I've got an error in my Matlab looks like a bug

Hi everybody I made a code like this :

m=(0:0.1:1.8);
m==0.6

and it returns ans =

1×19 logical array

   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0

but m(7) returns ans =

    0.6000

I've tried this with 'find' function and I've got the same, by the way my friend also checked if they are doubles or not but they both are ...it looks like to be a Matlab's bug can you check this in your own ? ...any help will be appreciated.... thank you

Answer 1

[OCDER]

https://www.mathworks.com/help/matlab/matlab_prog/floating-point-numbers.html Look at "Avoiding Common Problems with Floating-Point Arithmetic"

It's caused by a rounding error. Not a matlab bug, but an issue with any computer floating point math.

m=(0:0.1:1.8);
m==0.6
format longe
m(7)
ans =
   6.000000000000001e-01  %that "1" at the way end is causing the m(7) == 0.6 to fail.

You need to use some sort of tolerance check

ismembertol(m, 0.6, eps*3)
ans =
1×19 logical array
0   0   0   0   0   0   1   0   0   0   0   0   0   0   0   0   0   0   0

Answer 2

[Guillaume]

It's not a matlab bug. It is expected behaviour of floating point numbers that computers use and you need to learn the pitfall associated with these.

It's also a FAQ. See faq why is 0.3-0.2-0.1 not equal to zero

The solution, as explained in the FAQ is never to compare floating point numbers with == but compare their absolute difference against an arbitrary small value.