Math 函数
在线手册：中文 英文
PHP手册

floor

(PHP 4, PHP 5)

floor — 舍去法取整

说明

float floor ( float $value )

返回不大于 value 的下一个整数，将 value 的小数部分舍去取整。floor() 返回的类型仍然是 float，因为 float 值的范围通常比 integer 要大。

Example #1 floor() 例子

<?php
echo floor(4.3);   // 4
echo floor(9.999); // 9
?>

参见 ceil() 和 round()。

Math 函数
在线手册：中文 英文
PHP手册
PHP手册 - N: 舍去法取整

用户评论:

clchuah at apic dot com dot sg (24-Mar-2012 04:37)

I have this little truncate float function

function truncate1 ($num, $digits = 0) {
  $shift = pow(10 , $digits);
  return ((floor($num * $shift)) / $shift);
}

I hit a weird floor() behavior problem recently, it took very long time to debug, see this:

$e_num = 25/24;
echo $e_num." orig\n";   // get 1.0416666666667

$rnd = round($e_num, 6);
echo $rnd." round\n";   // get 1.041667

$trc = truncate1($rnd, 6);
echo $trc." truncate\n";   // get 1.041666 !! expected  1.041667

$mul = $rnd * 1000000;
echo $mul." multiply\n";   // get 1041667

$flr = floor($mul);
echo $flr." floor\n";   // get 1041666, expected 1041667

I think because $rnd is kept as float so it's not precisely 1.041667. So i tested this further

$flr1 = floor(1.041667 * 1000000);
echo $flr1." floor1\n";   // get 1041666

$flr2 = floor(1041667);
echo $flr2." floor2\n";   // get 1041667

So beware of float type !!!

dr+php at dakotaryan dot com (21-Sep-2011 07:22)

The better way to round up/down to the nearest hundred (or thousand, etc.) is with roundBetter():

<?php
// This function is backwards compatible with round(),
// but provides an optional extra argument which,
// if set and not null, allows you to override the default
// rounding direction. See examples below.
function roundBetter($number, $precision = 0, $mode = PHP_ROUND_HALF_UP, $direction = NULL) {   
    if (!isset($direction) || is_null($direction)) {
        return round($number, $precision, $mode);
    }
   
    else {
        $factor = pow(10, -1 * $precision);

        return strtolower(substr($direction, 0, 1)) == 'd'
            ? floor($number / $factor) * $factor
            : ceil($number / $factor) * $factor;
    }
}

// roundBetterUp(1999, -3) => 2000
// roundBetterUp(1001, -3) => 2000
function roundBetterUp($number, $precision = 0, $mode = PHP_ROUND_HALF_UP) {
    return roundBetter($number, $precision, $mode, 'up');
}

// roundBetterDown(1999, -3) => 1000
// roundBetterDown(1001, -3) => 1000
function roundBetterDown($number, $precision = 0, $mode = PHP_ROUND_HALF_UP) {
    return roundBetter($number, $precision, $mode, 'down');
}
?>

seppili_ at gmx dot de (05-Mar-2010 04:52)

I use this function to floor with decimals:
<?php

function floordec($zahl,$decimals=2){   
     return floor($zahl*pow(10,$decimals))/pow(10,$decimals);
}
?>

lewis at lewishenshall dot co dot uk (26-Feb-2009 07:12)

If you're wanting to round down to the nearest hundred:

<?php
function rounddowntohundred($theNumber) {
    if (strlen($theNumber)<3) {
    $theNumber=$theNumber;
    } else {
    $theNumber=substr($theNumber, 0, strlen($theNumber)-2) . "00";

    }
    return $theNumber;

}
?>

fragov at gmail dot com (15-Jan-2009 05:39)

Have solved a "price problem":

<?php
$peny = floor($row->price*1000) - floor($row->price)*1000)/10;
?>

benrr101 at gmail dot com (11-Jun-2008 05:39)

But, if you want the number closest to zero, you could use this:
<?php
  if($foo > 0) {
    floor($foo);
  } else {
    ceil($foo);
  }
?>

-benrr101

greene dot mc at removethispart dot gmail dot com (27-May-2008 09:08)

I believe this behavior of the floor function was intended.  Note that it says "the next lowest integer".  -1 is "higher" than -1.6.  As in, -1 is logically greater than -1.6.  To go lower the floor function would go to -2 which is logically less than -1.6.

Floor isn't trying to give you the number closest to zero, it's giving you the lowest bounding integer of a float.

In reply to Glen who commented:
 Glen
01-Dec-2007 04:22
<?php
  echo floor(1.6);  // will output "1"
  echo floor(-1.6); // will output "-2"
?>

instead use intval (seems to work v5.1.6):

<?php
  echo intval(1.6);  // will output "1"
  echo intval(-1.6); // will output "-1"
?>

jay at w3prodigy dot com (25-Mar-2008 08:01)

Note:

<?php
$int = 0.99999999999999999;
echo floor($int); // returns 1
?>

and

<?php
$int = 0.9999999999999999;
echo floor($int); // returns 0
?>

Glen (01-Dec-2007 09:22)

<?php
  echo floor(1.6);  // will output "1"
  echo floor(-1.6); // will output "-2"
?>

instead use intval (seems to work v5.1.6):

<?php
  echo intval(1.6);  // will output "1"
  echo intval(-1.6); // will output "-1"
?>

jolyon at mways dot co dot uk (10-Aug-2004 05:41)

Beware of FLOAT weirdness!

Floats have a mind of their own, and what may look like an integer stored in a float isn't.

Here's a baffling example of how floor can be tripped up by this:

<?php
$price = 79.99;

print $price."\r\n";     // correct result, 79.99 shown

$price = $price * 100;

print $price."\r\n";    // correct result, 7999 shown

print floor($price);    // 7998 shown! what's going on?
?>

The thing to remember here is that the way a float stores a value makes it very easy for these kind of things to happen. When the 79.99 was multiplied by 100, the actual value stored in the float was probably something like 7998.9999999999999999999999999999999999, PHP would print out 7999 when the value is displayed but floor would therefore round this down to 7998.

THe moral of this story - never use float for anything that needs to be accurate! If you're doing prices for products or a shopping cart, then always use an integer and store prices as a number of pence, you'll thank me for this later :)

illyena at musefish dot net (12-May-2004 11:53)

For calculating the number of days, hours, minutes and seconds to an event.

<?php
$then = date(mktime(8,0,0,6,25,2004)); //remember that mktime is hour,min,sec,month,day,year
$now = date("U"); // "U" is the number of seconds since the epoch, equivilant to using "YmdHis"

$time = $then - $now; //gets the number of seconds between now and the event
$days = floor($time/86400); //rounds down to the whole number, in this case # of days
echo $days." Days";
$time = $time - ($days*86400); //leaves you with the amount of time ramaining after subtracting the days
$hours = floor($time/3600); //rounds down to the whole number, in this case # of hours
echo $hours." Hours";
$time = $time - ($hours*3600); //leaves you with the amount of time ramaining after subtracting the hours
$min = floor($time/60); //rounds down to the whole number, in this case # of minutes
echo $min." Minutes";
$sec = $time - ($min*60); //leaves you with the amount of time ramaining after subtracting the minutes which is equivilant to the remainins seconds
echo $sec." Seconds";
?>

PHP Helper (20-Dec-2003 10:39)

floor basically truncates, or chops off everything to the right of a decimal. For instance, if you have a length of 5.1234, but just wanted the whole number, you could use the following code:

<?php
$length = 5.1234; //this is our original length
$length = floor($length); //length is truncated, original variable name is kept
print "$length"; //this prints our result
?>

This code would print the following: 5

Now, although there is a specific function in PHP for rounding, rounding can also be performed with the floor function. Let's say we wanted to round the length to the hundredths place.

<?php
$length = 5.1234;
$length = floor(($length) * 100 + .5) * .01;
print "$length";
?>

The result is: 5.12

This works because, first, the length is multiplied by 100, which moves the decimal point to the right two places, giving us the value of 512.34. Next .5 is added to the length, which gives us a value of 512.84. Then the floor function truncates the value, leaving us with 512. Lastly, to compensate for multiplying by 100 earlier, now we must divide by 100, or in this case, multiply by .01. This moves the decimal point back 2 places to it's original place and gives us the rounded value of 5.12.

We can also round to other values, such as the thousandths, by adjusting the code as follows:

<?php
$length = 5.1234;
$length = floor(($length) * 1000 + .5) * .001;
print "$length";
?>

Result: 5.123

(10-Sep-2002 11:08)

mathematical functions lack a floating point version of the modulo operation, which returns the difference between the floor() of the argument and the argument itself:

<?php
function fmod($value) {
  return $value - floor($value);
}
?>

Very useful with trigonometric functions to reduce the angle argument to a circle that includes angle 0.

Useful also to reduce an arbitrarily large floating point value into an entropy source, by first transforming this value into a pair using logarithm functions with distinct bases (add 1 if the function can return 0, to avoid floating point errors with logarithms!):

<?php
$f = 1 + @disk_free_space("/tmp");
$r = (int)(fmod(Log($f)) * 0x7FFFFFFF)
^ (int)(fmod(Log10($f)) * 0x7FFFFFFF)
?>

Then $r can be used as a good entropy source, if the free space in your temporary folder used by PHP is constantly evolving within a large range of values.

You can combine this value by xoring it with other values such as time(), (int)microtime(), ip2long($_SERVER['REMOTE_ADDR'], $_SERVER['REMOTE_PORT'], getmypid(), ...