PHP精度解决方案

http://php.net/manual/zh/function.bcadd.php

上述网址中有对于精度的解决方案.

取一个数的精度,可以采用该数与0相加后取对应的精度的解决方案

如下

bcadd('1.23456789',0,5);

上面网址还有个笔记写的不错,是一个函数

在下面转一下该笔记

I made this to add an unlimited size of numbers together.. 

This could be useful for those without the BCMath extension. 

It allows decimals, and optional $Scale parameter.  If $Scale isn't specified, then it'll automatically adjust to show the correct number of decimals. 

<?php 

function Add($Num1,$Num2,$Scale=null) { 
  // check if they're valid positive numbers, extract the whole numbers and decimals 
  if(!preg_match("/^\+?(\d+)(\.\d+)?$/",$Num1,$Tmp1)|| 
     !preg_match("/^\+?(\d+)(\.\d+)?$/",$Num2,$Tmp2)) return('0'); 

  // this is where the result is stored 
  $Output=array(); 

  // remove ending zeroes from decimals and remove point 
  $Dec1=isset($Tmp1[2])?rtrim(substr($Tmp1[2],1),'0'):''; 
  $Dec2=isset($Tmp2[2])?rtrim(substr($Tmp2[2],1),'0'):''; 

  // calculate the longest length of decimals 
  $DLen=max(strlen($Dec1),strlen($Dec2)); 

  // if $Scale is null, automatically set it to the amount of decimal places for accuracy 
  if($Scale==null) $Scale=$DLen; 

  // remove leading zeroes and reverse the whole numbers, then append padded decimals on the end 
  $Num1=strrev(ltrim($Tmp1[1],'0').str_pad($Dec1,$DLen,'0')); 
  $Num2=strrev(ltrim($Tmp2[1],'0').str_pad($Dec2,$DLen,'0')); 

  // calculate the longest length we need to process 
  $MLen=max(strlen($Num1),strlen($Num2)); 

  // pad the two numbers so they are of equal length (both equal to $MLen) 
  $Num1=str_pad($Num1,$MLen,'0'); 
  $Num2=str_pad($Num2,$MLen,'0'); 

  // process each digit, keep the ones, carry the tens (remainders) 
  for($i=0;$i<$MLen;$i++) { 
    $Sum=((int)$Num1{$i}+(int)$Num2{$i}); 
    if(isset($Output[$i])) $Sum+=$Output[$i]; 
    $Output[$i]=$Sum%10; 
    if($Sum>9) $Output[$i+1]=1; 
  } 

  // convert the array to string and reverse it 
  $Output=strrev(implode($Output)); 

  // substring the decimal digits from the result, pad if necessary (if $Scale > amount of actual decimals) 
  // next, since actual zero values can cause a problem with the substring values, if so, just simply give '0' 
  // next, append the decimal value, if $Scale is defined, and return result 
  $Decimal=str_pad(substr($Output,-$DLen,$Scale),$Scale,'0'); 
  $Output=(($MLen-$DLen<1)?'0':substr($Output,0,-$DLen)); 
  $Output.=(($Scale>0)?".{$Decimal}":''); 
  return($Output); 
} 

$A="5650175242.508133742"; 
$B="308437806.831153821478770"; 

printf("  Add(%s,%s);\r\n// %s\r\n\r\n",$A,$B,  Add($A,$B)); 
printf("BCAdd(%s,%s);\r\n// %s\r\n\r\n",$A,$B,BCAdd($A,$B)); 

/* 
  This will produce the following.. 
    Add(5650175242.508133742,308437806.831153821478770); 
  // 5958613049.33928756347877 

  BCAdd(5650175242.508133742,308437806.831153821478770); 
  // 5958613049 
*/ 

?> 

It was a fun experience making, and thought I'd share it. 
Enjoy, 
Nitrogen.