php arithmetic errors - modulo

Question

I am using PHP 5.5 (5.5.28) I have come across a really weird problem with PHP arithmetic

$value=29; $divisor=10;
$mod=$value%$divisor;
echo "$mod<br>";

returns 9 (as expected)

$value=(0.29*100); $divisor=(0.1*100);
$mod=$value%$divisor;
echo "$mod<br>";

returns 8 (????)

I can't find anything about this in the documentation Obviously its relatively easy to write a replacement function and use that However, the problem seems to be bigger, in that I am getting weird rounding errors which mean my ledger entries in the project I am working on are not balancing

Any suggestions?

Read, learn and inwardly digest, and read the big warning message on this PHP Docs page.... if you're writing a ledger, you should never truust the precision of floats in any programming language, always using integer — Mark Baker
– Mark Baker, Commented Sep 12, 2016 at 7:13
Thanks Mark ... naive of me to assume that common sense would apply — user3600150
– user3600150, Commented Sep 12, 2016 at 7:28
Common Sense? There's a reason infinitely large decimal values can't be stored in a fixed number of binary bytes, and you are duly presented with a large warning when you visit the page concerned! — Ben Hillier
– Ben Hillier, Commented Sep 12, 2016 at 7:35
Common sense would be to use the right type of opernand for the operator you are using and % expects integers. According to the manual: Operands of modulus are converted to integers (by stripping the decimal part) before processing.. You could of course also use fmod() - that actually returns 9 in your test case - but as @MarkBaker already mentioned, you should never trust the precision of floats when working with ledgers. — jeroen
– jeroen, Commented Sep 12, 2016 at 7:42
This isn't anything to do with common sense.... it's common across almost every programming language; and exactly the same problem exists when human beings use decimal, there are numbers such as 1/3 that cannot be represented in decimal, so we always work with "estimates" — Mark Baker
– Mark Baker, Commented Sep 12, 2016 at 8:06

Paul Coldrey · Accepted Answer · 2016-09-12 07:17:57Z

2

This is almost-certainly related to the way that floating point numbers are stored and you have the issue where .29*100 is not EXACTLY equal to 29 and 0.1*100 is not EXACTLY equal to 10. Hence if one is slightly smaller or larger than you expected then you get a result that is out by one. If you print your intermediate values out with 10 decimal place you will probably be able to observe the cause

This is exactly the same issue you have when trying to compare equality of floats and you end up having to do something like: if(fabs(a-b) < 0.0001) rather than saying: if(a==b)

answered Sep 12, 2016 at 7:17

Paul Coldrey

1,44913 silver badges20 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

php arithmetic errors - modulo

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related