Floating number precision

Asked Jan 07 '17 at 18:02

Active Jan 07 '17 at 18:02

Viewed 45 times

I started wondering why floating numbers sometimes behave in such a strange way. I found out that it is because a number has to be represented with either 32 or 64 bits, so at some point the number needs to be simplified and rounded up.

So I experimented a bit on Python to get a better idea of how floats behave:

1-0.000000000000000001 == 1 returns True as expected.

However 0.000000000000000001 == 0 and 0.9999999999999999 == 1 return False.

Remarkably 1-0.000000000000000001 == 0.9999999999999999 also returns False

Is there something I am missing?

asked Jan 07 '17 at 18:02

Eleftheria Chatziargyriou

1

Not necessarily an exact duplicate, but reading that post will likely answer your question. – Carcigenicate Jan 07 '17 at 18:03
Did you try the repeating 9s with one more 9? – Mark Ransom Jan 07 '17 at 18:09
Think about 1/3 presented with decimal notation. Here you have decimal value in binary notation. You can't store that if you have limited space. – Michał Zaborowski Jan 07 '17 at 18:12

Floating number precision

0 Answers0