$number = 3000/365*365-3000
echo $number

Output:
4,54747350886E-13

No, Just No. I really like you PHP but thats supposed to be zero. I don't got time for your Tantrum's. I got work to do!

PS: Does anyone know why this happens? Solved it by rounding on 10 decimals but prefer it would just answer 0 instead of me having to force it back.

in java it gives -80.

@kimailis and with floats?

so write it like this
(3000/365)*365-3000
and you will be fine?!

My guess is that it rounds at certain amount of decimals halfway. Which should be solved if you write 3000*365/365-3000 which is the same.

@aronmik without floats, if number declared as int, with floats it gives the same answer the you get. from what i understand it's a screwup with the binary counting system because 3000/365*365 gives 3000.00000000000005 or something like that so it has to be round down. i guess most math engines do the round down or round up like google, it gives 0.

@kimailis Yeah just found it out myself.

@elDeth I even tried this:

$number = 3000/365;
echo $number;
$number = $number*365;
echo $number;
$number = $number-3000;
echo $number;

Output:
8,21917808219
3000
4,54747350886E-13

@ctrlz That worked :) But formulas are user input. Would have to make a lot of changes to force that or sort that.

In that case just do a round of the result. Should not matter at all. Im guessing you do not need 14 decimals. @aronmik

Just to be sure... you probably know but that E-13 means *10^-13. So this basically is 0.00000000000000456355

@ctrlz Yeah, prob. Maybe the analist don't agree. But then they should find more budget/time for it. But it makes you wonder. Which other code that I wrote has this problem. Before this I did not know this was possible.

@aronmik i definitely understand the confusion. I love php but things like this make it hard to defend everything about it. Although apparently java isnt consistent either. Good luck with it!

learn a bit about the standard used to encode decimal numbers... (double précision floating point)
It's because of an approximation of representation (we can't represent exactly all the decimal numbers in binary)

I tried with my Android's calculator, it gave 5,00000000E-13...

@Lenka my android calculator result for 3,000÷365×365-3,000 is: 0

@corscheid: your Android is better than mine 🤔

Its based on data types and how computers process those numbers. This same problem exists in most languages.

Idk how mine did that without reproducing the issue. I'm using the default calculator app on the LG G4. *shrug*

neat, it does it on my android and windows also. if you're really concerned about something like this you could use

round (3000/365*365-3000,10)

it's dirty but it should work and shouldn't effect any real calculations.

@jckimble I already did That. Like I said in the rant. But thanks for the tip ^^

@aronmik look up IEEE 754 format for floating point numbers. The way they are stored and represented introduces these small errors.

@thmnmlst

Ooh I had already forgot this one :) Dirty round() trick is still doing the trick.