[RESOLVED] Prevent minus zero with number_format

[phpSimon]

Hi all

I'm comparing $total and $previoustotal, which are figures that end in two decimal places. I want to display the result with no decimal places.

$result = $total-$previoustotal;
print ("number_format($result,0)");

This works fine but if the compared numbers are the same, for example 1234.37 and 1234.37 - the result is sometimes minus zero, and sometimes zero.

I wondered if the rounding up might cause this but it doesn't seem to matter if the digits after the decimal point are over .50 or not.

I could do something like:

if (number_format($result,0) == -0) $result = 0;

... but I'm sure there's a more elegant way and I'm not sure -0 would be recognised anyway.

Any tips?

cheers

[phpSimon]

Re-reading this - maybe if I code so number_format is not applied when the result is zero:

$result = $total-$previoustotal; 

if ($result != 0) {
$result = number_format($result,0); }

print ("$result");

I'll try this but still interested in any comments!

[Weedpacket]

I'd never seen a negative zero, though I guess you could get that if you're doing floating-point arithmetic since presumably $total and $previoustotal are calculated differently. Lessee...

echo number_format(-0.0000000001,0);

Yup, there it is.

Since you want zero decimal places, you ought to be able to [man]round[/man] before using number_format.

phpSimon wrote:

maybe if I code so number_format is not applied when the result is zero:

Wouldn't work: the problem happens when $result is less than zero, not equal to it.

[phpSimon]

Excellent, thank you!

$result = round($total-$previoustotal, 2);
print ("number_format($result, 0)");

Works fine in all cases. I don't know enough about floating-point arithmetic to know if I'm using it or not. $total and $previoustotal both come from a MySQL datatabase. The fields themselves are DECIMAL(12,2).

However your ROUND advice was the answer, thanks again.

[johanafm]

As soon as you are dealing with non integers, you are dealing with either floating point or fixed point arithmetic. I don't know which, why, when or how they are used in PHP. But you should always refrain form equality comparisons when dealing with such numbers. For example

$p = 0.2 * 3;		// we know that 0.2 * 3 = 0.6 (but what does your computer say?
$j = 0.6;

// truly equal?
if ($p == $j) {
	echo $p . ' is the same as ' . $j . '<br/>';
}
else
	echo $p . ' is different from ' . $j . '<br/>';

// equal with acceptable precision?
if (($p - 0.0000001) <= $j && $j <= ($p + 0.0000001))
	echo $p . ' is the same as ' . $j . '<br/>';
else
	echo $p . ' is different from ' . $j . '<br/>';

[phpSimon]

Interesting. A bit of Googling for FP arithmetic brings up some very complicated information.

In the project I'm working on, data is extracted from and written to a MySQL table where it is stored as DECIMAL(12,2). When the data is assigned to a variable in PHP I don't know if it would be considered fixed or floating but I suppose floating by default.

Whenever I've performed mathmatical functions on this data in PHP I've always been careful to add the 'actual' data - I only use number_format to remove decimals when displaying data that requires that format. I'm now using round as well, if I'm doing an equality comparison; the only problem is in the display when it's wrong, it's not as if the data is getting saved anywhere - but best to get it right.

Thanks for your added advice.