TL;DR:
The PHP native function bcdiv seems to do precisely what is required, and properly.
To simply "truncate" a number, bcdiv($var, 1, 2);
where 2 is the number of decimals to preserve (and 1 is the denomenator - dividing the number by 1 allows you to simply truncate the original number to the desired decimal places)
Full Answer (for history)
This turns out to be more elusive than one might think.
After this answer was (incorrectly) upvoted quite a bit, it has come to my attention that even sprintf will round.
Rather than delete this answer, I'm turning it into a more robust explanation / discussion of each proposed solution.
number_format - Incorrect. (rounds)
Try using number format:
$var = number_format($var, 2, '.', ''); // Last two parameters are optional
echo $var;
// Outputs 2.50
If you want it to be a number, then simply type-cast to a float:
$var = (float)number_format($var, 2, '.', '');
Note: as has been pointed out in the comments, this does in fact round the number.
sprintf - incorrect. (sprintf also rounds)
If not rounding the number is important, then per the answer below, use sprintf:
$var = sprintf("%01.2f", $var);
floor - not quite! (floor rounds negative numbers)
floor, with some math, will come close to doing what you want:
floor(2.56789 * 100) / 100; // 2.56
Where 100 represents the precision you want. If you wanted it to three digits, then:
floor(2.56789 * 1000) / 1000; // 2.567
However, this has a problem with negative numbers. Negative numbers still get rounded, rather than truncated:
floor(-2.56789 * 100) / 100; // -2.57
"Old" Correct answer: function utilizing floor
So a fully robust solution requires a function:
function truncate_number( $number, $precision = 2) {
// Zero causes issues, and no need to truncate
if ( 0 == (int)$number ) {
return $number;
}
// Are we negative?
$negative = $number / abs($number);
// Cast the number to a positive to solve rounding
$number = abs($number);
// Calculate precision number for dividing / multiplying
$precision = pow(10, $precision);
// Run the math, re-applying the negative value to ensure returns correctly negative / positive
return floor( $number * $precision ) / $precision * $negative;
}
Results from the above function:
echo truncate_number(2.56789, 1); // 2.5
echo truncate_number(2.56789); // 2.56
echo truncate_number(2.56789, 3); // 2.567
echo truncate_number(-2.56789, 1); // -2.5
echo truncate_number(-2.56789); // -2.56
echo truncate_number(-2.56789, 3); // -2.567
New Correct Answer
Use the PHP native function bcdiv
echo bcdiv(2.56789, 1, 1); // 2.5
echo bcdiv(2.56789, 1, 2); // 2.56
echo bcdiv(2.56789, 1, 3); // 2.567
echo bcdiv(-2.56789, 1, 1); // -2.5
echo bcdiv(-2.56789, 1, 2); // -2.56
echo bcdiv(-2.56789, 1, 3); // -2.567