1 + $f));
}
return
$val;
}
?>

Originally posted in http://stackoverflow.com/a/12710283/1596489
up
29
slimusgm at gmail dot com
10 years ago
If you have negative zero and you need return positive number simple add +0:

$number = -2.38419e-07;
var_dump(round($number,1));//float(-0)
var_dump(round($number,1) + 0);//float(0)
up
9
Mojo urk
6 years ago
Solving round_down() problem:
-----------------------------
Use of <?php floor(pow(10, $precision) * $value) / pow(10, $precision); ?> fails in some cases, e.g. round_down(2.05, 2) gives incorrect 2.04.
Here is a "string" solution (https://stackoverflow.com/a/26491492/1245149) of the problem (a negative precision is not covered):

<?php
function round_down($value, $precision) {
$value = (float)$value;
$precision = (int)$precision;
if (
$precision < 0) {
$precision = 0;
}
$decPointPosition = strpos($value, '.');
if (
$decPointPosition === false) {
return
$value;
}
return (float)(
substr($value, 0, $decPointPosition + $precision + 1));
}
?>

Solving round_up() problem:
---------------------------
Use of <?php ceil(pow(10, $precision) * $value) / pow(10, $precision);?> fails in some cases, e.g. round_up(2.22, 2) gives incorrect 2.23 (https://stackoverflow.com/a/8239620/1245149).
Adapting the above round_down() "string" solution I have got this result (a negative precision is not covered):

<?php
function round_up($value, $precision) {
$value = (float)$value;
$precision = (int)$precision;
if (
$precision < 0) {
$precision = 0;
}
$decPointPosition = strpos($value, '.');
if (
$decPointPosition === false) {
return
$value;
}
$floorValue = (float)(substr($value, 0, $decPointPosition + $precision + 1));
$followingDecimals = (int)substr($value, $decPointPosition + $precision + 1);
if (
$followingDecimals) {
$ceilValue = $floorValue + pow(10, -$precision); // does this give always right result?
}
else {
$ceilValue = $floorValue;
}
return
$ceilValue;
}
?>

I don't know it is bulletproof, but at least it removes the above mentioned fail. I have done no binary-to-decimal-math-analysis but if `$floorValue + pow(10, 0 - $precision)` works
always as expected then it should be ok.
up
12
serg at kalachev dot ru
10 years ago
Excel-like ROUNDUP function:

public static function round_up($value, $places)
{
$mult = pow(10, abs($places));
return $places < 0 ?
ceil($value / $mult) * $mult :
ceil($value * $mult) / $mult;
}

echo round_up(12345.23, 1); // 12345.3
echo round_up(12345.23, 0); // 12346
echo round_up(12345.23, -1); // 12350
echo round_up(12345.23, -2); // 12400
echo round_up(12345.23, -3); // 13000
echo round_up(12345.23, -4); // 20000
up
12
jongbumi at gmail dot com
8 years ago
PHP 5.3, 5.4, 5.5
<?php
$fInfinty
= pow(1000, 1000); // float(INF)
$fResult = round(123.456, $fInfinty); // double(123)
?>

PHP 5.6
<?php
$fInfinty
= pow(1000, 1000); // float(INF)
$fResult = round(123.456, $fInfinty); // float(0)
?>

PHP 7
<?php
$fInfinty
= pow(1000, 1000); // float(INF)
$fResult = round(123.456, $fInfinty); // null
?>
up
21
djcox99 at googlemail dot com
10 years ago
I discovered that under some conditions you can get rounding errors with round when converting the number to a string afterwards.

To fix this I swapped round() for number_format().

Unfortunately i cant give an example (because the number cant be represented as a string !)

essentially I had round(0.688888889,2);

which would stay as 0.68888889 when printed as a string.

But using number_format it correctly became 0.69.
up
18
esion99 at gmail dot com
10 years ago
Unexpected result or misunderstanding (php v5.5.9)

<?php

echo round(1.55, 1, PHP_ROUND_HALF_DOWN); // 1.5
echo round(1.551, 1, PHP_ROUND_HALF_DOWN); //1.6

?>
up
10
christian at deligant dot net
13 years ago
this function (as all mathematical operators) takes care of the setlocale setting, resulting in some weirdness when using the result where the english math notation is expected, as the printout of the result in a width: style attribute!

<?php
$a
=3/4;
echo
round($a, 2); // 0.75

setlocale(LC_ALL, 'it_IT@euro', 'it_IT', 'it');
$b=3/4;
echo
round($b,2); // 0,75
?>
up
20
twan at ecreation dot nl
24 years ago
If you'd only want to round for displaying variables (not for calculating on the rounded result) then you should use printf with the float:

<?php printf ("%6.2f",3.39532); ?>

This returns: 3.40 .
up
9
craft at ckdevelop dot org
10 years ago
function mround($val, $f=2, $d=6){
return sprintf("%".$d.".".$f."f", $val);
}

echo mround(34.89999); //34.90
up
16
Anonymous
14 years ago
Here is function that rounds to a specified increment, but always up. I had to use it for price adjustment that always went up to $5 increments.

<?php
function roundUpTo($number, $increments) {
$increments = 1 / $increments;
return (
ceil($number * $increments) / $increments);
}
?>
up
7
michaeldnelson dot mdn at gmail dot com
15 years ago
This function will let you round to an arbitrary non-zero number. Zero of course causes a division by zero.

<?php
function roundTo($number, $to){
return
round($number/$to, 0)* $to;
}

echo
roundTo(87.23, 20); //80
echo roundTo(-87.23, 20); //-80
echo roundTo(87.23, .25); //87.25
echo roundTo(.23, .25); //.25
?>
up
3
greghenle at gmail dot com
7 years ago
/**
* Round to first significant digit
* +N to +infinity
* -N to -infinity
*
*/
function round1stSignificant ( $N ) {
if ( $N === 0 ) {
return 0;
}

$x = floor ( log10 ( abs( $N ) ) );

return ( $N > 0 )
? ceil( $N * pow ( 10, $x * -1 ) ) * pow( 10, $x )
: floor( $N * pow ( 10, $x * -1 ) ) * pow( 10, $x );
}

echo round1stSignificant( 39144818 ) . PHP_EOL;
echo round1stSignificant( 124818 ) . PHP_EOL;
echo round1stSignificant( 0.07468 ) . PHP_EOL;
echo round1stSignificant( 0 ) . PHP_EOL;
echo round1stSignificant( -0.07468 ) . PHP_EOL;

/**
* Output
*
* 40000000
* 200000
* 0.08
* 0
* -0.08
*
*/
up
4
dastra
12 years ago
round() will sometimes return E notation when rounding a float when the amount is small enough - see https://bugs.php.net/bug.php?id=44223 . Apparently it's a feature.

To work around this "feature" when converting to a string, surround your round statement with an sprintf:

sprintf("%.10f", round( $amountToBeRounded, 10));
up
9
php at silisoftware dot com
22 years ago
Here's a function to round to an arbitary number of significant digits. Don't confuse it with rounding to a negative precision - that counts back from the decimal point, this function counts forward from the Most Significant Digit.

ex:

<?php
round
(1241757, -3); // 1242000
RoundSigDigs(1241757, 3); // 1240000
?>

Works on negative numbers too. $sigdigs should be >= 0

<?php
function RoundSigDigs($number, $sigdigs) {
$multiplier = 1;
while (
$number < 0.1) {
$number *= 10;
$multiplier /= 10;
}
while (
$number >= 1) {
$number /= 10;
$multiplier *= 10;
}
return
round($number, $sigdigs) * $multiplier;
}
?>
up
8
martinr at maarja dot net
16 years ago
Please note that the format of this functions output also depends on your locale settings. For example, if you have set your locale to some country that uses commas to separate decimal places, the output of this function also uses commas instead of dots.

This might be a problem when you are feeding the rounded float number into a database, which requires you to separate decimal places with dots.

See it in action:
<?php
echo round('3.5558', 2);
setlocale(constant('LC_ALL'), 'et_EE.UTF-8');
echo
'<br />'. round('3.5558', 2);
?>

The output will be:
3.56
3,56
up
1
Anonymous
7 years ago
Note that PHP 5.3 didn't just introduce $mode, it rewrote the rounding implementation completely to eliminate many kinds of rounding errors common to rounding floating point values.

That's why round() gives you the correct result even when floor/ceil don't.
For example, floor(0.285 * 100 + 0.5) VS round(0.285*100 + 0.5). First one gives 28, second one gives 29.

More details here: https://wiki.php.net/rfc/rounding
up
1
feha at vision dot to
14 years ago
Here is a short neat function to round minutes (hour) ...

<?php

function minutes_round ($hour = '14:03:32', $minutes = '5', $format = "H:i")
{
// by Femi Hasani [www.vision.to]
$seconds = strtotime($hour);
$rounded = round($seconds / ($minutes * 60)) * ($minutes * 60);
return
date($format, $rounded);
}

?>

You decide to round to nearest minute ...
example will produce : 14:05
up
0
goreyshi at gmail dot com
6 years ago
When you have a deal with money like dollars, you need to display it under this condition:
-format all number with two digit decimal for cents.
-divide 1000 by ,
-round half down for number with more than two decimal

I approach it using round function inside the number_format function:

number_format((float)round( 625.371 ,2, PHP_ROUND_HALF_DOWN),2,'.',',') // 625.37
number_format((float)round( 625.379 ,2, PHP_ROUND_HALF_DOWN),2,'.',',') // 625.38
number_format((float)round( 1211.20 ,2, PHP_ROUND_HALF_DOWN),2,'.',',') // 1,211.20
number_format((float)round( 625 ,2, PHP_ROUND_HALF_DOWN),2,'.',',') // 625.00
up
-2
omnibus at omnibus dot edu dot pl
13 years ago
Beware strange behaviour if number is negative and precision is bigger than the actual number of digits after comma.

round(-0.07, 4);

returns

-0.07000000000000001

So if you validate it against a regular expression requiring the maximum amount of digits after comma, you'll get into trouble.
up
-1
Astro
4 years ago
Okay, final version of my function:

function NumberPrecision($n, $precision=0, $is_round=true)
{
if ($is_round)
{
$r = 5 * pow(10, -($precision+1));
$n += (($n < 0) ? -$r : $r);
}

$comma = '.';

$r = 5 * pow(10, -($precision+2));
$n += (($n > 0) ? -$r : $r);
$n = number_format($n, $precision+1, $comma, '');

$n .= $comma;
list($n, $frac) = explode($comma, $n, 2);
$n = rtrim(rtrim($n, $comma) . $comma . substr($frac, 0, $precision), $comma);
return ($n);
}

It can be useful in come cases when built-in function like round() or number_format() returns unexpected results. Works with positive and negative numbers, zero, numbers like 1/12, 0.3, numbers in scientific notation etc.
up
-1
terry at scribendi dot com
20 years ago
To round any number to a given number of significant digits, use log10 to find out its magnitude:

<?php round($n, ceil(0 - log10($n)) + $sigdigits); ?>

Or when you have to display a per-unit price which may work out to be less than a few cents/pence/yen you can use:

<?php
// $exp = currency decimal places - 0 for Yen/Won, 2 for most others
$dp = ceil(0 - log10($n)) + $sigdigits;
$display = number_format($amount, ($exp>$dp)?$exp:$dp);
?>

This always displays at least the number of decimal places required by the currency, but more if displaying the unit price with precision requires it - eg: 'English proofreading from $0.0068 per word', 'English beer from $6.80 per pint'.
up
-2
Anonymous
14 years ago
This functions return ceil($nb) if the double or float value is bigger than "$nb.5" else it's return floor($nb)

<?php
function arounds_int($nb) {

if(!
is_numeric($nb)) {
return
false;
}

$sup = round($nb);
$inf = floor($nb);
$try = (double) $inf . '.5' ;

if(
$nb > $try) {
return
$sup;
}

return
$inf;
}
?>
up
-3
Anonymous
7 years ago
This function has strange. behaviors:

<?php
echo round(0.045, 2); // 0.05
echo round(0.45, 1); // 0.5
echo round(1.045-1, 2); // 0.04 !!!
echo round(1.45-1, 1); // 0.5
up
-1
Hayley Watson
5 years ago
It should just be noted that what is called "precision" on this page is more correctly called accuracy; precision is the total number of significant digits on both sides of the decimal point, while accuracy is the number of digits to the right of the point. It's a common confusion.
up
-3
spectrumcat at gmail dot com
10 years ago
In case someone will need a "graceful" rounding (that changes it's precision to get a non 0 value) here's a simple function:

function gracefulRound($val, $min = 2, $max = 4) {
$result = round($val, $min);
if ($result == 0 && $min < $max) {
return gracefulRound($val, ++$min, $max);
} else {
return $result;
}
}

Usage:
$_ = array(0.5, 0.023, 0.008, 0.0007, 0.000079, 0.0000048);
foreach ($_ as $val) {
echo "{$val}: ".gracefulRound($val)."\n";
}

Output:
0.5: 0.5
0.023: 0.02
0.008: 0.01
0.0007: 0.001
0.000079: 0.0001
0.0000048: 0
up
-4
maxteiber at gmail dot com
17 years ago
the result of this function always depends on the underlying C function. There have been a lot of compiler bugs and floating-point precission problems involving this function. Right now the following code:

<?php
echo round(141.075, 2);
?>

returns:

141.07

on my machine.
So never really trust this function when you do critical calculations like accounting stuff!
Instead: use only integers or use string comparisons.
up
-3
php at persignum dot com
9 years ago
Because this function is missing round up and round down constants and the top note doesn't really show you how to round up or down to the nearest number, here is an easy way to always round up or always round down to the nearest number.

int is the number you want to round

n is the nearest number you want rounded to.

Round up to the nearest number

function round_up($int, $n) {
return ceil($int / $n) * $n;
}

And to round down to the nearest number

function round_down(int, $n) {
return floor($int / $n) * $n;
}
To Top