Solutions to inaccurate floating point calculations, comparisons and rounding in PHP_php basics
- WBOYOriginal
- 2016-05-16 09:00:002192browse
Comparison of floating point calculation results
An example of floating point calculation is as follows:
$a = 0.2 0.7;
$b = 0.9;
var_dump($a == $b);
The printed result is: bool(false). In other words, the calculation result of 0.2 0.7 here is not equal to 0.9, which is obviously against our common sense.
Regarding this issue, the official PHP manual once stated: Apparently a simple decimal fraction such as 0.2 cannot be converted to an internal binary format without losing a little precision. This has to do with the fact that it is impossible to express certain decimal fractions exactly with a finite number of digits. For example, 1/3 in decimal becomes 0.3333333….
We print the above variables in double precision format:
$a = 0.2 0.7;
$b = 0.9;
printf("%0.20f", $a);
echo '
';
printf("%0.20f", $b);
The output results are as follows:
0.89999999999999991118
0.90000000000000002220
Obviously here, as floating-point data, part of its accuracy has been lost and cannot be completely accurate. So never trust that a floating-point number result is accurate to the last digit, and never compare two floating-point numbers for equality. It should be noted that this is not a problem with PHP, but a problem with the computer's internal processing of floating point numbers! The same problem will be encountered in languages such as C and JAVA.
So to compare two floating point numbers, we need to control them within the precision range we need before comparing, so use the bcadd() function to add the floating point numbers and convert the precision (to a string): p>
var_dump(bcadd(0.2,0.7,1) == 0.9); // Output: bool(true)
Floating point number rounding
In the article "PHP rounding functions ceil and floor", there is an example:
echo ceil(2.1/0.7); // Output: 4
?>
After the above discussion on floating-point number calculations, we know that this is caused by the inaccurate floating-point number calculation results:
printf("%0.20f", (2.1/0.7)); // Output: 3.00000000000000044409
?>
After the above discussion of floating-point number calculations, we know that this is caused by the inaccurate floating-point number calculation results, so we can use the round() function to deal with it:
echo ceil( round((2.1/0.7),1) );
?>
Although the round() function rounds according to the specified precision, retaining one decimal place has no effect on our rounding result.