0.9-0.89 is 0.01000000​0000000000​2081668171​17217

Hi, why

0.9-0.89 is 0.010000000000000000208166817117217?

I want to use it in:

if a-b<c sprintf('error') end......

when I use vpa, than I get error in if..

thank you

Answer 1

[Image Analyst]

One of the most frequently asked questions, so see the FAQ: http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

[James Tursa]

@John, For your particular example:

>> num2strexact(0.9)
ans =
0.90000000000000002220446049250313080847263336181640625
>> num2strexact(0.89)
ans =
0.89000000000000001332267629550187848508358001708984375
>> num2strexact(0.9-0.89)
ans =
1.00000000000000088817841970012523233890533447265625e-2

As you can see above, none of the numbers involved can be represented exactly in IEEE double precision floating point. You can find NUM2STREXACT here:

http://www.mathworks.com/matlabcentral/fileexchange/22239-num2strexact-exact-version-of-num2str

[john]

OK , thank you